Medicare: A Strategy for Quality Assurance, Volume II: Sources and Methods (1990)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

8 The Utilization and Quality Control Peer Review Organization Program Kathleen N. Lohr and Allison I. Walker INTRODUCTION In the early 1980s, the Utilization and Quality Control Peer Review Or- ganization (PRO) program replaced the Professional Standards Review Organization (PSRO) program as the Medicare peer review effort. As with the PSROs, the purpose of the PROs is to ensure that services rendered through Medicare are necessary, appropriate, and of high quality. PRO activities extend well beyond those emphases into many aspects of the Medicare program, including implementation of the Medicare diagnosis- related groups (DRG) prospective payment system (PPS) for hospitals. PROs serve different purposes for different parties, not all of whom have the same interests or concerns In the words of one interested observer, " . . . PROs are quickly becoming all things to all people . . . " (Webber, cited in GIG, 1988b). The program is important to the Office of Inspector General (OIG) in the Department of Health and Human Services (DHHS), the Executive Office of Management and Budget, the Medicare beneficiary community (and consumer groups more generally), and providers of many sorts (espe- cially local practicing physicians). This chapter describes the development of the PRO program up until the summer of 1989, with most emphasis being placed on the required activities that reflect the considerable expansion and complexity of responsibility and activity of the program. For more extended discussions of the early im- plementation of the PRO program, see Lohr (1985) or Dans et al. (1985~. Volume I, Chapter 6 is excerpted extensively from this chapter, although the committee conclusions and recommendations in that chapter are not given here. 343

344 KATHLEEN N. LOHR AND AI BISON J. WALKER PRO LEGISLATION AND REGULATIONS Legislation The role of Congress in the development of the PRO program should not be underestimated. Several pieces of legislation successively broadened the mandate of the program, often beyond the capacity of the Health Care Financing Administration (HCFA) to implement them as expected; they also attempted to turn the program toward a greater emphasis on quality of care. The key act was the Tax Equity and Fiscal Responsibility Act (TEFRA) of 1982 (more specifically, the Peer Review Improvement Act, Title I, Subtitle C of TEFRA), which amended Part B of Title XI of the Social Security Act. Other important legislation included the Social Security Amendments of 1983; the Deficit Reduction Act of 1984 (DEFRA); the Consolidated Omnibus Budget Reconciliation Act (COBRA) of 1985; the Omnibus Budget Reconciliation Acts (OBRA) of 1986, 1987, and 1989; and the Medicare and Medicaid Patient Program Protection Act of 1987. TEFRA (P.L. 97-248) established the PRO program. It changed the funding arrangement from a system of federal grants to a 2-year fixed-price competitive contract, and it extended eligibility for the PRO program to for- profit groups and to payer organizations such as insurers and fiscal interme- diaries (FIB) if there is no other available entity. It also strengthened the PROs' ability to sanction providers, thus improving the potential effective- ness of the program. The original 195 PSRO areas were consolidated into 54 areas. The Social Security Amendments of 1983 (P.L. 98-21) established Medicare's PPS scheme for hospital payment, which increased the number of required activities of the PROs as well as their visibility. The legislation required each PPS hospital to contract with a PRO as a condition of partici- pation in the Medicare program. PROs were assigned many mandatory review tasks, mostly relating to monitoring the behavior of hospitals fol- lowing PPS; the most obvious concern was increased unnecessary admis- sions. Required activities of PROs were aimed chiefly at the necessity and appropriateness of admissions and invasive procedures rather than inade- quate or poor technical care. As part of DEFRA (P.L. 98-369), PROs became the quality and utiliza- tion monitors for the new reimbursement system established by the Social Security Amendments of 1983. This act provided for the continued funding of PSROs from the Medicare Trust Fund until PRO contracts could be signed. It also extended the deadline for hospitals to sign a contract with a PRO. COBRA 1985 (P.L. 99-272) and OBRA 1986 (P.L. 99-509) considerably expanded PRO responsibilities. COBRA mandated pre-admission and pre

UTILIZATION QUME CONTROL PRO PROGRAM 345 procedure review of certain surgical procedures (based on studies of medi- cal practice variations) and pre-procedure review of cases requiring surgical assistants at cataract surgery. It also gave PROs authority to deny Medicare payment to physicians and hospitals for substandard quality (which would come to be known as "quality denials". This authority was expected to complement, not compete with, the PROs' ability to sanction providers in other ways. OBRA 1989 clarified the quality denial activity somewhat. OBRA 1986 considerably extended PRO review responsibility beyond the inpatient setting to include review of the following: ambulatory ser- vices; services provided in hospital outpatient departments and ambulatory surgical centers (ASCs); care rendered in skilled nursing facilities (SNFs); care rendered by home health agencies (HHAs); inpatient and outpatient care from Medicare risk-contract health maintenance organizations (HMOs) and competitive medical plans (CMPs); and all written complaints from Medicare beneficiaries. These activities were to be phased in over several years. OBRA 1987 (P.L. 100-203) extended contract cycles from 2 to 3 years and allowed a contract extension (up to 2 years) for existing contracts to achieve more efficient renewals. Accordingly, a contract renewal phase-in schedule (generally referred to as the third Scope of Work) was developed in which one-fourth of the PROs began the new contract cycle on October 1, 1988, and the remaining PROs on April l, 1989. OBRA 1987 made numerous other changes in the PRO program, including a mandated mini- mum level of on-site review of rural hospital care and greater emphasis on education and other instructional activities for rural practitioners. The Medicare and Medicaid Patient Program Protection Act of 1987 (P.L. 93-100) expanded both sanction and civil monetary penalty authorities for those programs. It also required the reporting of disciplinary actions made by state medical licensure boards to the Secretary of DHHS, with the latter being responsible for disseminating information on these actions to state boards and to other state and federal officials. It did not, however, mandate that PRO s report their disciplinary actions to state boards. Regulations Apart from legislative acts, numerous regulations and other directives govern the administration and operation of PROB. The Administrative Pro- cedure Act (APA) requires that regulations be promulgated through notice and comment rulemaking procedures. HCFA considers the APA procedures to apply to some but not all of its directives and publishes in the Federal Register those procedures for which it seeks public comment. As an alter- native or adjunct to the rather cumbersome regulatory rulemaking mecha- nism, HCFA also relies extensively on PRO Manual transmittals, contracts

346 KATHLEEN N. LOHR ARID ALLISON J. WALKER and contract modifications (and their Scopes of Work), and other, less for- mal instructions. Administration HCFA's Health Standards and Quality Bureau (HSQB) administers the PRO program. HSQB functions are carried out in both the Central Office (CO) and in the 10 HCFA Regional Offices (ROs). CO staff establish policies for the program, perform a great deal of data analysis, negotiate PRO contracts (with help from RO staff), and set evaluation criteria. RO staff have appreciable involvement in PRO activities. They transmit pro- gram requirements to PROs, oversee implementation of those requirements, evaluate, and generally provide oversight and technical assistance. Some of HCFA's guidance for the PRO program is promulgated in this way; thus, inconsistencies among the ROs in interpreting HCFA policies and guide- lines can pose a considerable problem for the program (OIG, 1989~. PRO ORGANIZATIONAL CHARACTERISTICS In organizing the PRO program, one change from the earlier PSRO pro- gram was to consolidate the many PSRO regions into 54 PROs (all the states, the District of Columbia, Puerto Rico, the Virgin Islands, and a combined area of American Samoa, Guam, and the Commonwealth of the Marianas). In a legislative attempt to retain some semblance of "local peer review," Congress specified that, to qualify as a PRO, a statewide organi- zation must either demonstrate sponsorship by being composed of at least 10 percent of the physicians practicing in the area (physician-sponsored organizations) or have at least one physician in every generally recognized specialty in the area available for PRO review (physician-access organiza- tions). Third-party payers can obtain PRO contracts if no other eligible organization is available; at the time of this study only one such organiza- tion had a PRO contract. A ~ ~ ~it. ~.~. . . . . . ~ row may not De a neaten care rac~l~ty or other entity subject to review. This avoids financial conflicts of interest with providers that may be the subject of review. The contractor must have at least one consumer repre- sentative on its governing board and must operate with objectivity and without apparent or real conflict of interest. PRO CONTRACTS Unlike their PSRO predecessors, PROs are financed by contracts, not grants. In principle, contracts make the program more manageable centrally and more consistent nationally, because the contracting agency can specify

UTINZATION ED QUME CONTROL PRO PROGRAM 347 in great detail precisely what it expects its contractors to do and can then evaluate them on how well they meet those contract specifications. PROs carry out a very complex set of review and intervention tasks that are specified in minute detail in their HCFA contracts. The contracting proce- dure is based on a formal Request for Proposal (RFP), which includes a "scope of work" (SOW) that becomes part of the contract between the government and the PRO. PROs began with 2-year contracts, and an effort was made to have all PROs start more or less at the same time. This proved very difficult to implement. (The first contracts became effective over a 5-month phase-in period from July to November 1984 that corresponded to the implementa- tion of Medicare's PPS.) Therefore, OBRA 1987 extended contract periods to 3 years, to permit somewhat more stability in anticipated financing and planning. Timing of PRO contract periods is now staggered so that HCFA does not have to negotiate 54 contracts simultaneously. Contracts can be renewed triennially or canceled and put up for competi- tive bidding if the existing PRO is judged not to qualify for a noncompeti- tive renewal. In some cases these contracts have been won by the original PRO because no competition emerged or because the bidding process galva- nized the existing PRO into a credible renewal effort. PRO contracts may be terminated by either the PRO or the secretary of DHHS. The secretary may terminate or choose not to renew a PRO contract when officials determine that the PRO has not met or is not meeting its obligations in a satisfactory manner. A complex set of procedures is speci- fied by which termination or nonrenewal can be accomplished. The secretary's decisions in this regard are not subject to judicial review, and the secretary has the absolute right to terminate a contract (rightfully or wrongfully) without the possibility of the decision being overturned later in court. In addition to contracts between the PRO and DHHS, PROs must main- tain written agreements with hospitals and with FIs and carriers. In hospital agreements, PROs must include their review plans, criteria, and procedures (including frequency of reviews, documentation to be required of the hospi- tal, and time and location of reviews). The review process must be consis- tent with the requirements placed on PROs in their own contracts with lICFA, and the PRO-hospital agreements must be coterminous with PRO- HCFA contracts. The agreements may be revised as needed to conform with changes in statutes, regulations, and HCFA policies and directives for the PRO program, meaning that changes to the PRO program can directly and materially affect the hospital industry. PROs must also enter into agreements with the FIs serving providers in their areas. These agreements specify procedures to coordinate review ac- tivities of the respective organizations. PROs are responsible for establish

348 KATHLEEN N. LOHR AND ALLISON J. WALKER ing procedures to collect and process data in order to assure that they are complete, accurate, and promptly reported. FIs must provide internally consistent and prompt data each month so that PROs can conduct timely reviews, and they are responsible for both the internal consistency and promptness in delivery of the information that they are required to produce. PROs are not permitted to collect or have collected for them any informa- tion that duplicates information that FIs are responsible for, although the PROs can negotiate to purchase data not currently collected by FIs. These relationships among PROs, hospitals, and FIs emphasize commu- nication and data sharing. PROs must determine the accuracy of informa- tion that hospitals provide to their FIs, and PROs and FIs must establish a system of data sharing that permits PROs to inform FIs of data lacunae and inaccuracies and then obtain completed and corrected information. After completing their various reviews, PROs must also report to both hospitals and FIs any claims that require payment adjustments. PRO SCOPES OF WORK The SOW details the specific obligations of the PRO that will be incor- porated into its contract. It defines the duties and functions of the Medicare review for a specific contract cycle. The first SOW was used during the first contract cycle (1984-1986~; the second during the 1986-1988 contract cycle; and the third covers the present period. No new SOWs are contem- plated. To reinforce the sense of stability and common expectations for the program, all future changes will be made through a contract modification process. All PROs will be expected to implement changes at the same time. If HCFA considers the contract modification to be significant, the agency will publish it in the Federal Register 30 days in advance of its intended start date. In the event that the contract modification requires more fund- ing, PROs and HCFA will have to agree on the additional level of support before the PROs begin the work. Although most PRO activities have remained fairly constant over the three SOWs, some tasks have changed dramatically. For example, the first SOW emphasized controlling inappropriate utilization, whereas the second and third SOWs direct more attention to assuring quality. The second and third SOWs remained fairly similar as a result of efforts to achieve consis- tency with minimum disruption to ongoing review activities; much of the second SOW remains in the third, but with variations in the size of sarnples.2 All three SOWs are described in the following sections, but the third SOW is described in greater detail because it is the current guide for PRO work. Table 8.1 outlines the main activities of the third SOW, and Table 8.2 compares the three SOWs on certain key review and other requirements.

UTILIZATION AND QUALITY CONTROL PRO PROGRAM TABLE 8.1 Elements of Required Peer Review Organization (PRO) Activities for the Third Scope of Work 349 I. Prospective Payment System (PPS) Hospital Casesa A. Random (the 3-percene sample) B. Transfers 1. PPS to PPS hospitals 2. PPS to exempt psychiatric units 3. PPS to exempt swing beds C. Readmissions in less than 31 days from discharge from a PPS hospital with review of intervening care 1. PPS hospital readmission a. Identifying all readmissions b. Review a random 25-percent hospital-specific sample 2. Intervening care a. Identify all cases in the 25-percent sample with care rendered by skilled nursing facilities, home health agencies, or hospital outpatient departments b. Review a 20-percent sample of each hospital's intervening care universe for quality of care (not medical necessity or overuse of services), with HCFA's generic quality screens D. Focused DRGs (100 percent review of DRGs 385-391, 472, 474, 475; 50 percent review of DRG 468; 25 percent review of DRG 462)b E. Day and cost outliers (25 percent random samples) F. Medicare code editor (12 principal diagnoses)C G. H. I. Hospital adjustments (any adjustments to higher weighted DRGs) Noncovered admissions (with covered level of care later in stay) PI and HCFA regional office referrals II. Specialty Hospitals A. Exempt units of PPS hospitals B. Exempt hospitals III. Ambulatory Surgery [Hospital Outpatient Areas and Ambulatory Surgical Centers (ASCs)] IV. Intensified Review V. Pre-admission and Pre-procedure reviews A. Ten procedures B. Assistants at cataract surgery VI. Review of Freestanding Cardiac Catheterization Facilities VII. Objectives (e.g., based on Generic Quality Screens) VIII. Development and Use of Explicit Written Criteria TABLE 8.1 continues

350 TABLE 8.1 Continued KATHLEEN N. LOHR AND AllISON J. WALKER IX. Reconsideration and Review of DRG Changes and Preparing Appeals Folders X. Data A. Reports submitted to HCFA on completed reviews B. Profiling 1. Hospital statistics (by 14 variables) 2. Physician statistics (by 4 variables) 3. Other provider statistics (HHA, SNFs, ASCs) 4. Internal quality control (monitoring of review decisions) XI. Beneficiary Communications A. Important Message to Medicare Beneficiaries (from hospitals) B. Hospital notices of noncoverage C. Community outreach 1. Hotline 2. Written inquiries responses 3. Education programs, seminars, and workshops 4. Informational materials 5. Coordination with beneficiary groups XII. Responsiveness to Inquiries and Complaints XIII. Interaction with Physicians and Providers A. Peer review B. Opportunity for consultation C. Education D. Criteria development and dissemination E. Communications F. Confidentiality and disclosure guidelines G. External relationships with concerned organizations H. Management responsibilities XIV. Sanctions XV. Confidentiality and Disclosure of Information XVI. Fraud and Abuse Review (of Cases referred by OIG or HCFA) XVII. Anti-Dumping Review (of Cases referred by HCFA) XVIII. Private Review XIX. Civilian Health and Medical Programs of the Uniformed Services (CHAMPUS) XX. Other Requirements A. Cooperation win HCFA B. Cooperation win tile SuperPRO C. Private review D. Internal quality control

UTILIZATION AND BUMS CONTROL PRO PROGRAM 351 aThe required review activities include: generic quality screens' discharge review, admission review, invasive procedure review' DRG validation, coverage review, and waiver of liability. bThe DRG categories are as follows: 385, neonates, died or transferred; 386, extreme immaturity, neonates; 387, prematurity with major problems; 388, prematurity without major problems; 389, full-term neonate with major problems; 390, neonate with other significant problems; 391, normal newborn; 462, rehabili- tation; 468, unrelated operating room procedures; 472, extensive burns; 474, acheostomy; and 475, mecharucal ventilation through endo~acheal incubation. CDiabetes mellitus, without mention of complication; noninsulin dependent and insulin dependent; obesity; impacted cerumen; benign hypertension; left bundle branch hemiblock; other bundle bench hemiblock; positive SRL/VRL HL3; elevated blood pressure reading without diagnosis of hypertension; other and unspecified complications of medical care, not elsewhere specified; and cardiac pacemaker (fitting and adjustment). Carotid endarterectomy and cataract procedures are required. Eight of the following 11 can also be selected: cholecystectomy, major joint replacement, coronary artery bypass graft, percutaneous transluminal coronary angioplasty, larninectomy, complex peripheral revascularization, hysterectomy, bunionectomy, inguinal hernia repay, prostatectomy, and pacemaker insertion. SOURCE: Attachment 33, HCFA, 1988. FIRST PRO SCOPE OF WORK (1984-1986 CONTRACT CYCLE) The first PRO contracts emphasized the detection of inappropriate utili- zation and payments under the new Medicare hospital PPS after October 1983. Contract activities, which concentrated on inpatient hospital care, included reducing unnecessary admissions, ensuring that payment rates matched diagnostic and procedural information contained in the patient rec- ords, and reviewing patients who were transferred or readmitted to an acute care hospital within 7 days of discharge. In addition, HCFA negotiated five generic "quality objectives" for each PRO: (1) reduce unnecessary hospital readmissions resulting from substandard care provided during the prior admissions; (2) assure the provision of medical services which, when not performed, have significant potential for causing serious patient complica- tions; (3) reduce the risk of mortality associated with selected procedures and/or conditions requiring hospitalization (initially denoted "reduce avoid- able deaths; (4) reduce unnecessary surgery or other invasive procedures; and (5) reduce avoidable postoperative or other complications. Corrective actions for physicians or hospitals included education and consultation, intensified review, and denial of payment for inappropriate or unnecessary admissions or readmissions. During the first SOW, PROs were

352 KATHLEEN N. WHR AND AI1ISON J. WALKER TABLE 8.2 Companson of the Three Scopes of Work (SOWs) with Respect to Selected Utilization and Quality Control Peer Review Organization (PRO) Activities (Ordered by Tasks Pertaining to the Third SOW)a I. Prospective Payment System Hospitals Cases Random Samples First SOW: 5-percent admission sample. DRG sample ranging from 3 to 100 percent based on hospital discharge size Second SOW: 3-percent sample (includes 1- and 2-day stays) Third SOW: Same as second SOW Transfers First SOW: From PPS to another hospital, exempt unit, or swing bed Second SOW: Same as first SOW, but lower level of review Third SOW: PPS to PPS, 50-percent sample; PPS to psychiatric, 10 percent; and PPS to swing bed, 25 percent Readmissions First SOW: All related readmissions within 7 days of discharge Second SOW: All related readmissions within 15 days of discharge Third SOW: 25 percent of readmissions within 31 days of discharge Intervening Care First SOW: Not in scope of work Second SOW: Not in scope of work Third SOW: 20 percent of all cases receiving home health agency, hospital outpatient, inpatient, or skilled nursing facility (SNF) care between sampled hospital admissions less than 31 days apart Focused DRGs First SOW: Review of DRG numbers 462 and 46B Second SOW: Review of DRG numbers 462, 468, and 088 Third SOW: Review of DRG numbers 462, 468, 385-391 472 and 474 475b . Day and Cost Outliers First SOW: Originally 100 percent; reduced to 50 percent during contact period Second SOW: 50 percent of day and cost outliers Third SOW: 25 percent of day and cost outliers Medicare Code Editor First SOW: 100 percent of 9 diagnoses with code editor rejects Second SOW: Same as first SOW Third SOW: 100 percent of 12 diagnoses with code editor rejectsC Hospital Adjustments First SOW: 100 percent of all cases adjusted to a higher-weighted DRG Second SOW: Same as first SOW Third SOW: Same as second SOW PI and HCFA Regional Office Referrals First SOW: 100 percent review of cases referred by FI or HCFA regional office for determination of medical necessity Second SOW: Same as first SOW Third SOW: Same as second SOW

UTILIZATION kD QUME CONTROL PRO PROGRAM 353 II. Specialty Hospital Review First SOW: Proposed by each PRO Second SOW: 15 percent of all discharges Third SOW: 15-percent random sample for PPS-exempt hospitals and units III. Ambulatory Surgery First SOW: Not in scope of work Second SOW: Not in scope of work Third SOW: 5-percent random sample of all cases Intensified Review First SOW: Trigger; 2.5 percent or 3 cases reviewed (whichever is greater). Review increased to 100 percent or subsets. Second SOW: Trigger; 5 percent or 6 cases reviewed (whichever is greater). Review increased to 100 percent or subsets (two consecutive quarters) Third SOW: Same as second SOW Preadmission Review First SOW: 5 procedures proposed by each PRO Second SOW: Pacemaker plus 4 procedures proposed by the PRO Third SOW: 100 percent of 10 procedures (cataract extraction, cartoid endarterectomy plus 8 of 11 others specified by HCFA)& Assistants at Cataract Surgery First SOW: Not in scope of work Second SOW: 100 percent review of cases for medical necessity of assistant at surgery Third SOW: Same as second SOW VII. Objectives First SOW: Three admission objectives arid five quality objectives. All proposed and validated by the PRO. Very limited areas for focusing . . O Electives Second SOW: Five objectives based on PRO data from first 90 days of generic quality screen review. HCPA-identified mortality rate outliers Third SOW: Objectives based on data from generic screens. May be statewide, or focused by physician, DRG, provider, etc. Hospital Notices of Noncoverage First SOW: 100 percent where patient or physician disagrees; 100 percent where patient is liable; 10-percent random sample Second SOW: Same as first SOW Third SOW: 100 percent where patient or physician disagrees; 100 percent where patient is liable Community Outreach First SOW: Not in scope of work Second SOW: Each PRO to propose its own program Third SOW: Minimum requirements to be met aRoman numerals refer to parts in Table 8.1. bFor definitions, see Table 8.1. CFor definitions, see Table 8.1. Nor listing, see Table 8.1. SOURCE: Adapted from unpublished HCFA documents.

