3

How Are Healthcare Systems Monitoring Quality Today?

A wide gulf exits between the ideal data system just described and the reality of cancer care quality monitoring today. Although the United States has no national comprehensive quality monitoring system, there is a patchwork of private and federal efforts to assess cancer care quality. Each initiative operates with different purposes, perspectives, and audiences. Many of the quality (and cost) monitoring activities are organized within hospitals or provider groups, usually in an effort to demonstrate value to the insurers and managed care organizations purchasing their services. Other quality monitoring activities are externally driven and have an accountability function—the government may want to ensure that publicly funded healthcare programs are adhering to best practices, or professional societies may want to demonstrate to the public that their care meets or exceeds accepted standards of care. This chapter first illustrates with a series of case studies the diversity of approaches to cancer care quality monitoring taken by selected individual providers, hospitals, health plans, provider groups, physician practice management companies, insurers, and purchasers. The chapter concludes with a discussion of the strengths and weaknesses of current approaches.

CASE STUDIES—ILLUSTRATIONS OF THE USE OF DATA TO MONITOR CANCER CARE QUALITY

In this section of the report, 10 case studies are presented to show how data are being used within various systems of care to provide information on the quality of cancer care (Table 3.1). The examples range from a single private



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care 3 How Are Healthcare Systems Monitoring Quality Today? A wide gulf exits between the ideal data system just described and the reality of cancer care quality monitoring today. Although the United States has no national comprehensive quality monitoring system, there is a patchwork of private and federal efforts to assess cancer care quality. Each initiative operates with different purposes, perspectives, and audiences. Many of the quality (and cost) monitoring activities are organized within hospitals or provider groups, usually in an effort to demonstrate value to the insurers and managed care organizations purchasing their services. Other quality monitoring activities are externally driven and have an accountability function—the government may want to ensure that publicly funded healthcare programs are adhering to best practices, or professional societies may want to demonstrate to the public that their care meets or exceeds accepted standards of care. This chapter first illustrates with a series of case studies the diversity of approaches to cancer care quality monitoring taken by selected individual providers, hospitals, health plans, provider groups, physician practice management companies, insurers, and purchasers. The chapter concludes with a discussion of the strengths and weaknesses of current approaches. CASE STUDIES—ILLUSTRATIONS OF THE USE OF DATA TO MONITOR CANCER CARE QUALITY In this section of the report, 10 case studies are presented to show how data are being used within various systems of care to provide information on the quality of cancer care (Table 3.1). The examples range from a single private

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care TABLE 3.1 Illustrative Case Studies of Using Cancer Care Data for Quality Monitoring Purposes Name (type of organization) Purpose Data Source(s) 1. Marin Oncology Associates (private oncologypractice) Monitor adherence to guidelines regarding screening, treatment, follow-up, supportive, and end-of-life care Medical chart 2. OnCare (physician practice management company [PPMC]) Monitor adherence to guidelines regarding treatment, follow-up, and end-of-life care Electronic medical chart 3. American College of Radiology Monitor patterns of care and adherence to treatment guidelines Medical chart abstraction from a national sample of radiation oncology providers 4. Sutter Health (integrated healthcare delivery system) Monitor adherence to breast cancer treatment guidelines Hospital cancer registry, administrative data, medical charts, patient surveys 5. Providence Health Plan (integrated delivery system) Monitor adherence to breast cancer treatment guidelines Hospital cancer registry, administrative data, medical charts, patient surveys 6. National Comprehensive Cancer Network (17 large cancer centers) Monitor adherence to breast cancer treatment guidelines Medical charts; reporting according to a uniform data set 7. Roswell Park Cancer Institute and private insurers in western New York Monitor adherence to breast cancer treatment guidelines Insurance claims, medical charts 8. Colorado Cancer Registry, University of Colorado, and State Medicare Peer Review Organization Monitor use of adjuvant therapies for breast and colorectal cancer State cancer registry, Medicare claims, medical charts 9. Central Florida Health Care Coalition (business coalition) Monitor quality of care for individuals with selected conditions including cancer Insurance claims, (hospital and outpatient), patient survey 10. National Cancer Data Base (database maintained by the American College of Surgeons and the American Cancer Society) Monitor quality of care for individuals with cancer Hospital cancer registries, cancer centers with uniform reporting requirements