354 KATHLEEN N. LI:)HR AND AI1ISON]. WAVER authorized to recommend sanction of providers or physicians to the GIG in one of two instances: (1) cases of a "substantial violation [of their Medicare obligations] in a substantial number of cases"3 and (2) single cases of a "gross and flagrant" violation.4 These categories of sanctionable problems continue to the present. The sanction process per se, however, has become more complex in response to legal and political challenges. SECOND PRO SCOPE OF WORK (1986-1988 CONTRACT CYCLE) During the 198~1988 SOW, eight PRO s reviewed two areas and one reviewed three. The following PROs conducted reviews in two states: West Virginia for Delaware; Maryland for the District of Columbia; Hawaii for Guam-American Samoa; Indiana for Kentucky; Rhode Island for Maine; Iowa for Nebraska; New Hampshire for Vermont; and Montana for Wyo- ming. The PRO for Washington State also reviewed Alaska and Idaho (OIG, 1988b). New or Expanded Responsibilities The second SOW emphasized quality review in addition to the detection of inappropriate utilization. This SOW lowered the level of review for hospitals with acceptable performance and increased it for hospitals with unacceptable performance. Contract goals were developed as guidelines for focusing quality review activities; these objectives were based on actual results of the first 90 days of review using generic ("occurrence") screens that were applied to a 3-percent random sample of medical records. Other new areas of activity in the second SOW included the following: reporting the results of review of short hospital stays in the routine sample; review of related readmissions to the same hospital within 15 days (up from 7 days in the first SOW); retrospective review of all sampled cases to assess the appropriateness of the discharge; review of hospitals that were high- mortality rate outliers in the first release of HCFA mortality rate data; an emphasis on statistically identifiable adverse outcomes such as premature discharge and death; and development of a community outreach program. Generic Quality Screens Hospital generic quality screens were the most visible addition to the second SOW (see Table 8.3~. Generic screens are widely used to detect what are regarded as the most common causes or manifestations of quality problems, and were adopted for use in the PRO program for all charts under review by the PRO for any reason. HCFA introduced them in the fall of 1986 without pilot-testing and issued interpretative guidelines in May of 1987. PROs were also permitted to develop their own screens to cover

UTIl1ZATION AND QUALITY CONTROL PRO PROGRAM 355 "adverse occurrences" that may reflect regional variations in practice pat- terns (and some PROs did), but locally devised screens could not change the intent of the federal screens. Generic quality screens are applied by nurse reviewers. They can either determine that no screen has been failed (or that a screen has been failed but that no potential quality problem exists) or refer the case to a physician advisor for further evaluation. Initially, nurse reviewers had to refer all screen failures to physician advisors; this produced considerable numbers of false-positive cases and appreciable frustration and anger for reviewers and the medical community. HCFA later permitted PRO nurse reviewers to use their professional judgment in not forwarding some screen failures for physician review. Only the physician advisor can "confirm" a quality prob- lem. The six required generic quality screens for the second SOW are as follows (see Table 8.3 for details and subitems): 1. adequacy of discharge planning 2. medical stability of patient at discharge 3. unexpected deaths 4. nosocomial infections 5. unscheduled return to surgery 6. trauma suffered in hospital THIRD PRO SCOPE OF WORK (1988-1990 CONTRACT CYCLE) The third SOW incorporates provisions from COBRA, OBRA 1986, and OBRA 1987. Phase-in of the third SOW took place over several months; all PROs were expected to be on the third SOW as of April 1, 1989. In making the shift to 3-year contracts, HCFA put PROs into one of four categories. Two of those categories (28 states in all) were to be awarded full 3-year contracts; the remaining categories (26 states and territories) were to be awarded either 6- or 12-month extensions before negotiation of their 3-year contracts. As of summer 19%9, however, four PRO contracts had not been awarded. Table 8.1 outlines the major required activities for care rendered in the fee-for-service (FFS) system, and some are described in more detail below. The focus is on inpatient hospital review, but the activities do not differ appreciably for nonhospital practitioners or settings. Required Review Activities for Hospital Inpatient Care The required PRO review activities for all inpatient hospital cases re- viewed are (1) generic quality screening, (2) discharge review, (3) admis- sion review, (4) invasive procedure review, (5) DRG validation, (6) coverage

356 KATHLEEN N. LOHR AND ALLISON J. WALKER TABLE 8.3 Generic Quality Screens Hospital inpatient 1. Adequacy of Discharge Planninga No documentation of discharge planning or appropriate followup care win consideration of physical, emotional and mental status needs at time of discharge. 2. Medical Stability of the Patient at Discharge a. Blood Pressure within 24 hours of discharge (systolic less than 85 or greater than 180; diastolic less than 50 or greater than 100) b. Temperature within 24 hours of discharge greater than 101 degrees Fahrenheit (38.3 Centigrade) oral, greater than 102 degrees Fahrenheit (38.9 Centigrade) rectal c. Pulse less than 50 (or 45 if the patient is on a beta blocker), or greater than 120 within 24 hours of discharge d. Abnormal diagnostic findings which are not addressed and resolved or where the record does not explain why they are not resolved e. Intravenous fluids or drugs after 12 midnight on day of discharge I. Purulent or bloody drainage of wound or open area within 24 hours prior to discharge 3. Deaths a. During or following any surgery performed during the current admission b. Following return to intensive care unit, coronary care or other special care unit within 24 hours of being transferred out c. Other unexpected death 4. Nosocomial Infectiona (Hospital-acquired infection) 5. Unscheduled Return to Surgery Within same admission for same condition as previous surgery or to correct operative problem 6. Trauma Suffered in the Hospital a. Unplanned surgery which includes, but is not limited to, removal or repair of a normal organ or body part (i.e., surgery not addressed specifically in the operative consent) b. Falla c. Serious complications of anesthesia d. Any transfusion error or serious transfusion reaction e. Hospital-acquired decubitus ulcer and/or deterioration of an existing decubitusa

UTILIZATION AND QUALITYCONTROLPRO PROGRAM 357 I. Medication error or adverse drug reaction (1) wide serious potential for harm or (2) resulting in measures to correct g. Care or lack of care resulting in serious or potentially serious complica- tions "Optional Screen" Medication or treatment changes (including discontinuation) within 24 hours of discharge without adequate observation aPeer review organization (PRO) reviewer is to record the failure of the screen, but need not refer potential severity Level I quality problems to physician reviewer until a pattern emerges. SOURCE: HCFA, 1988. review, and (7) determination of the application of the waiver of liability provision. These are described in more detail below. The cases selected for review through the 3-percent random sampling process plus cases under review for other reasons constitute almost 25 per- cent of all Medicare admissions. A cumulative data summary of PRO activity to the end of February 1989 gives the following national figures for combined retrospective review and pre-admission and prepayment review: of a universe of bills and cases numbering 26,747,451, a total of 6,407,967 were selected for review (24 percent); 6,993,179 were reviewed (26 per- cent).s The estimated number of reviews for the third SOW totals 10,541,730, including 7,600,006 hospital reviews, 877,739 HMO and CMP reviews, and 2,063,985 ambulatory surgery reviews (HCFA, l989b); the estimates are reached partly by projecting expected Medicare admissions in the many different categories and applying the review sampling percentages to those . . projections. Generic Screens In addition to the six generic screens from the second SOW, PROs can use an optional screen for medication or treatment changes (including dis- continuation) within 24 hours of discharge without adequate observation. The third SOW has also added an adequacy-of-care screen to the set of "trauma" screens to cover inappropriate or untimely assessment, interven- tion, or management resulting in serious or potentially serious complica- tions. As before, these are applied to every case under PRO review for whatever reason. Figures 8.1a and 8.lb illustrate the full process. With respect to generic screens, most of the major changes (from scope to scope) occur in the interpretive guidelines for the screens rather than in

358 KATHLEEWN. LOHR AND AI1ISON J. WALKER FIGURE 8.1a Overview of the Quality Review Process for Inpatient Hospital, Home Health Agency, and Outpatient Surgery Generic Screensa 1. 2. Nurse reviewer identifies screen failure Nurse reviewer makes quality determination ~I Potential quality problem 3. Severity Level I Case pended Quarterly profiling Threshold of 3 cases/quarter or 5 6. Severity Level II and III cases/bi-quarter \ Physician Advisor Review Potential quality problem No quality problems STOP 7. Refer to specialist or committee (if applicable) Potential quality problem l 8. Notice to responsible party(ies) for discussion 9. Confirmation/Nonconfirmation of quality problem 10. Notice of final determination to responsible party(ies) 11. No quality problem STOP No quality problem STOP Severity of weight assigned; Quality intervention process initiated aIncludes inpatient hospital screens 1, 4, 6b, and be and certain home health agency and outpatient surgery screens. SOURCE: HCFA, 1989a.

UTILIZATION AND QUALITY CONTROL PRO PROGRAM FIGURE 8.1b Overview of the Quality Review Process for Other Ge- nenc Screensa 359 1. Nurse Reviewer identifies a potential quality problem not related to generic quality screens 1 1 2. Referred to physician advisor for review Potential quality problem 3. Severity Level I Case Pended 5. Quarterly profiling Threshold of 1 case/qtr. Or 5 cases/bi-qtr. Severity Level STOP II or III 6. Refer to specialist or committee (if applicable) + Potential quality problem 7. Notice to responsible party(ies) 8. Confirmation/nonconfirmation of quality problem 9. Notice of final determination to responsible cartv(ies) Nurse reviewer identifies a quality screen failure other than those listed above Potential quality No quality Problem problem 1 STOP No quality problem ~ r No quality problem STOP ~ . , ~ , 10. Severity ~ Eight assigned: Quality intervention process initiated aIncludes inpatient hospital, home health agency, and outpatient surgery generic screens not covered by the process in Figure 8.1a. SOURCE: HCFA, 1989a.

360 KATHLEEN N. LOHR AND A~;ISON J. WALKER the screens themselves.6 Prompted by the finding that screens were not being uniformly applied across the country, starting in the fall of 1987 HCFA held a series of consensus-building meetings to discuss the use of inpatient hospital screens. The groups consisted of PRO medical staff, RO medical staff, and CO representatives. From this series of local meetings, two representatives from each meeting were asked to participate in a na- tional task force. That group revised the screens, especially by adding clarifying guidelines, to bring more consistent definitions to the application of the screens. These refined screens were implemented in October 1988. Other Required Review Activities Discharge review identifies potential problems with premature discharge of two types: (1) the patient was not medically stable at discharge or (2) discharge was not consistent with the patient's continued need for acute inpatient care. This task relates in part to concerns about possible unin- tended consequences of DRG-based PPS. Retrospective admission review identifies whether inpatient hospital care was medically necessary and appropriate; it involves reviewing reasons for admission against pre-established criteria devised or adopted by individual PROs and subjective physician review. Admission review in particular is a traditional utilization review effort that can be traced back to PSRO days and beyond. As a PRO requirement, it relates to concerns that PPS incen- tives would also induce hospitals to admit patients who were not sick enough to warrant hospitalization (i.e., for whom a full DRG payment would be made but who would require far less in hospital costs). Invasive procedure review retrospectively examines the medical neces- sity of invasive procedures that affect the assignment of a case to one DRG rather than another (which means virtually all invasive procedures done in the hospital setting). The review is applied to cases already selected for review, not to additional cases. If the procedure is not medically necessary or is noncovered, and if the procedure was the sole reason for admission, then payments for the admission and the procedure are denied. If the procedure is not medically necessary and is noncovered, but other reason- able and necessary services were provided and the admission was medically necessary, then only the physiciants payment for the procedure is denied and the DRG is changed. Throughout the program, the purpose of DRG validation has been to assure that Medicare payments under PPS are correct, because it was imme- diately (and correctly) understood that DRG coding was susceptible to some manipulation and slow upward "creep" to higher-weighted DRGs over time (Simborg, 1981; GAO, 1988a; Hsia et al., 1988~. A Registered Record Administrator or an Accredited Record Technician generally has the overall responsibility for the PRO's validation process. The result of DRG valida

UTILIZATION AND QUALITY COlITROL PRO PROGRAM 361 lion can be to leave the DRG unchanged or to upgrade or downgrade it, thereby affecting the hospital payment. This activity also includes review to assure that the requirements for physician attestation are met. Physician attestation requires physicians to sign a statement certifying that their narrative descriptions of the principal and secondary diagnoses and the major procedures are accurate and com- plete to the best of their knowledge. Originally, physicians were expected to sign such a statement for every discharge, but it rapidly became clear that this was a considerable burden and an implicit indictment of physicians; the requirement is now that physicians sign a statement once a year and that hospitals keep that affirmation on file. Coverage review determines whether items or services normally excluded from Medicare coverage are medically necessary. This review is performed only in instances when coverage can be extended for specific items and circumstances if certain medical conditions are met. In PSRO days, hospitals and over providers were assumed to have a favorable Bewailer of liability" status; waiver of liability meant that unless a hospital "knew or could reasonably have been expected to know" that the care it was providing was unnecessary, the costs of that care would still be reimbursed and the hospital was not financially liable. Only if the hospital's waiver was revoked or if the PSRO determined on a case-by-case basis that the provider could have been expected to know that the case was unneces- sary would it become financially at risk for days of care or services pro- vided to a beneficiary. Now, under the waiver of liability (also referred to as limitation of liability)' the favorable presumptive status has been elimi- nated, and the PRO must determine whether the beneficiary or provider should be held liable for care not covered under Medicare because either the beneficiary or provider knew, or could reasonably have been expected to know, that such care was not covered. Other Aspects of Hospital Inpatient Review In addition to the above required review activities, several other hospital inpatient tasks are now either expanded or required. Each PRO will also be required to publish (at least annually) a report that describes their findings about care that was unnecessary, was inappropriate, was given in an inap- propriate setting, or did not meet professionally accepted standards; this report is to be widely distributed to providers and practitioners whose serv- ices are subject to review. Pre-admission and Pre-procedure Review PROs are also required to review 10 procedures, generally on a pre- adrnission or pre-procedure basis, for necessity and for appropriateness of

362 KATH~F:EN N. LOHR AND ALLISON J. WALKER setting (e.g., inpatient or ambulatory). They must review all proposed ca- rotid endarterectomy and cataract procedures. In addition, they must either pick an additional 8 procedures from a list of 11 supplied by HCFA or document why a procedure not on the list should be subjected to 100 per- cent pre-admission review. The 11 conditions are cholecystectomy, major joint replacement, coronary artery bypass graft, percutaneous transluminal coronary angioplasty, laminectomy, complex peripheral revascularization, hysterectomy, bunionectomy, inguinal hernia repair, prostatectomy, and pacemaker insertion. PROs establish their own prior-authorization criteria for this purpose, sometimes in consultation with local or state physician specialty groups, and some sharing of criteria does take place across PROB. Nevertheless, PROs differ in the types of clinical factors or levels of patient functioning that they require to be present (or absent) before they will approve the procedure.7 Rural Providers Rural providers have vigorously asserted that they are not reviewed by "local peers" and that their style of practice and the constraints under which they function are not well appreciated or taken into account. To respond to these concerns and bring peer review more fully into areas that were rarely visited, the third SOW mandates that at least 20 percent of all rural hospi- tals shall be reviewed on-site. According to OBRA 1987, rural physicians (namely, those in counties of 70,000 or fewer residents or in officially designated rural health manpower shortage areas) are also given special protection during the sanctioning process. Except when the physician or provider is found to pose a "serious risk" to Medicare enrollees, exclusions are to be put on hold until full hearings before an Administrative Law Judge (ALJ) have been conducted. Some effort is being mounted to extend these protections to all physicians who live in counties of 140~000 or fewer residents, a move that would effectively extend special appeal rights to about two-thirds of practicing physicians (Vibbert, 1989d). Nonhospital Review Review of Skilled Nursing Facility and Home Health Agency Care: Intervening Care PROs have been directed to undertake review in several nonhospital set- tings. By and large, this review has not been very comprehensive. The main effort to date has been in response to OBRA 1986 requirements for

UTILIZATION AND QUALm CONTROL PRO PROGRAM 363 PROs to assess the adequacy of post-acute care. PROs are expected to review a small sample of cases receiving "intervening care," mainly care delivered by home health agencies (HHAs) and skilled nursing facilities (SNFs) between two related hospital admissions up to 31 days apart. (Care rendered in hospital outpatient areas, or HOPAs, is also included.) This effort had minimal pilot project pre-testing. Pilot studies in Massa- chusetts and Pennsylvania were conducted from mid-1987 and mid-1988, respectively. These efforts did not yield much in the way of quality defi- ciencies. By definition, this task does not cover the great bulk of post-acute care, because it focuses only on the interval between hospital admissions. With respect to SNF review, little is being done in the PRO program, but some experience has been gained in this area. PSROs reviewed care pro- vided in long-term-care facilities; at the peak, about 55 PSRO projects were under way in such facilities. A special effort assessed 10 PSRO demonstra- tion projects in long-term-care review. The assessment included pre-admis- sion, admission, and continued-stay review; quality assurance activities; Medical Care Evaluation studies (MCEs); and data systems development (Kane et al., 1979~. The net conclusion of the evaluators was that the progress made by the 10 demonstrations justified optimism "that the PSRO has potential for affecting the appropriateness of utilization and the quality of nursing home care . . . [and that] it is imperative to preserve the incentive for PSROs to enter the long-term care review field" (Kane et al., 1979, p. . .N Vll). Despite this injunction, PRO attention to this sector of the health care delivery system has been scant, in part because PROs (unlike PSROs) were mandated to review only Medicare (not Medicaid) services. PROs are al- lowed to contract with state Medicaid agencies to do Medicaid review, and as of mid-l989, 34 PROs had such contacts (H. Brook, personal communi- cation, 1989~. However, this does not represent a systematic federal effort by PRO s to review the care of nursing home residents. Review in Other Fee-for-Service Settings Initiatives in the other settings, especially ambulatory physician office care, are getting under way mainly as pilot projects. Plans for these proj- ects are described later. Nonhospital Generic Screens For the third SOW, generic quality screens have been developed to re- view care rendered in the following settings: (1) HHAs; (2) SNFs; (3) hos- pital outpatient departments (HOPDs); and (4) ambulatory surgery centers (ASCs) (see Table 8.4~. Also under development are screens for reviewing

364 KATHLEEONN. LOHR AND ALL;fSON J. WALKER TABLE 8.4 Nonhospital Generic Screen Categories Home Health Agencies 1. Adequacy of intake evaluation 2. Appropriate and timely interventions 3. Adequacy of restorative care (specialty therapies and nursing instructions) 4. Deaths within 48 hours of transfer to hospital 5. Possible indications of secondary infections 6. Issues related to patient care after the home health start of care 7. Documentation plan for appropriate follow-up care and discharge summary to physicians of record 8. Any other events or patterns of care resulting in adverse outcomes that should be evaluated Skilled Nursing Facilities 1. Appropriateness of hospital discharge to SNF and appropriateness of admission to and continued stay in that SNF 2. Appropriate notification, response, and further evaluation within specified time frames (e.g., 4 hours to 24 hours) for a series of problems (e.g., medications, vital signs, fall, infection, hydration and nutrition, mobility, pressure sores, etc.) 3. Abnormal results of diagnostic services addressed, resolved, or explained 4. Assessment, plan of care, evaluation, and discharge plan for appropriate disciplines 5. Deaths following transfer to hospital 6. Any other events or patterns of care resulting in adverse outcomes that should be evaluated Hospital Outpatient Departments 1. Stake assessment 2. Appropriate and timely interventions 3. Specialty therapies 4. Patient teaching 5. Death within 48 hours of admission to hospital 6. Issues related to provision of patient care 7. Documented plan for appropriate follow-up care or discharge 8. Any other events or patterns of care resulting in adverse outcomes that should be evaluated Outpatient Surgery 1. Adequate pre-operative assessment 2. Appropriate and timely interventions during surgery for significant and sustained problems 3. Issues relating to post-operative care 4. Appropriate documented discharge plan with provision for follow-up care 5. Adequate patient education SOURCE: Attachments, HCFA, 1988.

UTILIZATION AND QUALrlYCONTROLPRO PROGRAM 365 psychiatric care and rehabilitative services; the psychiatric screens were issued to all PROs in November 1989 and the rehabilitation screens are scheduled to be completed in fiscal year (FY) 1990. The new screens are similar to inpatient generic screens, but they are supposed to be more relevant to the particular setting. For example, the SNF screens deal with polyphannacy (multiple medication) issues and the mental stability of the patient. Although the original set of generic screens was developed by the contractor for SuperPRO (SysteMetrics, Inc.) in con- junction with HCFA, providers were heavily involved in development of the new screens for use in nonhospital settings. RESPONSES TO QUALITY OR UTILIZATION PROBLEMS: NOTIFICATIONS, CORRECTIVE ACTIONS, AND SANCTIONS PROs can pursue several possible interventions when they have con- firmed a quality or utilization problem. They can notify practitioners or providers of problems, put them on "intensified" review, require a wide variety of corrective actions, or institute sanction procedures. These re- sponses to problems are described in connection with the quality interven- tion plan required by the third SOW, but they have been available since the start of the PRO program. Quality Intervention Plan The quality intervention plan (QIP) is "a prescribed blueprint which re- quires PROs to implement specific interventions in response to confirmed quality problems" (Federal Register, 1989, p. 1966~. The QIP is new to the third SOW, was developed in an effort to achieve consistency among PROs, and is based on minimum requirements set by HCFA. It includes determi- nation of the source of the problem; assignment of "severity levels" and weights to quality problems; establishment of a time frame for completion of the review process; profiling; and quality interventions that are related to severity levels. Determination of the Source of the Problem All initial case reviews are completed by a nurse reviewer. If a nurse reviewer identifies a potential quality problem, then a physician advisor reviews the case. The nurse reviewer can determine that a potential quality problem does not exist, but only a physician advisor can determine that a quality problem does exist. After the PRO decides a quality-of-care prob- lem exists, it determines the source of the problem, such as an individual physician or hospital.8 After an opportunity for discussion, the medical director or the quality assurance committee at the PRO assigns the severity level.

366 Assignment of Severity Levels and Weights KATHLEEN N. LOHR AND ALLISON J. WALKER Severity levels are a way to categorize quality problems. They are as- signed according to the nature of the problem and the potential for causing adverse patient outcomes. The relevant phrase, significant adverse effects, is defined as unnecessarily prolonged treatment, complication, readmission, or patient management that results in anatomical or physiological impair- ment, disability, or death. Weights (numerical points) assigned to severity levels indicate when PROs must take various corrective steps. The levels (with weights in parentheses) and definitions are as follows: Severity Level I (1~: Medical mismanagement without We potential for significant adverse effects on the patients Severity Level II (51: Medical mismanagement with the potential for sig- nificant adverse effects on the patient Severity Level III (25~: Medical mismanagement wide significant adverse ef- fects on the patient Time Frames for Quality Review HCFA set a maximum time frame for quality review. If a potential Severity Level I problem exists, the case is held in a pending status until a pattern of problems emerges. For Severity Levels II and III, the maximum time frame for completion of the review is 135 days. To be timely, overall review must be completed within the specified periods. Maximum time frames are as follows. For Severity Levels I to III, receipt of FI data tapes through the initial physician advisor-reviewer deter- mination can take only 60 days. For Severity Levels II and III, there are three stages. Review by specialist or quality committee, if applicable, can take no more than 30 days. If the case has been determined to involve a potential quality problem, notification of the attending physician or pro- vider (or both) and opportunity for discussion can take no more than an- other 30 days. Finally, confirmation or nonconfirmation of a quality prob- lem and final notification to the attending physician or provider (or both) can take no more than 15 days. The sum (60 + 30 + 30 + 15) is 135 days. Profiling The purpose of profiling for the QIP is to identify areas for focused review or other corrective action. The PRO is required to produce a quar- terly profile of every physician and provider for every case it reviews. It must also profile (1) potential quality problems with a Severity Level I, by the source of the problem (e.g., provider or physician), to determine whether

urILlzATIoN AND QuALrryCONTROLPRO PROGRAM 367 the threshold levels of 1 case per quarter or 5 cases per 2 quarters (some- times called a bi-quarter) are reached; (2) a weighted severity level score for confirmed quality problems, by We source of the problem (i.e., provider or physician); and (3) all confirmed quality problems, by source of the problem. Quality Interventions When a PRO identifies a quality problem resulting from case misman- agement, then the PRO develops a corrective action plan using a variety of interventions. The six types of interventions listed below must be included in its contract review plan, but the PRO may add more. 1. Notification. The notice of the final determination must include a description of the confirmed quality problem, what the appropriate action should have been, the severity level, and what interventions will be taken. 2. Educational efforts. These include telephone and/or in-person discus- sions with the responsible parties, suggested literature reading, continuing medical education (CME) courses, and self-education courses. 3. Intensification of review through 100 percent retrospective review of all cases or intensified review of a focused subsample. 4. Other interventions include concurrent or pre-discharge review; prior approval or pre-admission review; and referral to hospital committees (e.g., infection control, tissue, quality assurance committees). 5. Coordination with licensing and accreditation bodies. The PRO must disclose confidential information to state and federal licensing bodies upon request when such information is required by those entities to carry out their legal functions, and the PRO may do so even without a request (e.g., when a practitioner or provider has reached a weighted score of 25 points in 1 quarter). 6. Sanction plans (discussed below). The PRO must use certain thresholds, called weighted triggers, to decide what intervention it should use. The threshold points are as follows: 1. Notification 2. Education 3. Intensification 4. Other interventions 5. Consideration of coordination with licensing bodies 6. Consideration of sanctions 1 per quarter or 5 per bi-quarter 10 15 20 25

368 KATHLEEN N. LOHR AND AI1ISON J. WALKER When cases have multiple problems in more than one severity level, the PRO uses the highest severity level and weighted trigger. These scores are computed at least quarterly for each problem source (e.g., physician or hospital); they are simply sums of the total points of Severity Levels I, II, and III given earlier. For instance, three Level II problems would produce a total score of 15; three Level II plus six Level I problems would produce a total score of 21. The SOW allows the PRO some flexibility to take inter- ventions before a trigger (e.g., a severity score of 25) is reached or to apply lower weighted interventions in special circumstances. For the last two interventions (coordination and sanctions), the PRO must consider but need not invoke them as long as it documents why it did not. Sanctions and Other Interventions The Secretary of DHHS, not the PROs, holds the authority to impose sanctions on Medicare providers, and the secretary has delegated that au- thority to the Office of Inspector General (OIG). The PROs' power lies in making sanction recommendations to the GIG. Figures 8.2a and 8.2b dia- gram this process. PRO Responsibilities The PRO recommends sanctions in either of two instances: (1) cases of "substantial violation" of a provider's Medicare obligations "in a substan- tial number of cases" and (2) a single case of a "gross and flagrant" viola- tion. If a PRO recommends exclusion, and the GIG does not act on that recommendation within 120 days, the exclusion automatically goes into effect until a final determination is made. No regulations define the criteria to be used by a PRO in determining whether a practitioner or provider has violated a Medicare obligation. The preamble to the PRO regulations states only that PROs must apply profes- sionally developed standards of care, diagnosis, and treatment based on typical patterns of practice in their geographic areas (Federal Register, 1985~. The PRO Manual also contains some material on the elements of a sanc- tionable offense. For cases in which the PRO determines that the provider or physician has failed to comply substantially with a Medicare obligation in a substantial number of cases, it sends the practitioner or provider an initial sanction notice.~° This notice gives the recipient 20 days to respond to the notifica- tion with additional information or to request a meeting with the PRO. If, after considering the additional information, the PRO confirms its original finding, it develops a corrective plan of action. If the practitioner or pro- vider fails to comply with that plan, the PRO sends a second sanction

UTILIZATION AND QUAL17Y CONTROL PRO PROGRAM 369 notice. In such cases, the provider or practitioner has a second opportunity to submit additional information or discuss the problem with the PRO (within 30 days of the second notice). If the concern is not resolved, the procedures at this point follow the pattern for gross and flagrant violations. Several specific procedures direct how the PRO should forward its recommendation to the OIG and how it should notify the individual or organization that is has done so, what the recommended sanction is, and how further information can be forwarded directly to the OIG (again within 30 days). The PRO must also give the practitioner or provider a copy of the material it used in reaching its deci- sion. At this point, the responsible sanctioning party is the OIG, not the PRO. OIG Responsibilities The OIG must determine whether the PRO followed appropriate proce- dures, whether a violation occurred, whether the provider has "demonstrated an unwillingness or lack of ability substantially to comply with statutory obligations" (known as the "willing and able" provision), and ultimately whether it agrees with the PRO recommendation (OIG, 198Bb). In these determinations, the OIG is expected to consider the type and severity of the offense, the previous sanction record, prior problems that Medicare may have had with the individual or institution, and the availability of alterna- tive medical resources in the community. In response to the PRO recom- mendation, the OIG can sustain the finding, alter the recommendation, or reject it. (See OIG, 1988b for a more complete description of this process.) If the OIG does not accept the PRO's recommendation, the sanctioning process stops. If the OIG does accept the PRO's recommendation, it must give notice that the sanction is to be imposed, effective 15 days after the notice is received by the practitioner or provider. The OIG notifies the public by placing a notice in a newspaper of general circulation in the individual's or institutionts locality. It also informs state Medicaid fraud control units and state licensing bodies, hospitals and other facilities where the practitioner has privileges, medical societies, carriers, FIs, and HMOs. DHHS Responsibilities The sanctioned provider or physician may appeal the OIG decision to an ALJ, who conducts a separate hearing starting essentially from scratch. The sanctioned individual or institution may call witnesses to testify under oath, cross-examine witnesses, submit documents and briefs, and present oral arguments. If practitioners or providers are not satisfied with the out- come of their hearing, they can request review by DHHS's Appeals Council

370 KATHLEEN N. LOlIR AND AI1`ISON J. WAILKER FIGURE 8.2a Overview of PRO/HHS Sanction Process for Substantial Violationsa __--- ~ I Initial sanction notice I | PRO decision . _ _ - Provider has 20 days to submit additional information and/or request a meeting i- ~ Not a substantial violation PRO decides that a substantial violation has occurred ~ Provider complies with corrective plan and develops a corrective plan of action within specified time period Provider does not comply with corrective plan r - - - J - - - _ I I Second sanction notice ! | PRO decision l _ _ - _ _ - _ _ I Final sanction notice; PRO recommendation on I I decision to Office of the Inspector General of HHS I | HHS Office of the Inspector General decision | ~ Do not sanction Provider appeals sanction to an administrative law judge of HHS - HHS Administrative Law Judge decision Provider has 30 days to submit additional information andJor request a meeting Not a substantial violation Provider has 30 days to submit additional information to Office of the Inspector General Provider accepts sanction Dismiss sanction Sustain (or modify) sanction ~ Provider accepts sanction Provider appeals sanction to the HHS Appeals Council Secretary of HHS Appeals Council decision Dismiss sanction ~- Sustain (or modify) sanction > Provider accepts sanction Provider seeks judicial review of HHS Appeals Council's decision to sustain sanction Court decision _ Sustain sanction | > Dismiss sanction  aA substantial violation is a pattern of care over a substantial number of cases that is inappropriate, unnecessary, does not meet recognized patterns of care, or is not supported by the documentation of care required by the PRO. SOURCE: Adapted with permission from OTA, 1988.

UTILIZATION AND QUALITY CONTROL PRO PROGRAM FIGURE 8.2b Overview of PRO/HHS Sanction Process for Flagrant , Violationsa ~ - ~ I Initial sanction notice I PRO decision | __ ! _ - r - - - ~ - _ _ _ _ _l I Final sanction notice; PRO recommendation on I I decision to Office of the Inspector General of HHS I | HHS Office of the 1~ Do not sanction I Sanction Provider appeals sanction to an administrative law judge of HHS L HHS Administrative Law Judge decision 371 Provider has 30 days to submit additional information and/or request a meeting Not a gross and flagrant violation Provider has 30 days to submit additional information to Office of the Inspector General Provider accepts sanction Dismiss sanction Sustain (or modify) sanction - ---- - ~ Provider accepts sanction Provider appeals sanction to the HHS Appeals Council r - - Secretary of HHS Appeals Council decision ---r --- _ I Sustain (or modify) sanction Provider seeks judicial review of HHS Appeals Council's decision to sustain sanction Court de_ _ Sustain sanction in- ~ Dismiss sanction Provider accepts sanction Dismiss sanction aA flagrant violation is a violation that has occurred in one or more instances and presents an imminent danger to the health, safety, or well-being of a Medicare beneficiary. SOURCE: Adapted with permission from OTA, 1988.

372 KATHLEEN N. LOHR AND ALLISON J. WALKER and then still seek judicial review of the decision at the level of a federal district court. If the OIG proceeds successfully through these steps, the Secretary of DHHS, through the OIG, can apply two kinds of formal sanctions: (1) exclusion from the Medicare program (which may be multi-year in dura- tion); and (2) monetary sanctions (which at present cannot exceed the cost of the services that were rendered). Successfully sanctioned (i.e., excluded) hospitals and providers must petition to be reinstated in the Medicare pro- grarn, and they can receive no payment for services rendered or items pro- vided during the exclusion period. Historical Record of Interventions and Sanctions The most frequent PRO intervention appears to be the formal letter of notification. By contrast, intensified review, formal education or similar programs, and sanction recommendations are used much less often, although during the second SOW more than 53 percent of hospitals were under inten- sified review for at least 1 quarter (HCFA, 1989c). PROs differ markedly in the rates at which they invoke various interventions. For instance, GAO (1988b) cites the following two ranges for letters of notification: 0 to 111 letters per 1,000 "new" physicians, and 0 to 396 letters per 1,000 "repeat" . . P nyslclans. PRO activity. Tables 8.5, a, b, and c, summarize intervention activity tabulated by HCFA for the second SOW, which is the most recent aggregate information. Of the more than 6.6 million completed reviews (mainly for the second SOW), PROs denied payment in over 4 percent of cases (includ- ing partial denials); the range across PROs was 1.2 to 25.5 percent. For about 33 percent of the denials the practitioner or provider requested a reconsideration (range, 0.6 to 69.6 percent). Of those reconsiderations, the denials were reversed in 44 percent of the cases (range, 15.1 percent to 100 percent); that is, the original decision was upheld 56 percent of the time. Through early 1989, the PROs had identified more than 87,000 physi- cians with some level of quality problem (Table 8.5b). Over 81,400 of those problems had been resolved, presumably through the more than 70,000 quality interventions carried out (HCFA, 1989c). HCFA data compiled from the start of the program through June 1989 shows that 43 PROs had sent a total of 1,065 first notices, the vast majority to physicians rather than hospitals (Table 8.5c). For physicians, more no- tices were for gross and flagrant violations than for substantial violations; the opposite was true for hospitals. The PROs had also recommended a total of 119 sanctions to the OIG, the vast majority for gross and flagrant violations by physicians. Many of the

UTILIZATION AlID BUMS CONTROL PRO PRO GRAD TABLE 8.5a Quality Intervention Activities of Peer Review Organizations (PROs) Through June 1989: Reconsiderations 373 Type of Action Number Percent of Completed Reviews Percent of Percent of Reconsiderations Denials Requested Completed reviews Payment denials Reconsiderations requested Reconsiderations upheld 6,655,505 278,294 4.2 91,268 1.4 32.8 51,252 0.8 18.4 56.2 SOURCE: HCFA, 1989b. TABLE 8.5b Quality Intervention Activities of PROs Through February 1989: Quality Interventions for Physicians Category Number of Cases - Newly Identified Repeat cases Physicians with quality problems Physicians with quality problems resolved Quality interventions taken 87,075 81,440 70,321 20,598 19,888 26,871 SOURCE: HCFA, 1989c. TABLE 8.5c Quality Intervention Activities of PROs Through June 1989: Sanctions Category of Activity Number of Physicians Providers First notices sent Substantial violations Gross and flagrant violations Second notices senta Cases referred to the Office of Inspector General  Substantial violations Gross and flagrant violations 907 335 572 68 109 29 80 9 158 109 49 17 10 1 aSecond notices are sent only in cases of substantial violations. SOURCE: HCFA, 1989b.

374 KATlILE~N N. LOHR AND ALLISON J. WALKER sanction cases date from earlier years of the program. The relatively lower numbers of sanction recommendations in more recent times has generated some debate and has been attributed to three factors: (1) revisions in proce- dures (prompted by the American Medical Association) that give practitio- ners the right to counsel during discussions with PROs of possible sanc- tions; (2) OIG directives that discouraged use of monetary fines as an alter- native to exclusion; and (3) possibly the high reversal rate of the ALJs who had upheld only 8 of 18 sanctions on appeal during this period (McIlrath, 1989~. In addition to these points, the growing confusion and tension caused by mixed signals from HCFA and the OIG concerning the relative emphasis to place on educational and disciplinary approaches to PRO implementation may have played a role in the sanction-recommendation picture. OIG activity. From PY 1986 through FY 1988, the OIG reported it had received 197 referrals (150 gross and flagrant, 46 substantial, and 1 lack of documentation). Of these, 79 cases (40 percent) were rejected. Of the remainder, two cases were closed because the physician died, three physi- cians retired before exclusion, and three cases were pending. A total of 110 sanctions had been imposed (56 percent). Of the latter, 83 were exclusions (82 physicians; 1 facility) and 27 were monetary penalties (25 physicians; 2 facilities). In short, the OIG accepts about three in five sanction recom- mendations from PROs, a figure that has been fairly constant across the years. Of cases rejected, about two in five were because the case did not meet regulatory requirements, about two in five because the practitioner could show that he or she was willing and able to improve, and one in five because medical evidence was adequate. Of the sanctions imposed, the great majority were exclusions from the program. Other PRO Required Activities Beneficiary Relations The PROs are required to act on behalf of Medicare beneficiaries in four ways not directly related to the technical quality of care rendered by provid- ers or physicians. They must monitor hospital distribution of An Important Message from Medicare (concerning patients' rights to appeal denials of hospital care) and hospital notices of noncoverage, respond to beneficiary complaints, and carry on general community education and outreach. Important Message to Medicare Beneficiaries. OBRA 1986 required all hospitals to provide Medicare patients (on or soon after admission) with a statement that explains (1) their rights to benefits, (2) their rights to appeal

UTILIZATION AND QUALITY CONTROLPRO PROGRAM 375 denials of benefits, and (3) the circumstances under which they will be liable for charges for those services (should a denial of benefits be upheld upon appeal). This two-page statement, An Important Message from Medi- care, must conform to language specified in DHHS regulations except for minor changes in how receipt of the Message is acknowledged. The PRO must monitor that all hospitals give beneficiaries the Message. Corrective action will be taken if the Message is not given to the benefici- ary or his or her representative, if it is given to the beneficiary but not at or about the time of admission, or if its content is altered in any way. The SOW leaves the development of the monitoring plan for each hospital up to the PRO. Hospital notices of noncoverage. A hospital can issue notices of non- coverage to Medicare beneficiaries when it determines that the patient's care is not (or will not be) covered because it is not medically necessary, is not delivered in the appropriate setting, or is custodial. These notices can be given before the patient is admitted or at any time during the hospital stay, according to a complex set of regulations. When hospitals issue such notices, beneficiaries are assumed to have knowledge that the services are not covered; they thus become liable for customary charges incurred after various periods of time have passed (e.g., at noon of the day following receipt of the notice of noncoverage). Similarly, providers who issue no- tices of noncoverage are assumed to know that the services furnished (or proposed) are not covered. In both these cases, issues relating to waiver (limitation) of liability come into play. The hospital cannot issue these notices without the attending physician's concurrence. Either the attending physician or the patient, or both, may dispute the hospital's finding. The PRO must review cases in the following instances: (1) all hospital requests for PRO review when the attending phy- sician does not agree that care is no longer required; (2) all cases where the patient disagrees with the notice issued by the hospital; (3) all cases where the patient is liable for charges for services furnished after notification; and (4) all cases where the hospital determined that the admission was non- covered. When the attending physician disagrees with the hospital's desire to is- sue a notice of noncoverage and the hospital requests PRO review, the hospital must notify the beneficiary, in writing, that it has made that re- quest. Before the PRO makes its review determination, it must make every effort to solicit the views of the patient, the physician, and the hospital. If the PRO concurs with the hospital, it issues a denial notice, which then becomes subject to reconsideration in accordance with the usual PRO re- consideration procedures. Apart from responding to hospital or beneficiary requests for review of

376 KATHLEEN N. LOHR AND AILISON J. WAVER these notices, the main responsibilities of PROs in monitoring hospital no- tices of noncoverage are to track the content of the notice, the accuracy of the hospital's determination, and the conformance of the hospital to the specified process. The procedures, sample sizes, and other aspects of this monitoring are spelled out in the SOW. Beneficiary complaints. All written complaints about the quality of care rendered by hospitals (inpatient or outpatient), SNFs, HHAs, ASCs, HMOs, and CMPs must be investigated by PROB. The focus is on care that does not meet professionally recognized standards. PROs do not review com- plaints involving underutilization, however, because that is the function of the FIs. Review of quality-related cases is to be based on medical record information, and procedures concerning disclosure of the review, the prob- lem, and its disposition are specified in the SOW. Community outreach. PROs must conduct programs to inform benefici- aries about Medicare PRO review and PPS; more specifically about the purpose of PROs and PPS, types of PRO review, and their right to appeal a PRO determination. PRO s are also expected to devise ways to explain how they ensure the quality of care and respond to complaints from beneficiar- ies. Each PRO must develop a detailed plan indicating how it will mount five community outreach activities. In addition, PROs are encouraged (not re- quired) to use radio and television public service announcements. The minimum required areas of activity include He following: 1. Maintain a toll-free hotline to respond to beneficiary inquiries be- tween 8:00 A.M. and 4:30 P.M. every working day. 2. Respond to written inquiries within 30 days of receipt. If this time frame cannot be met, an acknowledgement of receipt of the inquiry must be sent to the beneficiary within 10 days. 3. Conduct education programs, seminars, and workshops to inform beneficiaries about PRO review, PPS, and their appeal rights. 4. Develop and disseminate informational materials (e.g., brochures, slides, and tapes) that explain PRO review, PPS, hospital notices, and ap- peal procedures. 5. Coordinate with and involve concerned beneficiary and provider groups. Provider Relations The PRO program continues to stress peer review. In detailed specifica- tions concerning what types of physicians shall review whom, the SOW appears to reflect HCFA's understanding of peer reviewed It also calls for

UTII1ZATION ED QUME CONTROL PRO PROGRAM 377 an "interaction plan" that will enhance the relationships between the PROs and providers, physicians, and other practitioners. That plan must give details concerning how physicians will be afforded opportunities to discuss problems or proposed denials and how the PRO will carry out educational efforts. The latter include informing physicians and other providers about PRO review, PPS, the rights of all parties under Medicare, and certain PRO confidentiality and disclosure guidelines. The PRO is also required to publish and disseminate (at least annually) a report that describes what it has found about care that does meet Medicare obligations (i.e., necessary, appropriate, and of acceptable professional stan- dards). This task mirrors the requirement that DHHS should submit to the Congress an annual report on the administration, impact, and cost of the program; such reports have not been published to date, however. In gen- eral, the outreach activities envisioned for providers are similar to those required for beneficiaries (seminars, informational material, etc.~. Data Acquisition, Sharing, and Reporting PROs are required to exchange information with FIs and carriers, with other PROs, and with other public or private review organizations. (Shar- ing of information with FIs was briefly described in connection with PRO contracts and the conduct of review.) The rules governing acquisition, sharing, and disclosure are complex and open to some interpretation. PRO access to information. HCFA regulations authorize PROs to have access to and obtain any records and information pertaining to health care services rendered to Medicare beneficiaries that are in the possession of any physician or provider in the PRO area. The PRO may require the practitio- ner or institution to provide copies of such information. The preamble to the relevant regulations notes that quality problems that affect Medicare patients usually affect all patients, particularly in the con- text of acute care.~4 Often a quality problem may be adequately handled for Medicare patients only by addressing it for all patients. Thus, PROs may need to review both Medicare and non-Medicare patient records to resolve the problems for the former. The regulations thus provide that PROs may gain access to non-Medicare patient records if authorized by the physician or provider. PROs are required to contact state medical licensing boards to establish mechanisms by which the state medical boards will send to the PRO the names of physicians against whom the board has taken disciplinary action. The PRO is then required to review all the cases of such practitioners (except for services provided in the physician's office) for the 3 months after notification by the board. Although this requirement has been in

378 KATHLEEN N. LOHR AND AllISON J. WA=ER effect virtually since the start of the program, by the end of the second contract cycle fewer than half of the PROs had ever had any exchanges with their state boards, and fewer than one-quarter had more than two exchanges (GAO, 1988b). The problems with information-sharing are bidirectional. Among the more common complaints that the Institute of Medicine (IOM) study com- mittee heard during its site visits (from several different sources) was the lack of communication from the PROs about problem practitioners. The rules that govern information disclosure may play a part in this problem. Disclosure of "confidential" information. Generally, information or rec- ords acquired by a PRO are confidential and not subject to disclosure. HCFA regulations distinguish between confidentially and nonconfidential information. They also limit the release of patient-identifying and physi- cian-identifying information to that required for PRO review or for other statutorily required reasons. Disclosure of information that does not iden- tify individuals is permitted. PROs are required to disclose certain confidential information to appro- priate agencies if the PRO believes that not to do so would pose a risk to public health. Confidential information must also be disclosed to federal and state fraud and abuse agencies and to state licensure or certification agencies upon those agencies' requests. Other disclosure of information with individual identifiers is allowed when such disclosure is required by judicial or administrative proceedings, when it is needed by the GIG and the General Accounting Office (GAO) in pursuance of their statutory re- sponsibilities, or when it is necessary (because of a substantial risk to pub- lic health) to disclose the information to state or local public health agen- cies. Redisclosure by those agencies is permitted if information identifying patients and physicians is deleted. Persons who violate the disclosure pro- visions are subject to fines up to $1,000 or imprisonment for up to 6 months. or both, and must pay prosecution costs. Among the more confusing aspects of data disclosure has been whether PROs may (or must) disclose to a hospital (or hospitals) the fact that they have confirmed a quality problem or are otherwise investigating or tracking the performance of a physician with privileges at the specific hospital (or at several hospitals). The early disclosure rules did not give unambiguous guidance on this point, and PRO s apparently received guidance that varied among the HCFA ROs. Comments during study site visits also reflected complaints from hospitals that PROs never notified them of problem physi c~=s. In June 1989 (after study site visits were completed), HCFA clarified the rule, stating that information about a physician-specific quality concern, which can consist of as few as one confirmed quality problem, must be dis

UTILIZATION QUME CONTROLPRO PROGRAM 379 closed to the hospital if the hospital so requests and may be disclosed if the PRO so chooses even without a request (HCFA memorandum dated June 1, 1989~. The PROs cannot, however, divulge to hospitals potential quality problems or the corrective actions being considered or taken. This clarifi- cation goes some way to meet the recommendations made in June 1989 by the Administrative Conference of the United States (ACUS, a federal advi- sory panel), namely that PROs should be required to send involved hospi- tals information about "confirmed violations of quality of care standards" by physicians and about the related corrective action plans (Vibbert, 1989c). It falls short by leaving the responsibility for requesting such information with hospitals (which can presumably register a blanket request with the PRO or otherwise develop cooperative relationships with the PRO so that the PRO will provide information even without a request) and by not in- cluding the corrective action information. Finally, as noted above, PROs are expected to obtain information on disciplined physiciansfrom various state authorities, but their responsibility to provide information on physicians who are involved in quality interven- tions (corrective action plans) or in sanction proceedings to such state agen- cies is not clear. Nor are PROs at present required to submit information about physician sanction recommendations to the National Practitioner Data Bank (which is being established through the Health Care Quality Improve- ment Act of 1986~. Thus, another set of reforms suggested by ACUS specifies that sanction recommendations be forwarded to state medical boards and licensing authorities and possibly to the data bank as well (Vibbert, 1989c). Disclosure of aggregate data. PROs must provide to agencies respon- sible for health planning certain summary utilization and demographic sta- tistics. They may not include anything to identify patients or physicians. Various regulations provide for disclosure of nonconfidential aggregate sta- tistical information, including PRO interpretations and generalizations about the quality of care. They specifically mandate PRO disclosure, upon de- mand, of information on quality and the appropriateness of health care services in individual hospitals. PRO s may release information on average length of stay, death rates for individual departments and illnesses, the number and type of surgical procedures performed, and the number of pa- tients in each category who required readmission. Disclosure of hospital-specific information. Information that identifies hospitals is not considered confidential, and disclosure of information that identifies particular institutions is permitted. Provider institutions may dis- close information about themselves as long as identifiers of patients and physicians are excluded.

380 KATHLEEN N. LOHR AND ALLISON J. WALKER To protect the rights and interests of hospitals, regulations require a PRO to notify a hospital when it intends to disclose information about that insti- tution (other than routine reports sent to HCFA or FIs, information to or from PRO subcontractors, or information to or from the hospital itself). The PRO must notify the institution at least 30 days before release, supply the institution with a copy of the information to be released, and give the institution an opportunity to submit comments. If the disclosure concerns an investigation of fraud or abuse and the information relates to a poten- tially prosecutable offense, the PRO need not notify the institution before releasing the information. Investigative agencies other than the OIG or GAO must stipulate in writing to the PRO that the information they are requesting does relate to possible criminal prosecution. Freedom of Information Act. PRO s are granted by statute a flat exemp- tion from requirements of the Freedom of Information Act (FOIA) and are not subject to FOIA disclosure requirements, but the situation is complex. As implied earlier, PROs must disclose all information to DHHS that the department requests. Because all agencies of the federal government are covered by the FOIA disclosure provisions, seemingly the information sub- mitted by PROs to DHHS would be subject to disclosure under FOIA. However, the preamble to the relevant regulations notes that the informa- tion routinely provided to DHHS does not identify individual patients or practitioners, and DHHS regulations protect individual privacy. Also, the regulations provide that some information must be disclosed only on-site at the PRO. Furthermore, FOIA protects personal privacy by providing that information contained in personnel and medical files, the disclosure of which would constitute a clearly unwarranted invasion of personal privacy, is ex- empt from FOIA. In short, the protections favor nondisclosure of PRO . ~ . Information. COSTS The annual PRO program budgets have risen in absolute terms in the latter half of the 1980s, although in earlier years they did not keep pace with those of the PSRO program. According to HCFA, the PRO program is currently budgeted (excluding internal expenditures of HCFA) at approxi- mately $300 million a year, up from $157 million a year for the first round of PRO contracts (FY 1984-1986) and $217 million a year for the second round of PRO contracts (FY 198~1988~. In FY 1987, Part A Medicare benefits amounted to $50.8 billion and Part B outlays to $30.8 billion (for a total of $81.6 billion) (Committee on Ways and Means, 1989~. For FY 1987 PRO outlays, the Congressional Research Service (CRS) gives $187.5 million for all PRO activities; GAO cites $155

UTILIZATION AND QUALITY CONTROLPRO PROGRAM 381 million for inpatient review (GAO, 1988a). The first figure amounts to about 0.2 percent of all outlays, and the second approximates 0.3 percent of Part A expenditures. For FY 1989, Parts A and B expenditures are estimated at about $98.5 billions; the PRO budget was put at $300 million, about 0.3 percent of outlays. An estimate for FY 1990 puts expenditures at better than $112 billion and the PRO budget presumably remains at $300 million; equivalent estimates for FY 1991 are approximately $130 billion and $330 million, respectively (H. Brook, personal communication, 1989~. Again, the PRO budgets amount to about 0.3 percent of Medicare outlays. These appear to be large numbers for the PROs, and in some sense they are. It is instructive, however, to place them in the context of the expendi- tures on services, the quality of which the PROs are expected to ensure. Total Medicare expenditures are, of course, an imperfect denominator for gauging the adequacy of the Medicare quality assurance investment, but the ratio of PSRO and PRO dollars to total expenditures over time does indicate whether that investment is rising, falling, or remaining constant in the face of changing demands on the peer review program. As a proportion of Medicare expenditures, these figures are lower for the PRO program than for the PSRO program, although the emphasis has shifted over the years a little more toward quality assurance and away from utilization and cost control. Still, even if the $11 million intended for pilot projects were added to the estimates given above for the PRO program, its expenditures would not exceed those of the PSRO program as a percentage of expected Medi- care outlays. Given the expanded responsibilities of the PROs compared with the PSROs, the markedly changing environment of health care for the elderly, and the greater perception of threats to high-quality care in the future, some view this level of funding as parsimonious. Furthermore, even if the $300 mil- lion per year were adequate for all the varied activities presently required of PROs, the need for future congressional or executive branch assignments to be adequately budgeted should be clear. Because PRO budgets are tightly tied to the number of expected reviews which is driven by the numbers of Medicare enrollees in a given state, individual budgets for PROs range very widely. For instance, of the PROs awarded full 3-year contracts for the third SOW, the California PRO was awarded nearly $82,83S,000, the largest award in the county and a record for the PRO program (Vibbert, 1989a); the PRO for Wyoming was awarded $1,210,000. Of PROs awarded extensions of existing contracts for 6 or 12 months, the largest award was to Texas (just over $16.1 million) and the smallest award was to American Samoa and Guam ($24,120) (unpublished HCFA data, April 11, 1989~. PRO budgets are based on negotiated costs for "simple," "complex," and

382 KATHLE~ N. BOHR ID BISON J. WISER ambulatory reviews; for fixed administrative costs and some start-up costs (largely accounting system updates); for photocopy and postage pass-through costs; and for costs of Civilian Health and Medicare Programs of the Uni- formed Services (CHAMPUS) review, for those PROs doing such review. Costs per review average $17.03 for simple review (range, roughly $13 t $32), $33.29 for complex review (range, nearly $27 to over $48), and $9.16 for ambulatory review (range, $4 to almost $151. QUALITY REVIEW IN MEDICARE RISK-CONTRACT PLANS17 As of April 1989, 1 million Medicare beneficiaries were enrolled in 133 "risk contracts" held by HMOs and CMPs. That enrollment figure accounts for about 3 percent of the Medicare population (O'Kane, 1989~. The Medi- care program has been experimenting for a decade with the concept of enrolling Medicare beneficiaries into HMOs. In 1981, prepaid health plans delivered care, on a cost-reimbursement basis, to nearly 600,000 beneficiar- ies. The Social Security Amendments of 1982 sought, through Section 1876, to expand participation of HMOs through risk-contracting. Also in 1982, TEFRA established the strategy of paying HMOs 95 percent of the average adjusted per capita cost (AAPCC) of the Medicare program, but implementation was delayed until 1985 until a means for computing the AAPCC had been devised. By 1989, it seemed clear that the risk-contract program had not been successful for many HMOs, owing to the extent of financial losses (except in areas where the AAPCC is extremely high), possible adverse selection because of the generosity of some benefit pack- ages, and high use of Medicare beneficiaries who had previously been un- derusing services because of financial reasons. Some HMOs withdrew from the program; others stopped marketing to the Medicare population. The history of quality review for the care rendered to such beneficiaries by HMOs and CMPs is both complex and historic. It is historic chiefly because it ushered in (1) ambulatory care review (i.e., review of care pro- vided in private physicians' offices), (2) an attempt to review "episodes" of care, and (3) an effort to design a way to reduce required review for provid- ers having an adequate quality assurance plan of their own. Because of the importance of these developments, the history and provisions of the HMO- CMP portion of the PRO program are described in some detail. History Section 1876 of the Social Security Amendments of 1982 was the first attempt to entice prepaid group practices into risk-contracting with Medi- care, but the initial design of the program, particularly its payment pro- visions' was not attractive to the HMO community. To overcome these

UTILIZATION AND QUAL17Y CONTROL PRO PROGRAM 383 drawbacks, TEFRA amended Section 1876 to create a new payment strat- egy based on the AAPCC. Implementation was considerably delayed while methods for computing the AAPCC were worked out. TEFRA did not spe- cifically mandate PRO (or indeed any) review of quality of care per se, although it did call for monitoring the quality assurance programs of risk- contract HMOs and CMPs. (PROs were not required to review or monitor plans with cost-based contracts.) During this time HCFA was attempting to design a strategy for review- ing capitated systems. The agency's "white paper" focused on the financial incentives of capitation (underuse, untimely care, and inappropriate set- tings), inpatient and outpatient care, consumer access, and medical review. The final TEFRA regulations, which became effective February 1985, in- cluded provisions requiring risk-contract HMOs and CMPs to comply with requirements for PRO review. In short, before COBRA, no specific legislative requirements existed for the review of services provided to Medicare beneficiaries enrolled in risk- contract HMO s and CMPs, but review was nonetheless contemplated by HCFA. The HMO industry resisted the TEFRA provisions, arguing that review was unnecessary because all plans were required to have an internal quality assurance program in order to qualify for a risk contract. Addition- ally, the industry argued that if quality review was to be external, then using existing PROs would be inappropriate because they are staffed pri- marily by fee-for-service (FFS) physicians (i.e., not by "peers"~. Nevertheless, given the regulations promulgated after TEFRA, some common understanding needed to be reached. HCFA thus convened a meet- ing that led to the designation of a PRO-HMO Ad hoc Committee that would draft an industry proposal for HMO review by PROs.~8 It was envi- sioned that this collaborative effort would guide the design of HMO-CMP review for Medicare. The industry proposal had the following main elements. First, a new national quality review organization (similar to the existing National Com- mittee for Quality Assurance [NCQA], an HMO industry body) would be established. It would have the following functions: (l) establish screening criteria for the quality review process; (2) establish qualifications of physi- cian reviewers; (3) define review, analysis, and reporting activities for each PRO's (to be established) HMO committee; (4) train HMO and CMP physi- cian reviewers; and (5) render technical advice to PROs and to HMOs and CMPs on proper interpretation and application of review criteria. Further, each PRO would create a committee of HMO and CMP physicians. This committee would be responsible for quality review and would (1) perform reviews, (2) analyze data to identify areas of concern, (3) validate areas of concern through chart review, and (4) recommend corrective actions. The PRO-HMO Ad Hoc Committees proposed scope of PRO review

384 KATHLEEN N. LOHR AND AI1ISON J. WALKER was quite broad. It included (1) examination of the validity and effective- ness of internal quality assurance programs at HMOs and CMPs (using NCQA standards); (2) monitoring plans, with review based on screening criteria developed by the national committee and focused review of poten- tial problems; (3) medical record review of areas of concern; and (4) recom- mendations for corrective action to be made to the HMO and CMP. Both institutional and ambulatory settings would be covered. The committee proposed to develop an approach to ambulatory review based on 15 "senti- nel conditions" as screens for possible inadequate ambulatory care.~9 In the midst of this activity, the executive Office of Management and Budget (OMB) overruled HMO review, citing the lack of legislative author- ity. In the aftermath, the work of the PRO-HMO Ad Hoc Committee was effectively lost. Because of continuing concern for underutilization in risk-contract pro- grams, OBRA 1986 solved the legislative-authority problem by mandating review of care rendered in HMOs and CMPs effective for services provided after January 1, 1987. "Comparable review"20 was to focus on quality, especially appropriate treatment and setting, access and timeliness of ser- vices, and the potential for underutilization of services. The review process was heavily oriented toward medical record review and required a consider- able volume of review. The basic approach was modified, at the instigation of the executive OMB, to permit a level of reduced review (which came to be known as "limited review") for those HMOs and CMPs that requested such status and could demonstrate they had properly functioning internal quality assurance programs. Rather than use existing models developed by NCQA, the Joint Commis- sion on Accreditation of Healthcare Organizations (the Joint Commission), and similar groups, HCFA defined de nova a set of "areas of focus" by which HMO and CMP internal programs would be evaluated. Characteris- tics of concern to HCFA were whether the plan (1) reviews individual cases of patient care, (2) reviews all settings, (3) includes physician review of medical records, and (4) uses reasonable sampling methods to select cases for review; whether the plan has been operating long enough for it to be able to demonstrate actual performance; and whether physicians make final Decisions on quality Issues. The industry widely regarded these concerns as rather old-fashioned and lacking in an understanding of what HMO and CMP quality assurance plans actually do. One other factor complicated the implementation of this review. OBRA 1986 allowed review of HMO and CMP services by entities other than PROs in the area as a means of stimulating competition among review organizations. These other entities are referred to as Quality Review Or- ganizations (QROs). QRO contracts were limited to no more than half the states, covering half the Medicare population. In March 1987, a Request

UTILIZATION AND QUALITY CONTROLPRO PROGRAM 385 for Proposals was issued for QRO review of HMO and CMP services in 25 states. Quality Quest (a subsidiary of InterStudy, located in Minneapolis, Minnesota) was awarded the review contacts for the states of Illinois, Kansas, and Missouri; California Medical Review was awarded the review for Ari- zona and Hawaii.22 All remaining HMO and CMP reviews is done by PROs in the state itself.23 Hereafter, PRO is understood to include the QRO. Three Types of HMO and CMP Review HMO and CMP reviews has three levels of review: limited, basic, and intensified (outlined in Table 8.6~. Basic review is the core approach to HMO review. Limited review is intended to reduce the volume of active TABLE 8.6 Summary of Activities for Health Maintenance Organization (HMO) and Competitive Medical Plan (CMP) Review, by Requirements for Limited, Basic, and Intensified Review Sample Size for Review Type of Review Limited Basic Intensified Thirteen sentinel conditionsa 50% - RSb 50% - RS 100% (of only 4 conditions) Hospital admissions3 % - RS3 % - RS 6 % - RS TransfersC100 %100 % 100 To Readmissions within less than 31 days25 % - RS50% - RS 100% Nontraurna deaths in5 % - RS10% - RS 100% all care settings aThese conditions, which are defined by ICD-9-CM codes, include: diabetic complications (ketoacidosis, hyperosmolar coma, other coma, and hypoglycemic coma); acute appendicitis with generalized peritonitis or peritoneal abscess; hy- pertensive problems (several categories, including occlusion and stenosis of precere- bral arteries and transient cerebral ischemia); gastrointestinal catastrophes (acute, chronic, or unspecified gastric ulcer with hemorrhage without obstruction; chronic duodenal ulcer with hemorrhage without obstruction; unspecified intestinal obstruc- tion); gangrene of the extremity; operations for breast malignancy (including certain biopsies and unilateral radical mastectomy); malignant neoplasm of the genitouri- nary organ; adverse drug reactions (several categories, mainly poisoning by specific pharmacologic agents); other cellulitis and abscess; malignant neoplasm of colon; hypokalemia; septicemia; and pulmonary embolus. bRS is random sample. CTransfer category eliminated August 1989. SOURCE: HCFA, 1988.

386 KATHLEEN N. I~HR AND ALLISON J. WALKER PRO review for the plan, mainly by requiring smaller sample sizes than are prescribed for basic review. Intensive review has the same general meaning as in FFS settings (i.e., it is invoked when a threshold for a quality problem is reached), and sample sizes are larger than for the other two levels of review (usually in 100 percent of relevant cases). The three levels are not a true continuum, however, because for plans on limited review, quality prob- lems that reach specified thresholds trigger intensified, not basic review. Limited Review Limited review is available only to those HMOs and CMPs that request it and pass an assessment of their internal quality assurance program to determine that it has the capacity to identify and correct quality problems. It has two main components. First, if the PRO judges the risk-contract plan's quality assurance pro- grarn adequate according to the "areas of focus" mentioned above, then it reviews a subsample of cases already reviewed by the plan to validate the plan's judgments. This is done when the plan is first assessed, and quar- terly thereafter. The purpose is to monitor the plan's internal program, not to provide a generalized statement about the quality of care provided. One of four "outcomes" for the subsample review effort is possible: 1. According to the PRO, care was deficient (in quality, appropriate- ness, or access) and the plan had not made this determination. 2. Care was deficient and the plan had already so determined. If a pat- tern is identified, the PRO would then monitor the plan's corrective action activities. 3. Care was not deficient and the plan had so decided. 4. Care was not deficient and the plan had previously found that it was. If a pattern of problems (relating to Outcomes 1 and 4 above) becomes apparent, the PRO then monitors the plan's corrective actions and may assign the HMO to intensified review. Second, medical record review is required in four main areas25: (1) a 50- percent random sample of four conditions selected from among 13 specified "sentinel" conditions,26 (2) a 3-percent random sample of inpatient admis- sions, (3) a 5-percent random sample of nontraumatic deaths in all health care settings, and (4) a 25-percent random sample of readmissions within 31 days. For the 13 sentinel conditions, both pre- and post-hospitalization ambulatory care is reviewed through the use of criteria developed by HCFA and the PRO. A fifth area is focused review of ambulatory care, for which PROs were given 6 months to develop a methodology. In addition, benefi- ciary complaints are also reviewed, and PROs must perform community

UTILIZATION AND QUAL17Y CONTROL PRO PROGRAM 387 outreach activities for risk-contract enrollees similar to those for FFS bene ~ . . . octanes. For plans on limited review, the total number of cases selected for the random validation subsample plus the total number selected for the remain ing reasons for review cannot exceed the number that would have been reviewed under the basic plan. Basic Review Plans not opting for, or not eligible for, limited review are placed on basic review. It includes the same five areas covered by the second compo nent of limited review but uses larger samples, and it requires the benefici ary complaints and outreach activities. Not included in basic review is the first component of limited review; that is, the extra quarterly review of charts. Intensified Review If a plan comes under intensified review, the sample of cases reviewed is again larger: 6 percent of all admissions, 100 percent of all cases among any of the 13 conditions, 100 percent of nontrauma deaths, and 100 percent of all readmissions. Plans on limited review move to intensified review in one of two instances: (1) if 5 percent or 6 cases selected for the subsample validation review in a quarter have Outcome 1 (in which the PRO found a problem but the plan did not); and (2) if 5 percent or 6 cases of all other cases reviewed have problems related to standards of quality, appropriate ness of care, or access. Basic plans are put under intensified review only in the second instance. Intensified review continues for 6 months before the plan's status is reviewed. The Review Process By and large, the process for reviewing care rendered to Medicare bene- ficiaries in risk-contract HMOs and CMPs is similar to the process for reviewing care in traditional FFS settings (e.g., use of generic screens, assignment of severity levels, physician or plan notification, and the like). That process has already been described. The main difference is that HMO-CMP review has attempted to imple- ment episode-of-care review based on two types of cases, simple and com- piex. A "simple case" is one in which services being reviewed were pro- vided in only one setting and during only one admission (if inpatient) for instance, those in the 3-percent inpatient sample. A "complex case" is one

388 i] KATHLEEN N. LOHR AND AlllSON J. WALKER n which services being reviewed were provided in more than one setting or involve more than one hospital stay; for example, cases selected under the 13 sentinel conditions would normally be classified as complex. For com- plex cases, PROs are expected to review the care rendered in all relevant settings (ambulatory, hospital, and post-acute). Arguably the most significant step is the requirement for ambulatory review, which leaves to each PRO the responsibility of developing a fo- cused review methodology and establishing clinical screening criteria to be used in reviewing the 13 sentinel conditions. Because the HMO industry, the PRO community, and HCFA believed that the possibility of dozens of different approaches to ambulatory review was not an attractive proposi- tion, these groups agreed that an industry-PRO task force would develop model methods to offer the PRO community. Because of the lack of uni- form data among PROs, the task force recommended a random sample of HMO Medicare enrollees for this "focused ambulatory review" activity. As of mid-1989, experience with focused ambulatory review was limited, but the process of collaboratively developing acceptable tools for such an effort was considered valuable (O'Kane, 1989~. CURRENT AND FUTURE INITIATIVES OF THE PRO PROGRAM Uniform Clinical Data Set HCFA began in 1987 a complex project to develop a data set and accom- panying decision rules for use by PROs and the wider research community that would contain far more detailed clinical data than was heretofore avail- able in the HCFA data files. Known as the Uniform Clinical Data Set (UCDS), this project is one of a number of steps intended to expand and improve the ability of the agency to assure the quality of care delivered to Medicare beneficiaries, using the PROs as the principal mechanism. The genesis of the UCDS was in the recognition that the way PROs make judgments about He necessity, appropriateness, and timeliness of care var- ies and is too subjective. One objective of the UCDS, therefore, was to put in place a mechanism to make PRO review more systematic. The basic principle is to screen cases by applying a uniform set of computerized decision rules that are based on more complete clinical data. Such a com- mon set of guidelines would also permit a more effective assessment of the work and performance of PROB. The second purpose of the UCDS is to permit the development of more and better information about the practice of medicine, so that PROs, among others, will be able to base decisions about quality, appropriateness, and medical necessity of care on systematically and objectively evaluated col- lective data rather than on individual experience. The availability of the

UTILIZATION AND QUALITY CONTROL PRO PROGRAM 389 extensive clinical information collected in the UCDS formats would support more thorough and detailed analysis of patterns of interventions and out- comes than is possible simply with billing data. Thus, for patients with particular medical problems, a large body of information about alternative interventions could be made available to PROs and the medical community. The agency plans to make the UCDS data available for intramural and extramural analysis. The basic operating premise of the UCDS is that relevant clinical data will be abstracted from medical records of all inpatient admissions reviewed by the PROs for whatever reason. (This amounts to about 20 to 25 percent of all Medicare admissions in a year, or approximately 2.0 to 2.5 million admissions; of these, about 3 percent are a truly random sample of admis- sions, and the remainder are cases mandated for review for specific reasons, largely related to the probability of problems related to PPS fiscal incen- tives or to quality of care.) Quality-of-care algorithms have been developed to screen cases for potential quality problems automatically. Nurse review- ers who flag instances of potential quality deficiencies for more in-depth review will have more organized, objective, clinical information before them, and physician reviewers likewise will have better organized information on which to base their decisions. More broadly, HCFA hopes to be able to set national and individual PRO goals for improving quality of care and to measure the success of PROs in reaching those goals (Morford, 1989a, 1989b). The total number of data elements available on the UCDS is about 1,600, although not every data element is needed or relevant for every case. The contents of the UCDS fall into 10 major categories (for details, see Table S.7~: I. Patient Identifying Information II. Patient History and Physical Examination and History and Physical Exam Findings III. Laboratory Findings IV. Imaging Findings and Other Diagnostic Test Findings V. Endoscopic Procedures VI. Operative Episodes VII. Treatment Interventions VIII. Medication Therapy in Hospital IX. Recovery Phase X. Patient Discharge Status and Discharge Planning Medical records will be abstracted by PRO abstracters either on-site or at a central office; data will be entered via desktop or laptop computers. At present, data abstraction requires about 1 hour per case, but that time re- quirement is expected to decline as software is improved and experience is

390 KATHLEE7'I N. LOHR AND ALLISON J. WAVER TABLE 8.7 Elements of the Uniform Clinical Data Set I. Patient Identifying Information Patient's Medical Record Number Health Insurance Claim Number Provider Number Physician Number Date of Admission Date of Discharge Patient's Date of Birth Patient's Sex Source of Admission (location of patient just before admission to inpatient bed) Readmission Code Ambulatory Care Care-giver on Admission Patient's Race Patient's Occupational Status Patient's Insurance Coverage II. Patient History and Physical Exam Part 1. (Source of data: physician and nursing notes) Activities of Daily Living Prior to Admission (continence; mobility) Patient Weight Patient Height Vital Signs (temperature, pulse, respiratory rate, blood pressure) Current Medications History of Drug Reaction History of Drug and Substance Use Medications Administered in Emergency Department History of Total or Partial Excision of Major Organ History of Amputation of Major Limb History of Replaced Body Structure (e.g., cardiac valve; hip) History of Organ Transplant History of Congenital Absence of Organs Part 2. (Source of data: all available information concerning findings documented within the first 24 to 48 hours following admission or up to 6 weeks before admis- sion, with emphasis on whether He patient is "under current management or moni- toring") History of chronic neurological diseasets) History of neurological surgery Current neurological examination findings (e.g., Glasgow Coma Score; cerebrovas cular accident; speech deficit; syncope) History of chronic cardiac diseases) (e.g., chronic hypotension) History of chronic vascular disease History of cardiovascular surgery (coronary artery surgery or bypass graft; angioplasty; intracardiac surgery; vascular surgery; peripheral vascular sur gery)

UTILIZATION AND QUAL17Y CONTROLPRO PROGRAM 391 Current cardiovascular exam findings (22 findings such as shock, pulmonary edema, peripheral edema, arrythmia, stasis ulcers, or steady chest pain) History of chronic pulmonary diseases) History of pulmonary surgery Current pulmonary findings (e.g., tachypnea; cyanosis, cough, strider) History of chronic immunological disease (e.g., HIV positive; other autoimmune disease; systemic infection) History of endocrine diseases (e.g., diabetes mellitus; hyperthyroidism~a History of endocrine system surgery (e.g., thyroidectomy) History of cancers History of chronic gastrointestinal disease (upper GI disease, lower GI disease, hepatobiliary disease, pancreatic disease, GI bleeding, colicky pain, weight loss ~ History of gastrointestinal surgery Current gastrointestinal findings (e.g., ascites; GI bleeding; abdominal disten- sion, rigidity; rectal blood, mass; recent persistent nausea, vomiting, etc.) Current cutaneous findings (e.g., skin ulcer; burn, cellulitis) History of musculoskeletal surgery (spine, hips/knees, and long bone fracture considered separately) Current musculoskeletal findings (e.g., congenital deformity; fracture; soft tissue trauma) History of chronic urologic conditionsa History of urologic surgery (prostate, bladder, ureter, kidney) Current urologic exam findings (e.g., enlarged prostate; flank or genitalia tender ness) History of chronic psychiatric disorders Current psychiatric exam findings (e.g., suicidal; major affective disorder; de- mentia or mental retardation) History of chronic ob/gyn disorders History of gynecologic surgery (in relation to pregnancy and to other gyneco- logic conditions) Current ob/gyn findings (pregnant; not pregnant) Current neonatal exam findings (e.g., gestational age; Apgar) III. Laboratory Findings (Listed laboratory values at three points in time: first (or worst) within 24 hours of admission (or pre-admission up to six weeks before admission), worst interim value, and last test) Chemistry/hematology (26 tests such as alkaline phosphatase, bilirubin, creatinine, glucose, hematocrit, platelets, potassium, and white blood cell count; plus 7 additional tests within 48 hours, including CPK, SOOT, and thyroid tests) Urinalysis: microbiology (e.g., protein, red cells, bacteria) Microbiology - cultures (e.g., blood, cervix, cerebrospinal fluid, sputum, stool, urine) Cytology/histology (e.g., bronchial wash, aspirates; joint fluid) TABLE 8.7 continues

392 TABLE 8.7 Continued KATHLEEN N. LOHR AND AI1ISON J. WALKER IV. Imaging and Other Diagnostic Test Findings (Dates of examination and specific findings are recorded; pre-admission tests up to six weeks before admission are acceptable.) Chest x-ray (with or without contrast) Upper GI, barium enema/swallow, gallbladder KUB/abdominal x-ray IVP/urogram Ultrasound Extremity/skeletal/spinal x-ray or myelogram CT scan (head and neck; spine; chest; body) Magnetic Resonance Imaging (head and neck; spine; chest; body) Nuclear medicine/isotope studies (thyroid, lung, bone, hepatobiliary, cardiac) Electrocardiogram Cardiac catheterization, ventriculogram Echocardiogram . . . Arterlogram, anglogram Electroencephalogram Pulmonary function test Gastrointestinal motility (manometric) V. Endoscopic Procedures (Dates of examination and specific findings are recorded; pre-admission tests up to six weeks before admission are acceptable.) Procedures: arthroscopy, cystoscopy/cystogram; upper GI endoscopy, lower GI endoscopy, endoscopic retrograde cholangiopancreatography (ERCP), laparoscopy, hysteroscopy, bronchoscopy and laryngoscopy. Also recorded: anesthesia type; anesthesia risk grade; and unexpected intra-endo- scopic events (e.g., apnea, hemorrhage, cardiac arrest, perforation) VI. Operative Episodes All procedures recorded (using ICD-9-CM codes) Also recorded: anesthetic type anesthesia risk (American Society of Anesthesiologists classification system) vascular access line surgical wound classification (clean, clean-contaminated, con- taminated, dirty) . .. . tissue mc lugs occurrence of unexpected intra-operative events (e.g., respiratory arrest, acute myo cardial infarction, blood loss, stroke) return to operating room VII. Treatment Interventions (This section covers various data items relating to noninvasive procedures.) Blood products

UTILIZATION AND QUALITY CONTROL PRO PROGRAM 393 Inhalation therapy Professional services (e.g., physiotherapy, respiratory therapy, social work, psychi- atric counseling, pastoral care) VIII. Medication Therapy in Hospital Current medication at admission (See Section II, Part 1) Medications administered in the hospital (this section covers name of medication, route of administration, and dates medication was initiated, stopped, or route changed; dosage is not recorded.) Medication at discharge (name) Adverse reactions to medications during hospital stay (e.g., abnormal drug reaction; expected toxic side effect) Delivery systems for medications (i.e., other than oral) IX. Recovery Phase Special care unit days (e.g., coronary care, general intensive care, neurosurgical) Do not resuscitate order (date) Unexpected inpatient events (a lengthy list of major or catastrophic complications with possible outcome of death or chronic disability, such as myocardial infarc- tion, pulmonary embolism, acute bowel obstuction, septicemia shock, respiratory failure, wound infection, urinary retention, pneumonia) Trauma suffered in hospital (e.g., fall or accidental injury with significant untoward effect; decubitus ulcer, attempted suicide) Nosocomial infections Prolonged stay (e.g., awaiting placement into chronic care facility; inability or re- fusal of family to care for patient) X. Patient Discharge Status Discharge vital signs Discharge physical exam findings (cardiovascular, neurological, pulmonary, gastro- intestinal, cutaneous, urologic, last chest x-ray, last abdominal x-ray) Last EKG Discharge Activities of Daily Living (urinary incontinence; walking; alimentation, elimination) Other discharge therapies (medications, IV therapy other than alimentation, oxy- genation, monitoring, mechanical implants, dialysis) Discharge plan (care-giver; follow-up) Discharge disposition (e.g., home, transfer, home health service, died) Full and final listing of diagnosis (first 15 final diagnoses in ICD-9-CM codes) aResource manual refers abstracters to separate lists that are embedded in the software and called up on the computer screen for reference. SOURCE: "Resource Manual for Uniform Clinical Data Set (U.C.D.S.)," pre pared by Case Mix Research, Queen's University, Department of Community Health and Epidemiology, Kingston, Ontario, Canada, in association with the Wisconsin Peer Review Organization (WIPRO), Madison, Wisconsin, 1988.

394 KATHLEEN N. LOHR AND ALLISON J. WAITER gained with this approach. The detailed guidelines that describe precisely the data to be acquired (for an example, see Table 8.8) have been developed by Case Mix Research, Queen's University, Ontario, Canada, in association with the Wisconsin Peer Review Organization, Madison, Wisconsin. The computer algorithms fall into several categories (for details, see Table 8.9~: surgery (12 specific procedures), specific diseases (12 condi- tions), organ systems (10 systems), generic quality screens (6 classes of problems), and discharge screens. They have been developed with the help of a task force of clinicians; initial programming was done in part through KePRO, the Pennsylvania PRO. As of April 1989, the UCDS project was in a small pilot-test phase. Field testing of the whole approach, including use of algorithms to assist in the selection of cases for physician review, is expected to begin during the winter of 1989-1990. An assessment and recommendation about whether to go forward with this approach as an integral part of the PRO quality review task is expected late in 1990. Pilot Projects for PROs Apart from the UCDS activity, HCFA and the PRO community are em- barking on a series of pilot projects designed to begin the implementation of several new review activities required by legislation (e.g., OBRA 1987~. These are all projects that should form a bridge to a more comprehensive quality assurance program in the future. These efforts fall into two catego- ries: (1) reduced hospital review (sometimes referred to as alternative hos- pital review) and (2) noninstitutional setting review, specifically outpatient (physician office-based) review and post-acute (HHA and SNF) review. Approximately $9 to 11 million in Medicare Trust Fund monies will be set aside over 3 years to pilot new review methods. Only PROs will be eligible for funding through contract modifications, although they can and will subcontract with each other and with outside research and academic groups for needed technical assistance. Two formal requests for contract modification proposals (for noninstitutional and alternative hospital review methods) were released in May and July 1989. Noninstitutional Review A pilot project on noninstitutional review began on December 1, 1989, in response to the request issued in mid-May 1989. The emphasis is on ambu- latory (office-based) care. The Wisconsin PRO serves as the lead PRO; six other PROs are included in the effort. The project will evaluate approaches to monitoring and assessing care in the office setting, taking four factors into account: (1) severity of illness, sequence and patterns of care, variety

UTIlIZATION AND QUALITY CONTROL PRO PROGRAM TABLE 8.8 Example Of Data Element Recorded for Uniform Clinical Data Set: Cardiac Cathetenzation and Ventriculogram 395 Formal report of the first cardiac catheterization and/or ventriculogram performed during the admission or up to 6 weeks before admission; if more than one test, use the one closest to admission. Catheterization takes precedence over ventriculogram. Any procedure done in an operating room, minor treatment room, at the bedside, or in the radiology suite can be included. All approriate categories of specified find- ings are checked. The general rules for recoding information for the UCDS are to change default values on the computer screen ("F") to '`T." The specific findings to be recorded for catheterization or ventriculogram are the following. The reviewer changes F to T unless a percentage is called for, in which case the worst percentage is recorded, or other information is specified. Normal AV shunt Ventricular/atrial septal defect Valvular defects: Aortic stenosis (<1 sq cm) Aortic regurgitation (moderate or severe) Mitral stenosis (<1 sq cm) Stenosis: left main (%) Stenosis: left anterior descending (%) Stenosis: circumflex (%) Stenosis: right (%) Cardiac output (liters/minute) (%) LV ejection fraction (%) Abnormal chamber size/wall motion Ventricular aneurysm Congerutal anomalies (patent ductus, ventricular septal defect) Aortic aneurysm Dissecting aortic aneurysm Other abnormal findings Coronary artery grafts number Number with >70~o stenosis Pressures Left ventricular systolic; diastolic Aortic-systolic; diastolic; mean Pulmonary artery (including Swan-Ganz) systolic; diastolic SOURCE: "Resource Manual for Uniform Clinical Data Set (U.C.D.S.)," pre- pared by Case Mix Research, Queen's University, Department of Community Health and Epidemiology, Kingston, Ontario, Canada, in association with the Wisconsin Peer Review Organization (WIPRO), Madison, Wisconsin, 1988.

396 KATHLEEN N. LOHR AND ALLISON J. WALKER TABLE 8.9 Selected Categories of Quality-of-Care Algorithms for the Uniform Clinical Data Set I. Surgery Algorithms Prostatectomy Major joint replacement Gastrointestinal surgery Cardiac revascularization Cholecystectomy Hernia repair Vascular bypass and endarterectomy Hysterectomy Mastectomy Cataract surgery L aminec to my Thoracotomy II. Disease-Specific Algorithms Ischemic myocardial disease and chest pain Cardiac pump failure Acute lower respiratory infection Ischemic cerebrovascular disease Obstructive airway disease and respiratory failure Cardiac dysrhythmia Diabetes mellitus Urinary tract infection Malignant pulmonary neoplasm Septicemia Thrombophlebitis and pulmonary embolism Malignant neoplasm of female breast III. Organ System Algorithms Cardiac system Vascular system Pulmonary system Gastrointestinal system Genitourinary system Bone, joint, and muscular system Neurological system Infectious diseases Endocrine system IV. Generic Quality Screen Algorithms Adequacy of discharge planning Medical stability of patient at discharge Deaths Nosocomial infection Unscheduled return to surgery Trauma suffered in hospital V. Discharge Screens Discharge status/disposition

UTILIZATION AND QUALITY CONTROL PRO PROGRAM 397 of treatments associated with various medical specialties, and the contin- uum of care (rather than specific encounters); (2) methods for retrieving appropriate details from office medical records; (3) methods for organizing and integrating such information with other data (e.g., the Medicare claims files); and (4) the impact of care on the patient's condition over time. The aim is to be able to evaluate the effectiveness of the care and to test the practicality and usefulness of the techniques available for making that evalu- aiion. HCFA had called for two general approaches. The first is a statisti- cal assessment of the "impact of the physician's services on the condition of his overall patient population"; the second is an evaluation of cases with unexpectedly adverse risk-adjusted outcomes. Reduced Hospital Review The first pilot projects in this category were intended to develop alterna- tive mechanisms by which hospitals can be put on "reduced review"; that is, be "focused out" of the level of review they presently experience. This effort reflected two factors: (1) the wide recognition that the volume of hospital review, especially for hospitals with good performance records, is too burdensome; and (2) the judgment that the old PSRO approach of "dele- gating" review to hospitals was not entirely successful and would probably not be practical in today's more data-intensive review world. The main purpose of this set of projects was to test alternative review methods for hospitals that are based on outcomes. Hospitals would provide information on their patients that would let HCFA track risk-adjusted out- comes. The hospitals would provide information to the PROs; with outside expert assistance in epidemiology and statistics, the PROs were then sup- posed to assess hospital performance on the basis of risk-adjusted outcomes. Although an REP was published in July 1989, no contract modifications were made. This proposed activity was questioned because it appeared to call for "delegated review" similar to that of the PSRO program (which presently is forbidden by HCFA regulations). Although hospitals can be expected to be responsive if they believe it will help reduce the review burden for hospi- tals with low rates of payment denials or quality problems-opposition from Congress, presumably owing to disappointment about PSRO perform- ance in this arena, was expected (Vibbert, 1989e). Appropriateness Guidelines One of the more dramatic developments in the health sector in the late 1980s has been the acceptance of, and indeed the call for, national practice guidelines, appropriateness indicators, and similar utilization management

398 KATHLEEN N. LOIN AND AI1ISON J. WALKER instruments (PPRC, 1988, 1989; IOM, 1989~. These tools serve various purposes, such as providing guidance for decisions about pre-procedure and pre-admission review and for understanding whether large variations in population-based rates of use of services reflect large variations in appro- priateness or quality of care.27 HCFA will consider pilot projects designed to develop appropriateness indicators and practice guidelines through the PROB. Such projects could update existing procedure-specific guidelines or indicators (such as those developed in the mid-1980s by The RAND Corporation) or develop and test similar tools for other procedures and services. As with the other projects, one PRO will be the lead PRO, will negotiate a contract modification, and will subcontract with other PROs or academic and research institutions as necessary. Other Activities Small Area Variations. Perhaps the most ambitious PRO project cur- rently under way is a small area analysis of variation in utilization and outcomes of hospital care being conducted by the American Medical Re- view Research Center (AMRRC). The project began in October 1987 and is expected to continue until June 1990. It will compare rates of use of hospital services in 1984-1986 in approximately 4,800 hospital market ar- eas. Using these data, project investigators will (1) develop and dissemi- nate information on use and outcomes of hospital care; (2) engage 12 PROs in a complex pilot education program to review, interpret, and feed back information to physicians on identified practice patterns; (3) improve the use of small area analysis methods as an operational tool for PROs; and (4) examine various intervention strategies to determine how they might best be applied in both the public and the private sectors. The 12 PROs involved in the educational component of the project, known as Medical Assessment Program (MAP) pilots, are located in the following states: Arizona, Arkansas, Colorado, Connecticut, Illinois, North Carolina, Ohio, Pennsylvania, Texas, Utah, Virginia, and Washington. The physician study group phase will include five surgical conditions (coronary artery bypass graft, cardiac catheterization, carotid endarterectomy, male repro- ductive organ operations, and small and large bowel operations) and five medical diagnoses (chronic obstructive pulmonary disease, pneumonia, bron- chitis and asthma, acute myocardial infarction, and diabetes). All PROs are to receive data, technical training in small area analysis methods, and nec- essary computer software. Uniform needs assessment. OBRA 1986 mandated the development of a "uniform needs assessment instrument" to be used to evaluate the needs of patients for post-acute care such as home health services or other health

UTI~ZATION^D QU4~CONTROLPRO PROGRAM 399 related long-term-care' services. This instrument is to be used by discharge planners, hospitals, nursing facilities, HHAs, and other providers, as well as by FIs-to make decisions about postdischarge needs and payment for ser- vices provided to Medicare beneficiaries. The instrument was to be developed in consultation with an advisory panel. The panel was chartered in May 1987 and its members appointed in March 1988. It met several times to review preliminary drafts of the instru- ment and consider recommendations for its use. An extensive effort was made to solicit review and comment on He final draft before formal publi- cation, which was expected in July 1989. (Table 8.10 indicates the major dimensions of the draft instrument.) HCFA plans to develop a users' man- ual and a standard training process and to field test the instrument for reliability, validity, and administrative feasibility. EVALUATING PRO ACTIVITIES HCFA has three mechanisms to monitor and validate PRO medical re- view activities: (1) the PRO Monitoring and Protocol Tracking System (PROMPTS-2~; (2) the SuperPRO; and (3) the monthly and quarterly data TABLE 8.10 Content of the Uniform Needs Assessment Instrument I. Sociodemographics II. Health Status Physical Health Mental Health III. Functional Status Activities of Daily Living (ADL) Instrumental Activities of Daily Living (IADL) Medical restrictions on ADLs or IADLs IV. Environmental Barriers to Post-discharge Care V. Nursing and Other Care Requirements Skilled observation Monitorir~g/superv is ionJev aluation Therapeutic needs Educational needs Therapy/service needs Durable medical equipment needs VI. Family and Community Support VII. Patient/Family Goals and Preferences VIII. Summary of Assessment of Needs for Continuing Care

400 KATHLEE7;I N. LOHR AND Al BISON J. WALKER reports that PROs must submit. Some observers also attempt to draw con- clusions about PROs' effectiveness from their record of bringing successful sanction recommendations (which are discussed elsewhere). Sanctions data, however, ignore the far greater number of corrective action plans required by PROs for less serious problems; thus, counting sanctions is a very lim- ited way to evaluate PRO activities. In an effort to provide additional and more rounded information about the PRO experience, the American Medi- cal Peer Review Association (AMPRA, the national PRO association) mounted a survey of its membership in mid-1989. PROMPTS, SuperPRO, and AMPRA activities are described in some detail below; data reporting is discussed in the next section. These ap- proaches are oriented more toward evaluation of the performance of indi- vidual PROs than of evaluation of the impact of the PRO program. No overall evaluation has been undertaken to determine what effect the pro- gram itself has had on maintaining or improving the quality of care for Medicare beneficiaries. PROMPTS-2 The PROMPTS-2 system, developed under the second SOW, provides information about how well PROs have fulfilled their contractual obliga- tions. HCFA uses such information to determine each PRO's status at the time of contract renewal (for a description and evaluation, see GIG, 1989~. Specific attention is given to timeliness and accuracy of medical review, responsiveness to beneficiary and provider inquiries, personnel requirements, report generation, and cost-effectiveness through a series of yes/no ques- tions and some rereview of cases. Thus, the focus is more on "process" aspects of performance than on "outcomes," and more on contract deliver- ables and cost-containment efforts than on quality of care. The 400 points to be assigned to four categories of PRO activity (utilization review, quality review, data, and management) are allocated evenly (100 each) across the categories. PROMPTS-2 review is required twice during a contract cycle and is completed by RO staff. Staff select a random sample of cases from various review categories (i.e., readmissions, transfers, outliers, etc.) and determine whether the PRO is correctly applying generic screens and other review procedures. The minimum number of cases to be rereviewed is generally 25 per review category unless the PRO performs a larger volume of re- views, in which case larger rereview samples will be chosen. The RO staff record the results of the rereview, indicating the number of cases for which they disagree with the initial PRO determination. PROMPTS-2 does not generate information on the types of quality prob- lems the PROs detect (or fail to detect). It does alert HCFA to possible

UTILIZATIOlV A]VD QUALITY CONTROL PRO PROGRAM 401 problems with PRO performance of medical review activities and contract compliance. The process largely duplicates the SuperPRO effort, although on a considerably smaller scale. Questions have also been raised about inconsistency across ROs, the expertise of their medical reviewers, and the validity of their decisions, as well as about the ability of data so generated to discriminate adequately among PROs (GIG, 1989~. A new PROMPTS is being developed to ensure consistency among regions (H. Brook, personal communication, 1989~. SuperPRO The SuperPRO (SysteMetrics, Inc.) reviews records previously examined by each PRO to make independent judgments about necessity, appropriate- ness, and quality of care. HCFA then compares SuperPRO findings with results of the PROs to determine whether either the PRO program or indi- vidual PRO performance need improvement or modification. More specifi- cally, the basic objectives of SuperPRO are as follows: 1. To validate the determination made by the PROs, specifically on ad- mission review, discharge review, and DRG validation; 2. To validate the medical review criteria being used by nonphysician reviewers for admission review; 3. To verify that nonphysicians are properly applying the PRO's criteria for referring cases to physicians for review; 4. To identify quality issues that should have been addressed by the PRO. Original Procedures During the first few years of SuperPRO, the process was essentially the following. For each review cycle (13 in a year), SuperPRO received 400 hospital medical records randomly selected by each PRO from among the cases it reviewed during that cycle. SuperPRO used the same generic screens initially applied by the PRO to judge each case. SuperPRO nurse reviewers first reviewed the case and referred those not passing the screens on to SuperPRO physician advisors; no information was sought directly from the providers whose care may have been in question. Unlike the PRO nurse reviewers, SuperPRO nurses had more leeway in referring cases for physi- cian review. Cases identified by SuperPRO as having quality problems were reported to the PRO, which could further review the cases, appeal the judg- ment of SuperPRO, and provide additional information in its rebuttal. Approximately 25 percent of PRO appeals led to reversals of decisions in

402 KATHLEEN N. LOHR AND ALLISON J. WALKER favor of the PRO (data provided by HCFA reviewer of this chapter; see also GAO, 1988a). SuperPRO reports were, until mid-l989, considered advisory and did not affect payment of claims for Medicare services. For PRO evaluation, HCFA reviewed SuperPRO findings and took necessary action to follow up identi- fied problems. Early Results HCFA believes that SysteMetrics would have identified more quality problems than the PROs did, had it (rather than the PROs) been conducting all the review nationally (Morford, 1989a). SuperPRO data indicate that approximately 10 to 15 percent of hospital admissions were unnecessary; that is, the services could have been performed in alternative settings. For instance, for Cycle 6 of SuperPRO review (dated mid-December 1988), PROs collectively had a denial rate (based on lack of medical necessity) of 2.7 percent. Depending on what cases were included or excluded in the calculations, SuperPRO had denial rates of 6.5 percent, 10.1 percent, and 13.9 percent. SuperPRO contends that approximately 5 percent of the cases it rere- views have quality problems. Of this 5 percent, approximately 3.5 percent represent potential harm to patients; 1 percent represent minor, reversible harm, and 0.5 percent represent irreversible or life-threatening harm.28 Cycle 6 data showed that the PROs had a rate of quality problems of 1.3 percent as contrasted with the SuperPRO rate of 4.8 percent (or about 3.8 times the average PRO rate). The discrepancies for individual PROs were consider- able, however. The SuperPRO problem rate was 44 times higher than that for one PRO but only 1.4 times higher than the rate for another PRO. Nevertheless, these rates of problem detection (or the prevalence of qual- ity problems they might imply) cannot be compared directly with each other because the review methodologies and data (particularly the level of infor- mation from the attending physicians or hospitals) are not the same for PROs and SuperPRO. Moreover, SuperPRO reviews do not provide infor- mation about the incidence of quality problems in the Medicare population (because it only rereviews cases already reviewed by the PRO), nor do these reviews make determinations about how cases are selected by the PRO or whether cases not reviewed by the PRO should have been. The GIG notes that "it is impossible to assess a particular PRO's per- formance based on SuperPRO data" (OIG, 1989, p. 12), citing two reasons. First, no criteria have been established for assessing the seriousness of discrepancies between PRO and SuperPRO findings. Second, no national analyses of the variations in quality problems have been undertaken. The GIG thus calls for an evaluation of the appropriateness of SuperPRO reviewers

UTILIZATION AND QUALITY CONTROL PRO PROGRAM 403 and criteria, an examination of how SuperPRO activities complement or duplicate the PROMPTS-2 process, and an assessment of how several SuperPRO procedures might be improved (OIG, 1989~. Future Plans HCFA issued a new competitive contract for SuperPRO in mid-1989 and awarded a 42-month contract for more than $5.6 million to SysteMetrics, beginning October 1, 1989. The contract calls for approximately 60,000 rereviews of PRO cases (Vibbert, 1989f). It also makes several changes to the current SuperPRO efforts (information provided at HCFA Data Confer- ence, June 14, 1989), although it is not clear these address the issues raised by the GIG. First, HCFA (rather than the PROs) will select the random sample of cases, now to number 600 per 6-month cycle (217 inpatient ad- missions, 195 HMO cases, and 188 ambulatory surgery cases). Second, if the PRO disagrees with the SuperPRO decision and sends a rebuttal, Super- PRO will do a rereview that may include "local" criteria. HCFA intends that the PROs and SuperPRO should be on a "level playing field" and that SuperPRO should use the same information that the PRO originally had in making quality judgments that is, the material reviewed by SuperPRO in making its decision is the information the PRO obtained from the hospital or physician in reaching its final decision. Nevertheless, SuperPRO will still not seek additional input from the hospital or physician whose care is under question. (This can be the point at which PRO and SuperPRO proce- dures differ most, if the additional information used by the PRO physician reviewer in deciding that the quality of care was acceptable is not fully documented or available to the SuperPRO, perhaps because it was gained through face-to-face discussion with the physician in question. This diver- gence can produce a systematic bias against PROs, insofar as the PRO may determine on the basis of that information that the quality of care was in fact acceptable, whereas the SuperPRO will decide that the quality of care was not acceptable.) Third, and most important from HCFA's point of view, the agency will now use SuperPRO results as a formal rather than advisory part of its evaluation of PROs (Vibbert, 1989c) hence the strength- ening of the rebuttal process. Because HCFA expects PRO-SuperPRO disputes to arise, the agency will establish a referee in the form of a nationwide "Physician Consultant" contract. PROs will be able to appeal to this entity when they and the SuperPRO disagree about at least three cases with a common clinical cause. Appealable disagreements can include issues of necessity of admission and levels of quality-of-care citations but not DRG validation (Vibbert 1989c). Another option for adjudicating PRO-SuperPRO disputes is to ask the HCFA ROs to resolve such differences (OIG, 1989), although variable RO per

404 KATHLEE;bJN. LOHR AND AllISON J. WALKER formance for PROMPTS-2 may call this option into question. The physi- cian consultant contract may be a way to provide the necessary physician backup to the ROs if they are used when third opinions are needed in PRO-SuperPRO disputes. AMPRA 1989 Impact Survey Neither PROMPTS-2 nor the SuperPRO evaluation provides any con- crete assessment of how well PROs are doing in improving the quality of care rendered under Medicare, and no formal evaluation similar to HCFA's intensive 1979 evaluation of the PSRO program has been undertaken. To provide some information about impacts on quality and what PROs are doing to accomplish them, in mid-1989 AMPRA began the first of several contemplated surveys on the impact of PROB. Results were expected by late 1989. The first survey was intended to document PRO experience in the acute care setting for the first two contract cycles (i.e., roughly through 1988~. Data were sought for five major dimensions of PRO review: (l) the effect of Medicare's PPS on quality of care and the kinds of problems PROs found during hospital review; (2) rates of hospital use for certain procedures and medical conditions; (3) PRO quality objectives; (4) DRG validation; and (5) educational activities and successful quality interventions. A sixth part of the survey focused on desirable improvements to the PRO review program. DATA COLLECTION, MANAGEMENT, AND REPORTING REQUIREMENTS HCFA collects an enormous amount of data through the PRO program. The SOW requires specific areas of review responsibilities, and the data compiled or reported by the PROs reflect these responsibilities. This sec- tion briefly outlines three types of data generated by the PRO s on a periodic basis: profiles, PRO files (or PROFs), and management information re- ports. At fixed intervals (either monthly or quarterly), PROs electronically submit information to HCFA on each completed review, including the re- view determination. Owing to the large amount of data required, the PRO does not have to generate hard copies of all reports. Although the interval and method for submission vary depending on the number and type of cases reviewed by the PRO, the record format is standard and has been defined by HCFA. The data elements in the record constitute a minimum data set that must be collected for each review.

UTILIZATION AND QUALITY CONTROLPRO PROGRAM Profiles 405 PROs develop numerous provider- and practitioner-specific profiles for several purposes: to identify areas for focused review or objectives, to monitor PRO performance, to assess internal staff, and to identify aberra- tions in utilization or quality of care for further study. Those outlined in Table 8.11 are required as a minimum and must be produced quarterly. Under the first and second SOWs, the PRO was required to develop a profile of each provider, and for the immediate past 2-year period, to in- clude variables that can be computed using PRO or HCFA data. The aim is to identify aberrant providers and physicians so that the PRO can focus its activities where it expects potential problems. Profiling includes averaging hospitals and physicians so that those falling two standard deviations from (above or below) the norm (for length of stay, death rates, readmission rates, and transfer rates) can be identified. In addition to the provider pro- files, the PRO is to use its own experience in review and data analysis to design and produce profiles that will assist in the review process. PROFs A new system of reporting has been developed using the following seven PRO files, or "PROFs." PROFs 1 to 4 are produced monthly, the remainder quarterly. PROF 1, Pre-admission and Pre-discharge Review Record, collects all pre-admission review cases in which a final decision has been made; it includes all PRO interventions subsequent to notifying the provider or prac- titioner that a confirmed problem exists. This record also captures cases where review is performed before discharge. PROF 2, Universe Data Rec- ord, collects data on the universe of cases from which the random samples are drawn. PROF 3, Completed Review Record, collects the completed results of all cases reviewed either prepayment or retrospectively. PROF 4, Ambulatory Surgery Record, reports data on ambulatory surgery performed in HOPDs and ASCs, and it will reflect review of the required S-percent random sample. PROF 5, Quality Interventions Record, collects data on the number and sources of problems (e.g., physician, provider, other) where the quality intervention is education, intensification of review, other interven- tions, coordination with licensing and accrediting bodies, or sanction. (Notification of a quality problem does not require generation of a record to HCFA.) PROF 6, Cost Report Data, summarizes each PRO's accrued costs during a contract cycle, divided into costs related to review and costs not related to review. Costs for reviewing each type of case are also included in the report. PROF 7, Dynamic Objectives, summarizes new objectives

406 KATHLEEN N. LOHR AND ALLISON J. WALKER TABLE 8.11 Profiles that Peer Review Organizations (PROs) Must Perform A. Profiles of hospital statistics by: Average length of stay per diagnosis-related group (DRG) per quarter Average number of discharges per DRG per quarter Frequency of procedures identified for quality objectives Frequency of specific diagnoses identified for quality objectives Noncompliance with PRO review procedures Cases that fail admission and quality objectives Cases that fail generic screens, by screen and physician Mortality rates Day and cost outlier rates Transfer rates to other acute care facilities and exempt units Pre-admission denials Readmission rates Admission denial rates B. Profiles of physician statistics by: Dernal rates Admissions Readmissions Mortality rates Length of stay for premature discharge C. Profiles of other provider statistics, for example: Home health agencies, skilled nursing facilities, ambulatory surgical centers D. Profiles for internal quality control: Review determination denials, DRG validation findings, arid audit Review decisions by: Physician reviewers Review coordinators SOURCE: HCFA, 198B. that are proposed during the PRO contract and approved by the HCFA RO for resolution of utilization and quality problems. Mallagement Information Reports Management information reports have two purposes. The first is to help the PRO achieve efficient and effective operations, and PRO staff are to use the reports for decision making and monitoring contract compliance. Spe- cific reports are supposed to be developed throughout the life of the con

UTILIZATION AND QUALITY CONTROL PRO PROGRAM 407 tract in order to meet the changing needs of the particular organization. They are not regularly submitted to HCFA. The second purpose is to monitor review selection requirements. For example, the PRO is required to review a 3-percent random sample of all cases. To determine whether selection requirements have been met, the PRO needs to know the total number of and the percentage of cases selected each month. Management reports thus are supposed to reflect the same time periods as the related PROF report; for example, PROF 1 should be consis- tent with the PRO management report on pre-admission activities. PROIFI Data Exchange Reports Data exchange reports are produced monthly and monitor the frequency, volume, and effectiveness of the tape submission process from FIs. Infor- mation in these reports includes whether the tapes are processed within the 15-day requirement; whether the correct number of cases has been selected; error rates; types of errors; and whether all hospitals are reporting, and if not, why not. An adjustment report is also produced monthly to monitor the number of adjustments in data generated, forwarded, pending, and re- turned to the FI. Medical Review Activities Medical review activities generate four types of management informa- tion reports, which are produced on a monthly basis. Timing of review reports are used to monitor the adherence of the review process to required time frames. The number, type, location, and age (length of time in the pending category) of all pending cases by review category are included in the report. Review results reports monitor the nature and results of review determinations. Data on who is making the determinations, final determina- tions, and reconsiderations are included. Pre-admission review reports as- sist the PRO in monitoring the operation of the pre-admission review pro- cess through the identification of the number of phone calls expected and received, and the decision made by the type of procedure. Finally, sam- pling reports are initially generated on a monthly basis, but then only as needed once the sampling process has been determined to be adequate. The report includes the number of cases received for each review category, the number of cases selected, and the percentage selected. Quality Control The final category of required PRO management information reports in- cludes three types of reports. Monthly correspondence control reports

408 KATHLEEN N. LOHR AND AI1ISON J. WA=ER monitor the number and type of beneficiary complaints and inquiries to determine if a pattern is developing in a particular facility or with a particu- lar physician. Quality intervention reports monitor the various stages of quality interventions. These reports include the numbers of physicians or providers for which some intervention has been set in motion, the numbers of notifications, and the numbers of cases under formal education, intensi- fied review, and sanction, as well as the numbers that have been sent to the OIG. Finally, internal quality control reports, produced quarterly, concern information about the accuracy of decisions made by nurse reviewers and physician advisors that would enable PROs to identify problems in manage- ment. CONTROVERSIAL OR PROBLEMATIC ASPECTS OF PRO QUALITY-RELATED ACTIVITIES Issues Relating to PRO Activities in All Systems or Settings Generic Screens The initial experience with inpatient generic screens has come under considerable scrutiny and criticism. Recall that the screens are: (1) ade- quacy of discharge planning; (2) medical stability of the patient at dis- charge; (3) deaths; (4) nosocomial infections; (5) unscheduled return to surgery; and (6) trauma suffered in the hospital. Issues include extreme variation across PROs in rates of screen failures and confirmed quality problems, the low yield of confirmed quality prob- lems among cases reviewed, and the value of generic screening for cases that are already under review for other quality-related reasons. Several groups have compiled data on generic screening, including the OIG, GAO, the Prospective Payment Assessment Commission (ProPAC), and HCFA. For instance, the OIG calculated that the rate of screen failures ranged from 0.1 percent (for screens 1, 3, and 5) to 43 percent (for screen 2) and that the rate of confirmed quality problems ranged from 0 percent (screens 1 and 5) to 99 percent (screen 4) (OIG, 1988a). GAO calculations indicated that in the first 6 to 9 months of use ". . . in several PROs, less than 5 percent of cases failed any screen, while in others, more than 40 percent failed. The percent of confirmed quality problems found in cases with screen failures ranged from less than 5 to over 70 percent; about half of the PROs fell in the 20 to 50 percent confirmed failure confirmation range" (GAO, 1988a, p. 47~. HCFA itself noted that, after a year of experience, it had become apparent that the screens were not being uniformly applied. They found, for instance, that in some PRO areas the percentage of screen failures to confirmed problems was close to 100 percent for certain screens

UTILIZATION AND QUALITYCONTROLPRO PROGRAM 409 (e.g., the discharge planning screen) whereas in other PRO areas it was con- siderably lower. They concluded that the differences between PROs were too disparate to be attributable to differences in medical practice (H. Brook, personal communication, 1989~. In a separate report, GAO also noted that the PROs themselves rate generic screens behind nurse judgments and profiling and tied with intensi- fied review in terms of their effectiveness in identifying cases with possible quality-of-care problems (GAO, l988b). In this same report, GAO repeated its conclusion that the PROs vary considerably in the rates at which review coordinators fail cases and especially in the rates at which physician advi- sors confirm the quality problems.29 These aspects of generic screens were evidently still an issue through the period of this study. Calculations based on data compiled by HCFA through June 1989 reflect the wide variation across PROs in the incidence of screen failures and confirmed quality problems; depending on the specific screen, screen failures among cases reviewed ranged from 0.2 percent to 38.6 per- cent, and confirmed problems from 0 percent to 100 percent (Table 8.12~. One drawback of these rate calculations is that the percentages of con- firmed problems are based on a denominator of referred screen failures, not on the universe of cases reviewed. PROs that look quite different on the two measures may actually be detecting fairly similar rates of problems. What fraction of all charts reviewed actually reflects a true quality prob- lem? Table 8.13 gives the average rates of screen failures and confirmed problems among screen failures through June 1989. The upper panel is based on more than 6.3 million cases reviewed. The first two columns clearly reflect the highly dissimilar rates of screen failures and confirmed problems as a percentage of screen failures. The third column, which gives the percentage of confirmed problems among all cases reviewed, shows the very low overall yield of the screens. The most productive screens relate to adequacy of discharge planning and nosocomial infections. Another question sometimes raised about generic screens is related to the fact that they are applied to cases targeted for review for many reasons. A purer test of their usefulness might be to examine their yield in just the 3- percent sample, which can be said to represent the universe of Medicare admissions. The lower panel of Table S.13 reports on data for just the 3- percent random sample of all Medicare admissions. The yield for just that sample is roughly the same as that from all sources of reviewed cases. ProPAC (1989, Table 1.20) had earlier reported percentages of confirmed failures for the 3-percent sample that were lower than those among failures for all cases. The latter includes, of course, the randomly selected cases, cases selected for expected quality problems, and other cases picked for review that do not relate, presumptively, to quality problems (e.g., those required for review by virtue of being 1 of 12 Medicare Code Editor princi

410 KATHLEEPJN. LOlIR AND ALLISOlV J. WALKER TABLE 8.12 Range of Generic Screen Failures and Confirmed Problems (in Percentages), by Type of Screen Range of Generic Screen Screen Failures Confirmed Problems Adequacy of discharge planning 0.2-19.1 2.1-lOO.Oa Medical stability of patient at discharge 1.4-38.6 0.1-68.5 Deaths 0.~5.8 0.0-37.1 Nosocomial infections 1.1-20.4 0.4_95.7b Unscheduled return to surgery 0.2-2.6 0.0-66.7 Trauma suffered in hospital 1.0-24.1 2.2-65.8 aFour Peer Review Organizations (PROs) reported numbers that yielded a figure above 100%, the highest one squalling 109.1%. lone PRO reported numbers that yielded a figure of 106.8%. SOURCE: HCFA, 1989b. pal diagnoses). The data suggest, however, that even if the smaller num- bers in the random sample are taken into account, using generic screens on just that sample will not produce quality problems with any greater regular- ity than screening all cases reviewed. Another, and perhaps more pressing, issue is why PROs differ so dra- matically in the rate of referrals and confirmed problems. The process is supposed to be quite standardized (witness the evolution of interpretative guidelines described earlier). Nonetheless, the data strongly suggest that process does vary from PRO to PRO and from HCFA region to region. Believing that vagaries of smaller PROs might account for some of these differences, we analyzed data through June 1989 for just the eight largest PROs (tables available on request). Calculations were done as before for screen failures as a percentage of total cases reviewed, confirmed problems as a percentage of screen failures, and confirmed problems as a percentage of total cases reviewed. Several issues emerge from an examination of these numbers. The first issue is low yield. For instance, for the eight largest PROs, screen 1 fail- ures as a percentage of total cases reviewed ranged from less than 1 percent to less than ~ percent (national average, 3.05 percent); for screen 2, the range was from about 3 to 29 percent (national average, 12.47 percent). Further, confirmed problems as a percentage of screen failures was as low as 2 percent in several of the eight PROs (against a national average of 28.54 percent). The relatively low yields of screen failures and confirmed problems (al- though comforting in terms of the implications about quality of care that

UTILIZATION ED QUME CONTROL PRO PROGRAM 411 might be drawn) combine to produce very small yields of quality problems among all cases reviewed. For these eight PROs, screen failures and con- firrned problems ranged from a low of 0.01 percent (for several screens) to a high of about 8.3 percent (for nosocomial infections in one PRO). In short, differences across all PRO s do not seem to be a product of peculiari- ties of only the very small ones. A second issue is that of variability across screens. The figures for screen failures and confirmed problems by screen for the largest eight PROs show that the percentage of failures for individual screens range from less than 1 percent to more than 100 percent' again demonstrating the variability across PROs and across screens. Another facet of the differences across all PROs is that of quality prob- lems never detected (and hence never addressed) in any formal way. ProPAC TABLE 8.13 Percentages of Cases Failing Generic Screens and with Confirmed Problems, by Generic Screen and Universe of Cases Percent Percent Percent Confirmed Failing Confirmed Problems Generic Screen Screen Problems Among and Universe of Cases Review Among Failures Reviewed Cases All casesa Adequacy of discharge planning 3.05 71.27 2.18 Medical stability of patient 12.47 10.60 1.32 at discharge Deaths 1.49 7.50 0.11 Nosocomial infections 7.84 35.67 2.80 Unscheduled return to surgery 0.99 7.56 0.08 Trauma suffered in hospital 4.92 20.82 1.03 Cases in the 3-percent samples Adequacy of discharge planning 2.93 79.52 2.33 Medical stability of patient 12.87 10.76 1.39 at discharge Deaths 1.24 8.45 0.11 Nosocomial infections 6.53 31.84 2.08 Unscheduled return to surgery 0.62 7.07 0.04 Trauma suffered in hospital 4.05 21.46 0.87 aNumber of all cases reviewed: 6,309,839. bNumber of cases reviewed in 3-percent sample: 705,983. SOURCE: HCFA, 1989b.

412 KATHLEEN1`J. LOHR AbJD AllISON J. WALKER (1989), for instance, cited one study in which researchers estimated that as many problems were present among cases not flagged by screens (e.g., in about 5 percent of the cases reviewed) as were identified by the screens. The problems allegedly related more to insufficient attention or response to medical problems than to poor care actually provided. Several factors may account for these differences. For instance, nurse reviewers may differ in how narrowly or expansively they interpret the screens regardless of interpretative guidelines. What may be a failure for some may not be for others; nurse reviewers within PROs, and even indi- vidual nurse reviewers from day to day, may differ on what they see or do not see in a given medical chart. Moreover, because of the required case selection specified in their contracts' SOWs and the close relationship of the budgets to those required types of review, PROs may not select "extra" providers, physicians, or problems for review that they suspect may be delivering substandard care. (This may have been more true in the first and second SOWs than the third, because as of now HCFA will pay for reviews conducted for these reasons.) Furthermore, PROs will differ in the collection of cases to which they apply generic screens because they will have hospitals and physicians on 100-percent intensified review for different reasons and problems and be- cause they will have different mixes of hospital transfers to other types of units.30 Finally, some cases are targeted for review precisely because a quality problem is considered more likely (e.g., day or cost outliers; the first of a pair of admissions within 31 days; and most cases on intensified review). The question thus becomes the marginal productivity of the screens given that there is already reason to believe a quality problem might be present. The PRO community initially argued for this type of review tool to be used nationally (Lohr, 1985), and a majority of PRO officials and HCFA staff believe that screens have been at least moderately effective (GAO, 1988b; OIG, 1988a). Nevertheless, they have not proven to be entirely successful in efficiently identifying quality problems, and criticisms can be raised in three areas: their value overall; their productivity for detecting problems in a random set of admissions, which appears to be lower (or at least no better) than for a pool of cases where quality problems are already suspected; and their incremental utility in those latter, "suspicious" cases. Other difficulties remain. Their application is highly labor-intensive. Apparently they still result in considerable false-positives, regardless of the relaxation of the requirement that nurse reviewers must refer failures for physician review, and they have a nontrivial false-negative rate as well. Revisions to the generic screens that are incorporated into the third SOW are essentially untried as of this date. Furthermore, some PROs have found that their own additional screens do as good a job or better than the HCFA

UTIllZATION AND QUALlIYCONTROLPRO PROGRAM 413 screens.3~ Finally, there are numerous reasons why PROs can legitimately differ in the rate of cases detected by the screens, making conclusions about the uniformity of this tool difficult to draw. Thus, standard, well-known generic quality screens might permit HCFA and the public to track quality problems nationally (depending on how much one can assume about the reliability and validity of these, or any, generic screens). Less certain is whether they can or should be used to compare the performance of PROB. If the quality assurance process were expected to become more alike across the review entities (e.g., more national and more centrally directed), then the factors discussed above would have to be ad- dressed. If, however, the quality assurance program were to remain more responsive to "local" needs and circumstances, then flexibility and local autonomy about these elements of the review process are both more impor- tant and more acceptable. Whether generic screens are a strong and reliable tool on which to base a considerable part of the Medicare quality assurance effort seems problematic, and a rigorous evaluation of their utility and pro- ductivity is probably warranted. This experience with inpatient generic screens underscores the need for rigorous pilot-testing of similar instruments designed for nonhospital set- tings, where there is much less experience with them. Perhaps more impor- tantly, this experience argues for considerable testing and review of the computerized screening algorithms now being developed in the UCDS, in part because they are very different from what the PROs have used so far and in part because of the likely cost of implementing such an extensive data collection effort. Pre-procedure Review Whether PROs should be doing pre-procedure authorizations is part of a complicated issue concerning what entities should be doing physician re- view. It has generated considerable debate for the Physician Payment Re- view Commission (PPRC, 1989~. The debate reflects confusion in at least two areas. First is which entities (carriers, FIs, and/or PROs) have the most experience in prior authorization of procedures and are in the best position to carry it out on a wide scale. Carriers and FIs have a history of prepay- ment review of physician services more extensive than that of PROB. The second is whether this is first and foremost an exercise in utilization and cost control or in quality assurance. It may never be possible to draw a firm distinction between prior authorization activities (prospective utiliza- tion management as contrasted with retrospective utilization review) that serve a quality assurance function and those that control use of services. To the extent that the latter purpose is the preeminent one, however, the ques

414 KATHLEEN lI. LOHR AND ALLISON J. WAITER lion can be raised whether it detracts from the purported greater emphasis on quality intended for the PROB. Reviewing Physician Office-Based Care According to PPRC (1989), the following entities carry out the following tasks. Carriers process claims, conduct utilization review, and initiate sanc- tions for substandard or abusive practices (for inpatient physician visits and procedures, ambulatory physician visits and procedures, durable medical equipment, and diagnostic testing). Carrier reviews of submitted claims involve both prepayment and postpayment review, and medical reviews are done with close consultation of their own medical staffs. Fls do the same set of tasks, as well as limited quality review, for inpatient hospital care, outpatient hospital care, SNF and HHA care, hospice care, outpatient reha- bilitation facility care, and Part B outpatient physician therapy. PROs, of course, do utilization review, quality review, verify bill and DRG data, and initiate sanctions for substandard care for inpatient hospital care, TEFRA HMOs and CMPs, ASCs, and "intervening" EIHA and SNF care, with their role regarding physicians' services in office settings to be determined (PPRC, 1989, p. 239~. In short, the three groups have overlapping, or possibly conflicting, re- sponsibilities. They operate in different ways, with different data bases and different rules, such as when (before or after hearings) they can deny pay- ment and what information about review criteria and screen thresholds must be made public. They also collectively leave a big gap. According to PPRC (1989), none of these entities has specific responsibility for review- ing most Part B services for quality of care, although carriers have author- ity to deny payment and initiate sanctions for substandard, unnecessary, or inappropriate care. PROs, however, are charged with reviewing office- based (ambulatory) care, which they do not yet do (although the pilot proj- ect on office-based care has begun). In short, the picture of which agencies have what authority to review outpatient care for quality of care and to take action in the face of instances of poor care remains clouded. PPRC (1989) made four recommendations concerning Part B carrier and PRO utilization and quality review. First, HCFA should establish proce- dures to encourage input from carriers and PROs in designing utilization and quality criteria, in developing physician profiling methods, and in in- vestigating physicians suspected of providing inappropriate or substandard care or of billing inappropriately. Second, HCFA, carriers, and PROs should work together to delineate future roles of PROs in ambulatory care review. Third, PROs and carriers should consult with appropriate medical or- ganizations when developing review criteria (over and above what they are required to do now). Fourth, HCFA should designate a single entity to sup

urlLlzATloN AND QUAL17Y CONTROL PRO PROGRAM 415 port research, demonstrations, evaluations, and technical assistance for all three entities doing utilization and quality review. Peer Review In spite of the historical emphasis on peer review in federal programs and in the PRO SOW, physicians and hospitals heard from during this study widely contend that PRO reviewers are not "peers." The points in conten- tion concern rural practitioners and providers, physicians fully in private practice reviewed by physicians only partly still in practice (e.g., because they are semiretired), specialists not being reviewed by members of their own specialty, and physicians for whom the relatively low reimbursements for PRO review are an important portion of their income. Physicians in prepaid group practice settings reviewed by those in FFS settings are dis- cussed later. HCFA does require PRO physician reviewers to be in active practice and have admitting privileges in one or more hospitals in the PRO area. PROs we visited acknowledged the problems concerning rural areas and specialists but generally defended their record of using peers. They cited budget constraints as playing a large role in these problems: among these were not being able to maintain regional offices in rural areas and not being able to reimburse reviewers at competitive levels. The emerging debate about "quality denial letters" discussed elsewhere in this chapter is ex- pected to add to the problems of recruiting specialists and, especially, sub . . specla lStS. Sanctions Retention of PRO sanction authority. The role of PROs in the sanciion- ing process, and the role of sanctions in the quality assurance efforts of PROs, have both been misunderstood over the course of the program. With regard to the former, several points need to be emphasized. First, PROs can only recommend sanctions to the GIG, not invoke or enforce them. Second, in principle (although perhaps not in practice), practitioners and providers have a considerable number of due process steps open to them before a sanction is actually imposed. Consequently, PROs have little influence over the outcomes of a full sanction process, even when they have provided full documentation of their reasons for recommending a sanction. Nevertheless, PROs are virtually uniform in their view that having the sanction-recommendation capability is an indispensable tool in their deal- ings with providers and practitioners whose performance is unacceptable (GIG, 1988a). For instance, GAO reported that almost 70 percent of PROs rated sanctions as very or extremely effective for correcting quality prob

416 KATHLEEN N. LOHR AND ALLISON J. WALKER lems; the next most useful intervention was counseling, with about 45 per- cent of PROs rating this as very or extremely effective; about 30 percent of PROs gave those ratings to formal education and to communication with quality assurance committees and with hospitals (GAO, l988b). In short, PROs would not be willing to relinquish the sanctioning author- ity they now have in favor of simply greater educational or persuasive interventions. William Moncrief, M.D., representing AMPRA at a hearing before the Committee on Governmental Operations in April 1989, testified that the "PRO sanction activity is important to the effectiveness of the program," describing it as "a careful and extensive" process involving "ex- tended consultation and due process procedures." This was also the view of those PRO officials who provided information to this study through the public hearings and site visits, although some cited the costs (in terms of financial and human resources) as a significant barrier to their effective pursuit of all but the most egregious cases. Elimination of PRO sanction authority. The most radical suggestion about the PRO authority to recommend sanctions is to eliminate it. Several arguments have been advanced for this view (Jost, 1989~. Critics believe that the educational and enforcement dimensions of the program are not compatible and that PROs have overemphasized one function or the other. Others argue that PROs lack legal expertise, which has led to inadequate protection for the due process rights of sanction targets. Thus, one proposal is to turn the sanction function over entirely to the OIG.32 PROs would continue their quality assurance functions by monitor- ing medical records, investigating complaints, and requiring corrective ac- tion where quality problems were identified;-they could put into effect other interventions that they currently use, such as limiting practice, mandatory consultation, preceptorships, attendance in residency programs, obtaining board certification, and requiring oversight in surgery (Jost, 1989~. When, however, they discover problems on which formerly they would have taken sanction actions, now they would refer the matter without recommendation to the OIG for investigation and possible action. The OIG, either with its own medical resources (which it would have to develop) or with experts borrowed from the PRO (or perhaps elsewhere, such as HCFA ROs or the FIs), would build its case, decide on an appropri- ate penalty, and present the case to a specially trained ALJ. The ALJ would decide whether to impose a sanction and the nature and extent of the pen- alty (Jost, 1989~. The OIG would, however, be given summary suspension power where the situation indicated that, because the practitioner poses an immediate risk to beneficiaries, immediate action is warranted (Jost, 1989~. This revised system allegedly has several advantages. It would allow the PRO to concentrate solely on its primary function, peer review. It would

UTILIZATION AND QUALITY CONTROL PRO PROGRAM 417 simplify the currently complicated procedures and remove PROs from the disputes about due process (Jost, 1989~. It would also be preferable from the perspective of practitioners who may be subject to investigation and sanction, because they would receive a hearing before an impartial ALJ before any sanction could be imposed (Jost, 1989~; this improvement, how- ever, could be made without changing the PROs' sanction authority. In short, it is claimed that the revised system would permit the PROs to carry out their most important function promoting quality of care for Medicare beneficiaries by threatened action (i.e., referral to the OIG) against of- fending practitioners and providers while avoiding much of the hostility and mistrust that the provider community currently feels toward them. This proposal has met significant opposition. The PROs fear the loss of some of their enforcement powers. The administrative conference (ACUS) has rejected it (Jost, 1989~. In view of the difficulties of the PSRO pro- gram which did not have the same sanctioning (or related regulatory) powers that the PRO program has weakening these powers for the PRO program does not seem to be an attractive option. Correcting some of the other problems of the entire sanctioning process, however, does appear to offer ways to strengthen the government's ability to protect the quality of care delivered to Medicare beneficiaries. Monetary penalties. PROs strongly advocate that a broader, or more finely tuned, set of sanction recommendations be open to them, viewing the choice of exclusion of the provider or a nearly meaningless monetary pen- alty as essentially no choice at all. That is, no viable sanction measure short of exclusion has really been available to them.33 (In the visits to PROs as part of this study, this issue tended to be seen as part of a larger need to strengthen all the intervention options open to PROs, not just the sanction capabilities.) In its 1988 review, the GIG recommended that the monetary penalty option be strengthened by allowing PROs to propose a fine of up to $10,000 per violation of Medicare obligations (i.e., the provision of substandard, unnecessary, or uneconomical care). GAO seconded this stand by recom- mending that the relevant congressional committees enact legislation setting a fixed upper limit to monetary policies in place of the present cost-based limit. HCFA agreed to assist the GIG in drafting such legislation. The ACUS also recommended that DHHS be permitted to assess "a substantial" fine against providers and practitioners (Vibbert, l989c). Countering these proposals were recommendations from the House Commerce Committee to cap the monetary fines at just $2,500 (Vibbert, 1989d), a step that would effectively nullify the effort to shore up the monetary penalty option of sanctions. Exactly how this issue would be resolved was uncertain at the time this report was completed.

418 KATHLEEN N. LOHR AND ALLISON J. WIDER The "unwilling and unable" provisions. The requirement that a provider or practitioner be found "unable or unwilling" to meet their Medicare obli- gations (in addition to finding that they have not in fact complied with those obligations) has caused unending confusion and frustration with the sanction process. Among other etchings, it effectively required that PROs demonstrate that a doctor or institution was unable to conform to a correc- tive action plan, because the physician or institution can claim that it was both able and willing to meet the specified obligations or corrective action. Thus, although the "unable and unwilling" provision can be seen as a desir- able aspect of due process for physicians or institutions, it is a nearly im- possible task and one that can take a very long period of time (during which the practitioner or provider could continue to provide services). Providers, for their part, have complained that PROs do not explain the basis for the "unwilling or unable" finding and do not give physicians a chance to respond to such determinations before the PROs impose sanc- tions. In one case (Lavapies v. Bowen), the federal district court held that the matter of the practitioner's alleged unwillingness or inability to comply was a serious question and that she had been denied an opportunity to present her side of the case under existing procedures; on that basis, the court issued a preliminary Injunction against the Flu s sanction. The problems were sufficiently apparent and persuasive toward the end of 1988 that the GIG recommended that DHHS submit a legislative pro- posal to the effect that failure to comply with patient care obligations was sufficient basis for sanctioning (GIG, 1988b). The ACUS has endorsed a recommendation that would remove this requirement to prove that a physi- cian or institution is "unwilling and unable" to meet Medicare obligations before sanctioning. Because this seemingly constricts due process, the ACUS has also suggested that all physicians now be afforded the opportunity to seek pre-exclusion stays (not just physicians in rural areas, as is the case in mid-l989~. Such stays mean that sanctioned providers can have exclusions lifted pending the outcome of a full evidentiary hearing. The ACUS would, however, require that the burden of proof in such hearings be shifted from the government (to prove that the sanction should be upheld) to the sanc- tioned physician (to prove that it should not). Such stays may not be regarded as sufficient protection for sanctioned practitioners, however, and the American Medical Association is expected to push for the retention of the "unwilling and unable" provisions (Vibbert, 1989c). Moreover, the House Commerce Committee has recommended that the time during which a physician can claim to be willing and able be expanded (Vibbert l989d). Hence, as with monetary penalties, the question of whether PRO sanction capabilities will be strengthened or weakened remains open.

UTILIZATION kD QUME CONTROL PRO PROGRAM 419 Timing of sanctions before hearings. Practitioners and providers sanc- tioned by the OIG are excluded from the Medicare program before they can obtain an ALJ hearing. They have argued that the procedures to safeguard their rights are inadequate to protect their constitutional rights under the Due Process clause of the Fifth Amendment. Although many courts have agreed that either a constitutionally recognized property interest or liberty interest is implicated in this exclusionary process, they have uniformly held that the government's interest in expediting procedures to protect Medicare patients outweighs the physician's or the institution's interest in increased procedural protection. The multiple levels of case review before the OIG imposes sanctions are seen as affording substantial protection already. Adequacy of notice of grounds for sanction. The concept of not meeting "professionally recognized standards of care" has evidently been confusing to some parties (from PROs to ALJs). Despite specific requirements in the regulations that the practitioner or provider be informed of the nature of the violation, the basis for the PRO determination, and the procedure rights due, PROs have in some cases issued vague charges and in other instances raised new issues at sanction meetings that were not reflected in the original notice (Jost, 1989~. One possible result has been a high rate of reversals of OIG sanction actions by ALJs (10 of 18 cases by one recent count) (Vib- bert, 1989b). In addition, in an effort to correct this problem, HCFA has issued model notice letters for PROs to use (Jost, 1989), but because of their legalistic and punitive tone' these are not universally liked in the PRO community.34 Denials for Substandard Quality of Care COBRA and OBRA 1987 allow PROs to deny Medicare payment for substandard quality of care, although the provisions of the legislation were not implemented awaiting final regulations; a draft proposed rule to effect these requirements was published in January 1989 (Federal Register, 19891. It required payment to be denied when substandard care resulted in actual, significant adverse effects on the beneficiary or placed the beneficiary in imminent danger of health, safety, or well-being (i.e., places the beneficiary in a situation that constitutes a "gross and flagrant violation"~. The former condition (actual . . . effects) was defined quite broadly, as including unnec- essarily prolonged treatment, medical complications, readmission, physio- logical or anatomical impairment, disability, or death. To protect the concept of peer review, the proposed rule specified that physician advisors engaged in initial denial determinations of substandard quality be specialists in the same field as the attending or consulting physi

420 KATHLEEN N. LOHR AND AlllSON J. WALKER clan whose care is under question. The requirement could be met by having the initial determination made by a physician reviewer who is actually in the same specialty, by having a second level review by such a specialist when the original determination was made by a physician in a different specialty, or by having a committee of specialists, at least one of whom is in the same field. This requirement could be relaxed, however, when meet- ing it would compromise the effectiveness or efficiency of PRO activities. The proposed rule further provided that hospitals would be held finan- cially liable even if they did not contribute directly to the substandard care rendered by the physician. Thus, any denial of physician payment on these grounds would also result in a denial of reimbursement to the hospital. Furthermore, physicians may not charge patients for the care denied for these reasons and, if they have done so, must refund those payments to the beneficiary. The proposal then specified that the PRO shall notify the patient when such payment has been denied on the basis of substandard care. The key section was to read: "Our determination [concerning denial of Medicare payment of a hospital admission or physician services provided in connec- tion with that admission] is based on a review of your medical records, which indicates that the quality of services you received does not meet professionally recognized standards of health care. Denial decisions are made by the PRO physician. Your attending physician and hospital were given an opportunity to discuss your case with the PRO before the denial decision was made . . . " (Federal Register, 1989, p. 1966~. In the initial proposal, this letter would have been sent before providers were able to exercise their rights to appeal (i.e., to have the case reconsid- ered), rather than after a final determination had been made. The notion that the initial denial determination would (supposedly) have been made by a physician in the same specialty as the physician whose care was being denied and the fact that there was a 30-day opportunity to discuss it was evidently considered sufficient protection for that practitioner whose patient would receive such a letter. The entire quality denial process prompted much debate, especially dur- ing the time this study was being conducted. Among the concerns was the lack of protection for physicians if they could not invoke their full due process rights to reconsideration before their patients are notified of such denials, especially because of the high proportion (more than 45 percent) of denial decisions that are now overturned by the PROs after reconsideration (Vibbert, 1989b). The "same-specialty" requirement does not, in many observers' minds, adequately mitigate this problem.3S Most professional and consumer interest groups thus support early opportunities for appeal. The ACUS has recommended that DHHS proceed expeditiously to imple- ment PRO quality denial authority, but with appeals before patient notifica- tion (Vibbert l989c).

UTILIZATION AND QUALITY CONTROL PRO PROGRAM 421 Another criticism centered on the impetus such letters might provide for increased malpractice suits filed by beneficiaries and the impact of higher litigation on PRO activities. There are also concerns about the expected increased difficulties in recruiting the specialists that will be needed to participate in these reviews and decisions. A fourth concern focused on how much specific information the PRO should have to put in the letter to the beneficiary. Some wanted to keep the letters general but specify that care was substandard, others wanted more specificity about what was dis- covered that led to that decision, and yet others wanted the letters to say only that care did not meet Medicare payment guidelines (and not refer to the denial as a quality denial) (Vibbert, 1989b). OBRA 1989, passed in late November 1989 (after the main part of this study had been completed), addressed some of these issues (Congressional Record, House, November 21, 19893. First, it protected the physician or institutional provider from unwarranted notices to patients. Specifically, it provided that the PRO should not notify beneficiaries until after the PRO had notified practitioners or providers of its determination about the quality problem and their right to a reconsideration; if the practitioner or provider requests such a reconsideration, then one would be conducted before any notices to beneficiaries. Second, it softened the wording of the beneficiary notice, by saying that the letter need only state: "In the judgment of the peer review organization, the medical care received was not acceptable under the Medicare program. The reasons for the denial have been dis- cussed with your physician and hospital" (p. 9380~. Administrative Procedures and Public Oversight As noted at the outset of this chapter, the authority for PRO activities resides in several legislative acts; in a broad array of regulations, guide- lines, and directives; and in various documents such as the PRO Manual, the contract SOWs, and periodic instructions. The practice of relying on Manual transmittals, contracts, and other less formal instructions, instead of promulgating regulations through "public notice and comment" rulemaking as required by the APA, has raised serious questions. This is not unique to the PRO program, however, as DHHS has frequently sought to carry out its regulation of the Medicare and Medicaid programs by means less formal than this relatively burdensome rulemaking method. The APA requires that the public be given an opportunity to comment on proposed rules, that the comments be considered by the agency, and that the final rules be published at least 30 days before their effective date. Section 603 of the APA requires regulatory flexibility analyses, and executive or- ders require that the OMB review some rules. All these make notice and comment rulemaking under the APA time-consuming and burdensome, and it is not unusual for the promulgation of a rule to take a year Cost, 1989~.

422 KATlILEE~ N. LOHR AND ALLISON J. WALKER Experts argue, however, Hat sound policy reasons support using the more cumbersome rulemaking process (Jost, 1989). It promotes public participa- tion and fairness to parties who will be affected by the rules; it also forces the agencies to consider their proposals with greater care and to express them clearly. In foregoing the rulemalcing process, HCFA has opened the door to accusations that it is attempting to govern the PRO program through "a continual and confusing stream of instructions [that has] severely ham- pered their ability to carry out their mandate" Cost, 1989) and earned the hostility of those governed by He program. Legal suits36 and legislation in the last few years have clanged the situ- ation somewhat. OBRA 1987 calls for substantive changes in PRO contract provisions to be published in the Federal Register, although not necessarily according to APA procedures. Section 4035(a) specifically provides that "no rule, requirement, or other statement of policy (other than a national coverage determination) that establishes or changes a substantive legal stan- dard governing the scope of benefits, the payment for services, or the eligi- bility of individuals, entities or organizations to furnish or receive services or benefits under this title shall take effect unless it is promulgated by the Secretary by regulation ...." Nevertheless, the question of public access to, understanding of, and ability to comment on the myriad rules governing the PRO program remains important. One commentator suggests that in view of congressional dead- lines, DHHS should whenever possible publish PRO SOWs and any changes or modifications made during the conuact cycle for an abbreviated period for comments and should publish final provisions at least 30 days before their effective date (Jost, 1989~. Others have urged DHHS to make PRO contracts, interpretive rules, statements of policy, and guidelines of general applicability available in places of easy public access and to publish up- dated lists of these materials every three months (Jost, 1989~. The ACUS contends that DHHS should use formal rulemaking procedures "when seek- ing PRO program changes except when the agency has 'good cause' to believe the process is 'impracticable, unnecessary, or contrary to the public interest"' (Vibbert, l989c, p. 6), should publish SOW rules in the Federal Register with 30 to 45 days for public comment, and should publish PRO contracts, Manual instructions, and other guidelines quarterly in the Fed- eral Register (Vibbert, l989c). Although steps such as these might be seen as markedly constraining the ability of HCFA to administer the PRO program, the gains in terms of useful and timely reaction beforehand to proposed changes and modifica- tions to the program might be substantial. It might facilitate decision mak- ing as to when new program initiatives (e.g., new review or data collection instruments or methods) should be extensively pilot- and field-tested and when such rigorous examination can be foregone. Moreover the increased

UTILIZATION AND QUAL17Y CONTROL PRO PROGRAMS 423 sense that program administration is open to public scrutiny might be a considerable benefit to a program that is often poorly understood by benefi- ciaries and health care providers alike. Evaluation Considerable criticism can be leveled at how the PRO program itself is evaluated, especially in terms of its impact on the quality of care. Virtually no reliable or comprehensive examination of PRO program impact has been done within DHHS. The several careful external investigations by the GIG and GAO have tended to focus on specific operational aspects (e.g., useful- ness of generic screens; structural aspects of PROs). How HCFA evaluates individual PRO performance can also be criticized. Existing tools such as PROMPTS-2, although in transition to improved efficiency, have not been especially successful at providing a coordinated approach to evaluation (although PROMPTS-2 evidently is more productive that the SuperPRO effort). The GIG in particular has criticized HCFA's ability to assess PRO performance (OIG, 1989~. When combined with the lack of public over- sight and accountability noted above, these problems appear to have high priority for attention and correction. Collectively, these efforts underscore the need for considerably more innovative and comprehensive review of the impact of the PRO program on quality of care (as contrasted with narrow evaluations of how well individ- ual PROs appear to be applying review methods or complying with contract specifications). The requirement in the third SOW for PROs to publish an "annual report" may prove to be a useful step in that direction. The changes in PROMPTS-2 and perhaps the SuperPRO effort may also prove useful over the longer run. Another step that some have recommended is to appoint a "national council" of outside experts that could oversee the operations of the PRO program and report periodically (e.g., annually) to the Secretary of DHHS or to the Congress, or both, on the progress, successes, and problems of the program. Such a council might also be very useful in solving some of the public oversight and administration issues noted above. Precedent for such a step might be found either in the former PSRO National Council or in the appointment of the advisory commissions established since the advent of PPS, namely ProPAC and PPRC. Congress established ProPAC in 1983 in response to concern about the need to monitor and update the PPS system. The 17 members of ProPAC are appointed by the director of OTA for a 3-year term. The commission has two major responsibilities; first, it recommends annual changes in the hospital payment rates to the Secretary of DHHS and, second, it recom- mends changes in the DRG classifications. It has discharged these respon

424 KATHLEEN N. LOHR ACID) AI1ISON J. WALKER sibilities through regular commission meetings, which are generally open to the public, and detailed annual reports and technical documents. The stat- ute limits the number of staff to 26, and the annual budget for the ProPAC activities approaches $4 million. Physician payment under Medicare has come under close scrutiny in the last several years, and Congress established PPRC in 1986 to advise on reforms to the physician payment system under the Medicare program (PPRC, 1988, 1989~. PPRC is modeled on ProPAC (e.g., members appointed by the director of OTA for a 3-year term, and a staff limited to 26 persons). This commission plays four major roles (PPRC, 1988~. First, it provides advice to the Secretary of DHHS. Second, it seeks the views of physicians, benefi- ciaries, and others concerning its recommendations. The conduct of analy- ses to form the basis for policy decisions is the third role of the commis- sion. Finally, it undertakes the work necessary to implement the recom mended policy changes. PPRC has initiated several research projects to develop a Medicare fee schedule for physician payment based on the relative value of resources used by the physicians to produce the services rendered to the patient (see Hsiao et al., 1988~. The implementation of such a system, however, will not take place for several years. PPRC has also taken the lead in calling attention to the need for "practice guidelines" to help physicians understand better when services, especially procedures, are appropriate and when they are not. Issues Relating to HHA Review PROs that had begun HHA review during the site visits for this study noted two significant problems. First, selecting an appropriate sample for this task requires that hospitals bill for the two admissions in a reasonably timely way. At least one PRO noted, however, that some hospitals bill for two admissions more than 31 days apart (which would not constitute a reviewable readmission); only much later would the hospital bill for the admission that occurred within 31 days of the first admission (which would be a reviewable readmission). This practice severely complicates the iden- tification of 31-day readmissions and hence of cases that would constitute the potential pool of HHA care. A related sampling problem is simply that the pool of HHA cases for readmissions only is itself small, and whether it is representative of all HHA care is unknown. Second, at least one PRO noted that the HHA sector is undergoing great growth and change, which includes the emergence and disappearance of "fly-by-night" agencies. An agency might be out of business by the time the PRO knew what cases of HHA intervening care had fallen into its sample. Review in that case probably would be impossible and certainly

UTILIZATION AND QUALITY CONTROL PRO PROGRAM 425 would be moot. Other technical problems involving data and records have also been noted (Vibbert, 1989d). Issues Relating to HMO-CMP Review Peer Review Physicians in prepaid group plans have always argued that physicians in FFS practice are not well placed (and historically not well disposed) to judge the care from HMOs and similar plans on a "peer basis"; the premises underlying prepaid practice and the resulting styles of practice are simply too different. This perception reflects the historic concern of the HMO community that their care will be unfairly and improperly judged by the standards of the larger and more powerful FFS community. These concerns are heightened by the reliance to date on implicit physician review of cases referred to them through the nurse-reviewer screening process rather than on more explicit quality-of-care criteria developed with some appreciation of the different practice styles of the prepaid setting. Thus, there is some concern that using "local standards of care" may perpetuate existing prac- tice patterns and vitiate the potential of prepaid systems for innovation and improved service to the Medicare population. To meet this objection from the HMO industry, the HMO-C SOW requires the PRO to use HMO-CMP physicians to perform reviews when- ever possible. In addition, conflicts of interest can arise in several in- stances: when only FFS physicians review HMO-CMP care, when HMO- CMP physicians review care rendered by a plan from which they may re- ceive financial benefit, and when HMO-CMP physicians review care from a competing plan. Thus, PROs are required to develop a means of addressing these possible situations. In addition, PROs are expected to develop a quality control mechanism that will protect the integrity of the review pro- cess as it is applied to HMOs and CMPs (e.g., by creating a committee whose membership includes one representative from each plan under re- view in the state). Records and Case Selection Cases for inpatient review are to be selected on the basis of claims (the UB-82 Medicare hospital bill) submitted to FIs that are aggregated into a file (the "UNIBILL" file) forwarded to the PROs; the PROs in turn select their samples from this file. For the FFS system, this procedure works well; because payment is tied to the submission of a bill, this record of hospital admissions can be assumed to be quite complete. For the HMO and CMP system, this process does not work well at all

426 KATH~ENN. LOHR AND AI1lSON J. WALKER and results in a very inadequate '4universe" of inpatient claims for several reasons. HMOs and CMPs (and the hospitals with which they have various payment arrangements when they do not own their own) have no particular incentive to file such claims because payment is not related to them. More- over, in many cases FIs were not processing UB-82s submitted by risk- contractors or their affiliated hospitals or were not forwarding the UNIBILL file on to the PROB. Although efforts were made to force hospitals to prepare and submit these bills to the FIs, HCFA estimated that still only about half are being submitted (O'Kane, 1989~. To overcome this still inadequate pool of cases from which to select samples of cases for review, HCEA now requires HMOs to, keep a separate record of all admissions; those not reflected in UNIBILL data will be sub- jected to their own random sampling procedure so that the required level of review will be reached. Because HMO s and CMPs will differ in the propor- tion of their total hospitalizations subject to this form of random sampling, an additional source of variability has been added to review in the risk- contracting segment of the Medicare program.37 Apart from the problem of identifying which cases to review is the con- siderable set of obstacles that exist to acquiring the charts. Obtaining hos- pital charts is not appreciably more difficult for the HMO and CMP sector than for the FFS sector, and both systems contend that low reimbursement of copying costs (just over $0.049 per page) and lack of reimbursement for administrative costs are problems.38 For outpatient records, however, the problems can be extreme, when records for one plan must be retrieved from numerous health centers. Although the problem is manageable for most group- and staff-model HMOs and even for group network models, it pre- sents independent practice associations (IPA-model HMOs) with extraordi- narily complex logistics, because large plans of this sort may have hundreds of physicians practicing in individual offices. HCFA has indicated it would support legislation to allow HMOs to be reimbursed for administrative costs of retrieving such records, which should alleviate the problem to some degree. Limited Review One of the more contentious issues in HMO-CMP review has been the limited success of so-called limited review (only 11 of 133 risk contractors are currently on limited review), and several factors seem to be at work. First, the notion of reviewing an HMO's own quality assurance plan was unfamiliar to PROs, and HMOs may thus have been reluctant to put them- selves in the position of having their internal programs subjected to an unpredictable and uneven evaluation process. Second, PROs are expected to review all care rendered in a case that falls into the limited review sample,

UTILIZATION AND QU~ CONTROLPRO PROGRAM 427 even if, in the HMO's own program, only selected parts of that care were subject to review (e.g., as part of a focused audit study of some particular problem that the HMO was evaluating in depth). The HMO thus becomes liable for a failure ("Outcome No. 1," a quality problem found by the PRO but not by the lIMO) relating to care it had never reviewed as part of a quality assurance plan that the PRO had found acceptable. Third, the main argument for limited review was that it reduced the number of cases subject to review; in practice, however, HMOs subject to limited review can end up having as many cases reviewed (but no more than) as HMOs on basic review. Finally, because of the second factor (the scope of PRO review versus that of HMO review under its own quality assurance program), the HMO under limited review in theory can run more of a risk than the HMO under basic review of being subjected to intensified review on the basis of PRO findings on the "validation subsample" (i.e., cases having been reviewed by the HMO pursuant to its own quality assur- ance plan). Ambulatory Care Review With respect to ambulatory review, the interesting question is how physi- cians will respond to review of the care they provide in their own offices. In the HMO environment, physician resistance might add to the incentive for plans to withdraw from the risk-contract program, especially if the re- quirement for office-based review in the FFS setting is not fully imple- mented for some years. As noted earlier, FFS office-based review is ex- pecte~d to be started as a pilot project. Given the lack of experience and proven tools for ambulatory review, this is arguably a good strategy for the PRO program. It does, however, place the HMO community in a position that they can understandably regard as unfair. Accountability Who is responsible for quality problems is a question that arises in any health care delivery system, but it is especially salient in the complex world of prepaid group practice arrangements. For instance, for HMOs that do not own their own hospitals, that contract for certain types of care (e.g., subspecialty care), or that cover their members on a FFS or contractual basis for out-of-plan (e.g., out-of-region) care, the issue of whether they are accountable for care well beyond their ability to oversee or control becomes very complicated. In situations when an HMO's patients are not concen- trated in only a few hospitals or are otherwise widely dispersed across practitioners, the HMO can find itself held responsible by the PRO program for quality problems but without any authority or ability to monitor or

428 KATlILEE:N N. LOHR AND ALLISON J. WALKER control the performance of those providers, except over the very long run when a sufficient number of patients have had experience with those pro- viders. The question of accountability may be more complicated by legal prece- dent and rulings concerning whether entities that employ physicians and other professionals are held to different standards than those that only con- tract with physicians (essentially the distinction between group and staff models on the one hand and IPA-type models on the other). Some IPA physicians hold that they should not be subject to review until such time as all physicians who contract with Medicare (presumably meaning all FFS physicians) are reviewed. This may be especially salient in the light of a case (Harrell v. Total Health Care, Inc.~39 in which the court held that an HMO as a corporate entity could be liable for the negligence of contracting referral physicians. Other Issues Relating to HMO-CMP Review PROs differ dramatically in the proportions of quality problems they find in HMOs; one accounting showed a range of 1.8 percent in one state to upwards of 30 percent in three states (HCFA data cited by O'Kane, 1989~. Although it may be the case that HMO quality differs as much as 15-fold across the states, variations of that magnitude call into question more than the true quality of the care being rendered, most especially the validity of inferences drawn about comparisons of HMOs with each other and with the FFS system. It also raises the issue of whether HMO s operating in several or many states are subjected to the "same" review, because the PRO in each state is responsible for the state-specific portion of the HMO risk-contract care. In the site visits for this study, multistate HMOs had different views on this issue. At least one felt very strongly that it wished to deal with only a single PRO because it was experiencing considerable, unexplainable vari- ation in review from the different PROs in the different states in which it operated. By contrast, at least one other plan found PRO review suffi- ciently nontroublesome that differences across PROs were either not notice- able or not a problem. The question of valid comparisons is especially problematic for those states with only a single risk contractor, because information about numbers of cases reviewed and numbers and percentages of quality problems cannot be protected from public disclosure. For an HMO with an "unblemished" record, this obviously poses no problems, but for an HMO with a "poor" record the risk to its competitive position (vis-a-vis other HMOs in the state) could be considerable.

UTILIZATION AND QUALITY CONTROLPRO PROGRAM CONCLUDING REMARKS 429 This chapter has presented information on the current Medicare PRO program. It is intended mainly as a reference document for Volume I, Chapter 6, of the IOM study committee's report. The history of the pro- gram for review of care rendered in both FFS and prepaid group practice settings is recounted with emphasis on the most recent (third) SOW. Some problems related uniquely to HMO and CMP review are described, in addi- tion to current special projects and other activities of the PROs (or HSQB and HCFA). Among the issues discussed are the following: the usefulness of hospital generic screens, the practical question of ensuring review by peers, the limitations of the current sanctioning process and denials for substandard care, the adequacy of public oversight and adminis~aiive procedures for the program, and the lack of any systematic program evaluation to date. In formulating its mission of a model quality assurance program, the IOM study committee took many of these issues under advisement, intending that its proposed program overcome some of the more troublesome drawbacks (e.g., He adversarial "regulatory" aspects of the program and the lack of public accountability) while building on the considerable expertise, skills, and experience of the PRO community. NOTES 1. The authors wish to acknowledge the perceptive and timely assistance of Alan Kaplan, a Washington, D.C. attorney and consultant, at various stages of the prepa- ration of this overview of the PRO program. Harvey Brook and John Spiegel of the Health Standards and Quality Bureau were always responsive to questions and re- quests for data, and we thank them and their staff for a detailed and helpful review of this chapter. We also thank Andrew Smith and Maxwell Mehlman, authors of a related background paper for this study, for important legal and regulatory informa- tion on which this chapter in part relies. Finally, the commissioned paper on the history of and current issues relating to PRO review of risk-contract prepaid group practices, by Margaret O'Kane, was a valuable background document. 2. The GIG (1989) notes that HCPA appears not to have formally published the rationales for the sampling methods imposed on PROB. They also contend that HCFA apparently sets most of the specific requirements for review internally, usu- ally without pilot-testing. 3. A "substantial violation in a substantial number of cases" is a pattern of care that is inappropriate, unnecessary, does not meet recognized professional standards of care, or is not supported by documentation required by the PRO. 4. "Gross and flagrant violation" means a violation of an obligation (in one or more cases) that represents an imminent danger to a Medicare beneficiary's health, safety, or well-being, or that places a beneficiary at an unnecessarily high risk.

430 KATH~FEW N LOHR AND ALLISON J. WALKER 5. More cases are shown as reviewed than as selected for review. This may be an artifact of counting practices. For instance, transfer and readmission categories in the reviewed classification actually involve two or more cases, and Medicare code editor cases are reported under both pre-admission-prepayment and retrospec- tive review (HCFA, 1989b). 6. Interpretative guidelines for generic screens include two features, exclusions and explanatory notes. For example, exclusions for adequacy of discharge planning include death, transfer to an acute short-term general hospital or swing bed status, or patient left against medical advice. Explanatory notes for nosocomial infection are another example. They state: "A screen failure is not necessarily a confirmed problem. A screen failure occurs when more than one indicator of an infection is identified more than 72 hours after admission. Indicators [are]: temperature eleva- tion of 101 degrees Fahrenheit or greater (oral) (38.9 Centigrade); elevated white blood cell and/or left shift; isolation of organism from body fluids or specimens; appropriate radiographic imaging abnormalities; purulent draining; heat, redness, focal tenderness and/or pain; pyuria, dysuria; and productive cough. When the case has two or more indications of a nosocomial infection (i.e., a screen failure), you are encouraged to refer to the CDC's [Centers for Disease Control] guidelines to deter- mine whether there was a nosocomial infection. The presence of a nosocomial infec- tion is always a confirmed quality problem. Treatment of the nosocomial infection does not negate the confirmed quality problem." 7. Prior-authorization criteria vary by PROB. For cataracts, for instance, the Delmarva PRO indications for the procedure require visual acuity of 20/50 or worse in the affected eye for distance, or visual acuity of 20/40 or worse in the affected eye for near vision. Three other criteria are also (ambiguously) stated: visual acuity is interfering with the patient's lifestyle; cataracts are causing another ocular dis- ease; arid cataracts are preventing treatment for another disease. (Another set of specifications serve as indications for admission for the procedure.) The New York PRO indications for the procedure require the patient to meet one specific criterion (cataract removal will improve the patient's visual performance for daily activities, employment, and/or recreation) and one of a number of others, including vision in the operative eye less than 20/60, or the nonoperative eye is phakic and visual acuity is less than 20/40. 8. This chapter refers in various places to practitioners and providers; the distinction is a term of art for the PRO program in that practitioners would be used to refer to physicians or other individual clinical caregivers, and providers to facili- ties and institutions such as hospitals, SNFs, HHAs, as well as HMOs and CMPs. 9. Comments to study site visitors from PRO officials indicated that many Severity Level I cases ultimately turn out not to be quality problems as defined, because they are related to poor documentation. The "quality problem" is not confirmed when He target physician or provider provides satisfactory additional information. 10. HCFA prescribes the format and wording of these initial sanction notices (apart from the specifics of the case at hand). Presumably for legal reasons, they are very formal in tone and must contain the following information: (1) the obligations involved; (2) description of the activity resulting in the violation; (3) the authority and responsibility of the PRO to report violations of obligations; (4) a suggested

UTILIZATION AND QUALlTYCOlITROLPRO PROGRAM 431 method and time period for correcting the problem (at the discretion of the PRO); (5) an invitation to submit additional information or discuss the problem with the PRO within 20 days of the notice; and (6) a summary of the information used by the PRO in reaching a determination. 11. As part of a case (American Medical Association v. Bowen) settled 3 years ago, the OIG committed itself to seek a regulatory alternative to the practice of newspaper notices. It was intended to permit sar~;ctioned physicians to inform their Medicare patients personally that Medicare would no longer pay for the physicians' services (e.g., by mailing notices to individual patients). 12. Payment cannot be made to another party where services or items have been ordered by an entity excluded by sanction when that order was necessary as a precondition of payment. Payment may be- made for services provided up to 30 days after the effective date of the exclusion (1) when the recipient of the services or items is an inpatient in a hospital or SNF and was admitted before the effective date or (2) when home health care services are delivered under a plan established before the exclusion. 13. The relevant passages defying peer review include: ". . . whenever possible, the contractor [i.e., the PRO] [shall] use physician reviewers who practice in a setting similar to that in which the physician whose services are under review prac- tices. In . . . initial review of psychiatric and physical rehabilitation services, the contractor must arrange for review (to the extent possible) by a physician who is trained in the appropriate discipline.... In addition, the contractor would be re- quired to use board certified or board eligible physicians or dentists, in the appropri- ate specialty, to make reconsideration determinations, wherever practicable.... The contractor would consult with other health care practitioners (e.g., podiatrists) when appropriate (e.g., when reviewing services rendered by that type of practitioner)" (HCFA, 1988~. The qualifiers ("whenever possible," "to the extent possible," "wherever practicable") appear to give the opportunity to dilute the emphasis on peer review, although HCFA argues that the term "whenever possible" is used for cases where it is not administratively feasible for the PRO to use a specialist. Examples of such situations include (1) when the specialist is located on the other side of the state or (2) when the specialty is so unique that it is not possible to locate a specialist reviewer in the field. 14. Consistently throughout this study, respondents at site visits confirmed the observation that quality problems tend to affect all patients, not just the elderly. Especially in hospitals and prepaid systems, it was noted that most quality problems tended to be with "systems" that cut across units and patient age groups. Moreover, facilities and groups with well-established internal quality assurance systems delib- erately did not single out "the elderly" or "Medicare patients" for specific quality assurance attention (except insofar as they needed to meet PRO demands for records and similar requirements), believing that a more efficient and ultimately more suc- cessful approach to quality improvement would involve the entire institution, its entire staff, and its enure patient census. 15. Regulations define "confidential information" as information that explicitly or implicitly identifies an individual patient, practitioner, or reviewer; sanction re- ports and recommendations; quality review studies that identify patients, practitio- ners, or institutions; and/or PRO deliberations. "Implicitly identifies" means data

432 KATHLEEN N. LOHR AND ALLISON J. WALKER sufficiently unique or numbers so small that identification of an practitioner, or reviewer would be obvious. 16. Estimates for Medicare outlays are from the Committee on Ways and Means (1989, Table 15, p. 152), where slightly different figures are given depending upon whether the estimates are from the Congressional Budget Office (CBO) or from HCFA (CBO estimates tend to be higher). The figures cited here are roughly midway between the CBO and HCFA numbers. Estimates for PRO budgets are from HCFA staff (lI. Brook, personal communication, 1989~. 17. Material for the Medicare risk-contract section is based in part on a paper prepared for this study, "PRO Review of Medicare Health Maintenance Organiza- tions and Competitive Medical Plans," by Margaret E. O'Kane, Director of Quality Assurance, Group Health Association, Washington, D.C., May 1989. 18. The PRO-HMO Task Force had three representatives each from three organi- zations, the Group Health Association of America, the American Medical Care Re- view Association, and the American Medical Peer Review Association; the first two are prepaid group practice trade organizations, and the latter is the PRO national . . association. 19. The 15 sentinel conditions for ambulatory care review were diabetic coma or acidosis, ruptured appendix, stroke with hypertension, bleeding or perforated ulcer, gangrene, breast cancer, cancer of cervix, drug overdose/toxicities, mar-union of fracture of hip, cellulitis, bowel obstruction, bleeding secondary to anticoagulation, hypokalemia, septicemia, and pulmonary emboli. 20. The "comparable review" language for risk contracts was interpreted to mean that the number of cases reviewed must be at the same level as was occurring under PPS in the FFS system; this in turn implied a substantial level of record review. The legislation did not provide for pilot projects or staged implementation. 21. The 25 states for QRO review of HMO and CMP services were Alabama, Arizona, California, Georgia, Hawaii, Illinois, Indiana, Iowa, Kansas, Kentucky, Maryland, Massachusetts, Missouri, New Jersey, New Mexico, North Carolina, Penn- sylvania, South Carolina, South Dakota, Tennessee, Texas, Virginia, Washington, West Virginia, and Wisconsin. 22. As of December 1989, Quality Quest is the QRO only for Missouri. The Illinois PRO (Crescent Counties Medical Foundation) is the QRO for Illinois. The QRO contract for Kansas had not yet been awarded. 23. Each month, the PRO (or the QRO) receives a copy of the "Monthly Report of Medicare HMO/CMP Contracts and Applications." This list is reconciled against a list of established risk-based HMOs and CMPs to identify new plans in the area. For each new plan that requests "limited review," the PRO (QRO) performs an initial analysis to determine whether this level is appropriate. Absent such a request, the plan is automatically placed on basic review. 24. The SOW for QRO review is sufficiently similar to that for PROs that it is not described more fully here. 25. An additional requirement for HMO and CMP review, 100 percent of hospi- tal transfers, was eliminated in July 1989. 26. The 13 "sentinel" conditions for HMO and CMP review include: diabetic complications (ketoacidosis, hyperosmolar coma, other coma, and hypoglycemic coma); acute appendicitis with generalized peritonitis or peritoneal abscess; hy individual patient,

UTILlZATION^D QUME CONTROL PRO PROGRAM 433 pertensive problems (several categories, including occlusion and stenosis of precere- bral arteries and transient cerebral ischemia); gastrointestinal (GI) catastrophes (acute, chronic, or unspecified gastric ulcer with hemorrhage without obstruction; chronic duodenal ulcer with hemorrhage without obstruction; unspecified intestinal obstruc- tion); gangrene of the extremity; operations for breast malignancy (from other bi- opsy through unilateral radical mastectomy); malignant neoplasm of the genitouri- nary organ; adverse drug reactions (several categories, mainly poisoning by specific pharmacologic agents); other cellulitis and abscess; malignant neoplasm of colon; hypokalemia; septicemia; and pulmonary embolus. These are very similar to the 15 conditions proposed by the PRO-HMO Ad Hoc Committee; mar-union of fracture of hip is missing and bowel obstruction was evidently subsumed under GI catastro- phes. For limited review, one of the four conditions reviewed must be the hyperten- sive problems. In addition, two of the four must be selected from among the following conditions: diabetic complications, gangrene, adverse drug reactions, other cellulitis and abscess, hypokalemia, septicemia, or pulmonary embolus. 27. In 1988, HCFA proposed an Effectiveness Initiative intended to bring the resources of Medicare to bear on the question of what works in the practice of medicine (Roper et al., 1988; IOM, 1989~. Information generated from that initia- tive was expected to provide an expanded basis of knowledge for the development of practice guidelines and similar work, even if HCFA was not directly involved in that development. The UCDS was seen as an integral part of the initiative. At HCFA's request, the IOM convened one workshop of clinicians to advise on appro- priate clinical conditions most appropriate for priority attention in such research; that study recommended five conditions: stable and unstable angina, acute myocar- dial infarction (AMI), breast cancer, congestive heart failure, and hip fracture. HCFA then asked the IOM to convene three "research strategies" workshops on breast cancer, AMI, and hip fracture to advise on key patient management issues that should be addressed first; these workshops were conducted between February and July 1989 (IOM, 1990a-d). Because of greatly heightened interest in effectiveness and outcomes research on the part of Congress and DHHS, considerable attention was placed in mid-1989 on exactly where in the department resources for this work should be placed. OBRA 1989 (P.L. 101-239) created the Agency for Health Care Policy and Research in the Public Health Service to conduct such work, and relatively little would be done in HCFA. Consequently, the fate of the Effectiveness Initiative as envisioned by HCFA was unclear as of this writing, although it was expected that the PRO pilot projects relating to practice guidelines could proceed independently and that the UCDS might be a source of clinical data. 28. Figures cited in the text concerning the rates of quality problems found by the SuperPRO are from the oral testimony of Christy Moynihan, Ph.D., of SysteMet- rics, Inc., at the October 21, 1988, public hearing of the IOM Study to Design a Strategy for Quality Review and Assurance in the Medicare Program. 29. PROs sometimes negotiate review objectives with HCFA on the basis of their experience with generic screens, with the idea of specifying a target problem reduction level. Discharge planning is the most popular screen on which such objectives are based, and it plus hospital-based trauma are areas in which PROs are most successful in reaching or exceeding their targets (GAO, 1988b). Discharge

434 KATHLEE7VN. LOHR AND AIllSON J. WALKER planning and falls (as part of hospital-based trauma) are the two screens for which nurse reviewers need not refer cases to physician advisors. 30. In its cumulative data summary, HCFA (1989b) notes that"totals for all reviews for the generic quality screen data should not be used to project national rates of occurrence" of the various problems identified through the screens. The reason is that national totals include problems detected through focused review, which (if it is being conducted as intended) should detect more than the usual number of quality problems. Hence, the total rates of occurrence, if projected to all discharges without accounting for the higher rates among focused review, might yield rates higher than are actually the case. 31. With respect to additional screens developed by the PROs, the Peer Review Organization of Washington (PRO/W) reported that some of its specially developed screens identify more [allures and/or confirmed problems than do the HCFA screens. 32 An ancillary proposal is that the OIG be given the authority of a "national medical board" to protect beneficiaries. States have not been very effective in exerting discipline over the medical profession, and some suggest that such a na- tional system would permit uniformity and greater protection for Medicare benefici- aries (Jost, 1989~. 33. The OIG had advised PROs in July 1987 not to recommend fines unless they could demonstrate that they would be cost-effective. Apart from perhaps contradict- ing legal requirements that PROs bring sanction recommendations against all pro- viders that have violated their Medicare obligations, this advisory discouraged PROs from recommending monetary penalties and decreased the number of sanction fil- ings. This situation was to have been remedied by Spring of 1989 by a change in policy from the OIG to the PROs (McIlrath, 1989; Vibbert, 1989a). 34. Smith and Mehlman (1989) provided the basis for much of the discussion in this entire section. 35. The site visits for this study-took place during the time that the proposed rule was published for public comment, and thus PROs were just beginning to analyze how it would affect their local operations. A major problem mentioned to the site visitors was that implementation of quality denials as planned, particularly the alleg- edly harsh language required for the letters to beneficiaries, would seriously erode the ability of PROs to attract the necessary specialist physician advisors who would be expected to make the initial denial decisions. Moreover, PROs believe that the relatively low rates they can pay physician advisors, often only between $55 and $65 an hour, already markedly hampers their recruitment efforts, especially of cer- tain kinds of specialists and subspecialists; implementation of the quality denials, with the heavy reliance on specialists as described, was expected to exacerbate the problem. 36. Early in the PRO program the American Hospital Association (AMA) charged that HCFA's administrative practices violated the APA and requested that DHHS promulgate comprehensive PRO regulations (i.e., in accordance with APA require- ments). This request was not met, and the AHA filed a lawsuit (~American Hospital Association v. Bowen) claiming that DHHS had failed to meet its statutory duties under the APA. The district court for the District of Columbia agreed with AHA. DHHS appealed the decision, and the court of appeals reversed the district court ruling. The majority opinion held that the contracting process, the issuance of the

UTIllZATION AND QUALlIYCONTROLPRO PROGRAM 435 SOWs, and Manual transmittals were all covered by exceptions to the APA (Smith and Mehlman, 1989~. 37. Vibbert (1989c) reported that some legislative language may require that DHHS make HMOs, not hospitals, responsible for assuring that PROs obtain the inpatient discharge abstract information needed to draw the relevant review samples. The rationale was that the current random analysis approach is ineffective and intru- sive (a point with which HMOs might well agree); nevertheless, this option would not appear, from the point of view of the HMO industry, to overcome that particular problem. 38. The issue of reimbursing hospitals for costs incurred in photocopying medi- cal records has been especially contentious since 1985. A court order in mid-1989 required that HCFA reimburse hospitals retroactively for costs incurred in photo- copying; the American Hospital Association argues in favor of a $0.12 per page reimbursement level. Recently hospitals gained class action status in a federal suit involving photocopying costs (Vibbert, 1989g). 39. No. WD39809 (Mo. Ct. App. 19891. Lee Missouri Court of Appeals found that the doctrine of "corporate negligence" is not limited to treatment settings be- cause it is based on common law principles of negligence. Moreover, it determined that the defendant (the HMO) had a duty of care to protect members from unreason- able risk of harm by exposing them to unqualified or incompetent physicians (and the HMO had not done so because it had failed to make any investigation of the background of the physician in question). REFERENCES AHA v. Bowen, 640 F. Supp. 453 (D.D.C. 1986~. AMA v. Bowen, Civil Action No. 87-995 (D.D.C. 1987~. Committee on Ways and Means. Background Material on Programs Within the Jurisdiction of the Committee on Ways and Means. 1.989 edition. Committee Print WMCP 101-4, March 15, 1989. Washington, D.C.: Government Print- ing Office, 1989. Dans, P.E., Weiner, J.P., and Otter, S.E. Peer Review Organizations: Promises and Pitfalls. New England Journal of Medicine 313:1131-1137, 1985. Federal Register, Vol. 50, pp. 15364-15389, April 17, 1985. Federal Register, Vol. 54, pp. 195~1967, January 18, 1989. GAO (General Accounting Office). Medicare Improving Quality of Care Assess- ment arm Assurance. PEMD-88-10. Washington, D.C.: General Accounting Office, May 1988a. GAO. Medicare PROs Extreme Variation in Organizational Structure and Activi- ties. PEMD-89-7FS. Washington, D.C.: General Accounting Office, FJovem- ber 19B8b. HCFA (Health Care Financing Administration). 1988-1990 PRO Scope of Work. Baltimore, Md.: Health Care Financing Administration, Department of Health and Human Services, 1988. HCFA. Request for Proposal for the SuperPRO Contract, Attachment lI. Baltimore, Md.: Health Care Financing Administration, Department of Health and Hu- man Services, 1 989a.

436 KATHLEEN N. LOHR AND ALLISON J. WALX:ER HCFA. Peer Review Organization Data Summaries dated February, May, and Au- gust 1989, and Technical Notes to the Data Summaries. Baltimore, Md.: Office of Peer Review, Health Standards and Quality Bureau, Health Care Financing Administration, 1989b. HCFA. Utilization and Quality Control Peer Review Organizations Second Scope of Work. Executive Data Summary. Report through March 1989. Report dated July 5, 1989. Baltimore, Md.: Health Care Financing Administration, 1989c. Hsia, D.C., Krushat, W.M., Pagan, A.B., et al. Accuracy of Diagnostic Coding for Medicare Patients Under the Prospective-Payment System. New England Journal of Medicine 318:352-355, 1988. Hsiao, W.C., Braun, P., Yntema, D., et al. Estimating Physicians' Work for a Resource-Based Relative Value Scale. New England Journal of Medicine 319:835-841, 1988. IOM (Institute of Medicine). Effectiveness Initiative: Setting Priorities for Clinical Conditions. Washington, D.C.: National Academy Press, 1989. IOM. Effectiveness initiative: Setting Priorities for Research in Breast Cancer. Washington, D.C.: National Academy Press, 1990a. IOM. Effectiveness Initiative: Setting Priorities for Research in Acute Myocardial Infarction. Washington, D.C.: National Academy Press, 1990b, forthcoming. IOM. Effectiveness Initiative: Setting Priorities for Research in Hip Fracture. Washington, D.C.: National Academy Press, 1990c, forthcoming. IOM. Effectiveness and Outcome Research: Proceedings of a Conference. Wash- ington, D.C.: National Academy Press, 1990d, forthcoming. Jost, T. Administrative Law Issues Involving the Medicare Utilization and Quality Control Peer Review Organization (PRO) Programs: Analysis and Recom- mendations. Report to the Administratrive Conference of the United States. Washington, D.C.: Administrative Conference, November 8, 1988 (Reprinted in Ohio State Law Journal 50~1), 1989~. Kane, R.A., Kane, R.L., Kleffel, D., et al. The PSRO and the Nursing Home: Vol. I, An Assessment of PSRO Long-Term Care Review. R-2459/1-HCFA. Santa Monica, Calif.: The RAND Corporation, August 1979. Lavapies v. Bowen, 687 F. Supp. 1193 (S.D. Ohio 1988~. Lohr, K.N. Peer Review Organizations: Quality Assurance in Medicare. P-7125. Santa Monica, Calif.: The RAND Corporation, July 1985. McIlrath, S. Receding Tide of Physician Sanctions by Medicare PROs Triggers Debate. American Medical News 3: 57, April 21, 1989. Morford, T.G., Director, Health Standards and Quality Bureau, HCFA. Testimony before the U.S. House of Representatives Committee on Government Opera- tions, Subcommittee on Human Resources and Intergovernmental Relations, April 4, 1989a. Morford, T.G. UpDate. Federal Efforts to Improve Peer Review Organizations. Health Affairs 8~2~:175-178, Summer 1989b. OIG (Office of Inspector General, Department of Health and Human Services). The Utilization and (2uality Control Peer Review Organization (PRO) Program. Quality Review Activities. Washington, D.C.: Office of Inspector General, Office of Analysis and Inspections, August 1988a.

UTILIZATION AND QUAL17YCONTROLPRO PROGRAM OIG. ~~ ~~ ~~ 437 l he Utilization and Quality Control Peer Review Organization (PRO) Pro- gram. Sanction Activities. Washington, D.C.: Office of Inspector General, Office of Analysis and Inspections, October 1988b. OIG. The Utilization and Quality Control Peer Review Organization {PRO) Pro- gram. An Exploration of Program Effectiveness. ~ ~. ~ ~ ~ O - · ~ ~ _ ~ ~ . Washington, D.C.: Office or Inspector General, Truce or Analysis and inspections, January 1989. O'Kane, M.E. PRO Review of Medicare Health Maintenance Organizations and Competitive Medical Plans. Paper prepared for the Institute of Medicine Study to Design a Strategy for Quality Review and Assurance in Medicare, 1989. OTA (Office of Technology Assessment). The Quality of Medical Care. Informa- tion for Consumers. OTA-H-386. Washington, D.C.: Government Printing Office, June 1988. PPRC (Physician Payment Review Commission). Annual Report to Congress, 1988. Washington, D.C.: Physician Payment Review Commission, March 1988. PPRC. Annual Report to Congress, 1989. Washington, D.C.: Physician Payment Review Commission, April 1989. ProPAC (Prospective Payment Assessment Commission). Medicare Prospective Payment and the American Health Care System. Report to the Congress, June 1989. Washington, D.C.: Prospective Payment Assessment Cornmis- sion, 1989. Roper, W.L., Winkenwerder, W., Hackbarth, G.M., et al. Effectiveness in Health Care: An Initiative to Evaluate and Improve Medical Practice. New England Journal of Medicine 319:1 197-1202, 1988. Simborg, D.W. DRG Creep. A New Hospital Acquired Disease. New England Journal of Medicine 304:1602-1604, 1981. Smith, A.H. and Mehlman, M.J. Medicare Quality Assurance Mechanisms and the Law. Paper prepared for the Institute of Medicine Study to Design a Strategy for Quality Review and Assurance in Medicare, 1989. Vibbert, S., ed. Watchdog Criticizes HHS IG Over PRO Monetary Sanctions. Medi- cal Utilization Review 17~7~: 1, April 4, 1989a Vibbert, S., ed. PROs Denied 2 Percent of 1986-1988 Medicare Cases. Medical Utilization Review 17~9~: 1, May 2, 1989b. Vibbert, S., ed. Regulatory Activity. Legal Panel Backs PRO Reform Package. Medical Utilization Review 17~13~:5-6, June 27, 1989c. Vibbert, S., ed. PRO Sanctions Fight Moves to Senate. Medical Utilization Review 17 (15~:1-2, August 8, 1989d. Vibbert, S., ed. HSQB Softpedals Controversial Pilot. Medical Utilization Review 17 (18~:3 - , September 21, 1989e. Vibbert, S., ed. SysteMetrics Retains SuperPRO Contract. Medical Utilization Review 17 (19~:2-3, October 5, 1989f. Vibbert, S., ed. Hospitals Gain Class Action Status in Federal Suit Over PRO Photocopying. Medical Utilization Review 17~21~:1, November 2, 1989g.