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care medical office, to large integrated delivery systems, to states. These case studies are not meant to be truly representative of all cancer-related quality improvement programs, but instead were drawn to illustrate the variety of ongoing approaches to using data. The case studies were not identified in any systematic way. Most were identified through contacts with cancer-related health services researchers and IOM workshop participants. Others were identified through descriptions in publications (e.g., the President 's Cancer Panel) or presentations at professional meetings (e.g., American Society of Clinical Oncology). Launching a quality assurance program using only internal resources is unusual for a small group of private practitioners, but Case Study 1 (see box) provides an example of one such initiative. Barriers to quality monitoring identified by members of this small practice included high costs, limited staff resources, a lack of incentives, an absence of an accepted set of quality measures, and a lack of benchmarks or standards with which to gauge success (P. Eisenberg, physician, Marin Oncology Associates, Inc., personal communication, October 18, 1999). According to the practice physicians, the program has been effective in aligning the practice with accepted practice guidelines and in demonstrating the value of the group practice to managed care organizations and insurers (P. Eisenberg, personal communication, October 18, 1999). Quality assessment is expensive and can usually be accomplished more easily when development costs are spread across groups of providers. Increasingly, provider practice management companies have formed to provide administrative functions for their members, for example, billing and claims processing (Mighion et al., 1999). Case Study 2 is an example of one such company with a unique focus on quality. Its members are small, community-based oncology practices located throughout the country. The program is especially notable because of its use of an electronic medical record system with embedded guidelines, available to providers at the point of contact with patients. Such support systems for clinical decision making can significantly improve the quality of patient care (Classen, 1998; Hunt et al., 1998). Other physician practice management companies and disease management companies are developing electronic medical record systems, and most have a central shared database of clinical, patient, financial, and administrative information. About one-quarter of OnCare patients have a diagnosis of breast cancer, and, since 1997, OnCare has attempted to standardize the approach to adjuvant treatment to improve outcomes and reduce cost. In general, adherence to the guidelines among OnCare physicians was already good, but giving physicians information on their own performance relative to average practice improved the performance of a few “outlier” physicians. More deviation from guidelines has been found for colon cancer (e.g., overuse of adjuvant therapy for Stage I cancer). In addition to changing practice patterns, the availability of the electronic medical record has improved the recording of stage, comorbid illnesses, medication, prior treatment, and other prognostic indicators (K. Bergstrom, vice president, disease management, OnCare, personal communication, October 26, 1999).

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Case Study 1: Marin Oncology Associates—A Community-Based Private Oncology Practice Marin Oncology is a San Francisco Bay Area, community-based private practice that has monitored the quality of care provided by its three physicians. The practice has attempted to achieve a patient-centered, evidence-based, and cost-effective practice by establishing standards and a monitoring system. The physicians identified cancer care measures from guidelines, the medical literature, and specialty societies (e.g., American College of Physicians, American Society of Clinical Oncology). Several measures were deemed appropriate to monitor; however, specific benchmarks or targets to gauge success could not be found. The measures finally adopted by the practice address the efficacy of treatment (e.g., survival), processes of care (e.g., use of second-line chemotherapy), and supportive care (e.g., pain management, end-of-life care). Breast cancer Proportion of early breast cancer patients receiving mammogram within 12 months of their primary diagnosis. American Society of Clinical Oncology Breast cancer follow-up guidelines (e.g., use of routine blood tests, bone scans). Non-small-cell lung cancer Proportion of patients with widespread disease (Stage IV) receiving first-, second- and third-line chemotherapy. 1- and 2-year survival. Chemotherapy Use of second-line chemotherapy for which there is no evidence of therapeutic benefit. Use of colony stimulating factors during chemotherapy. Pain control Hospitalization rates. End-of-life care Hospice use, length of stay. Site of death (home, hospital, Extended Care Facility). Interval from last chemotherapy to death. Use of form documenting patient preferences for care (e.g., advanced directives, do not resuscitate orders). Data are retrieved through chart audit and are summarized on Excel spreadsheets. Much of the work is done by the clinicians and their staff, but some analytic support has come from Public Health residents from Loma Linda University. SOURCE: P. Eisenberg, physician, Marin Oncology Associates, Inc.,personal communication, October 18, 1999.

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Case Study 2: OnCare—A Physician Practice Management Company OnCare is a privately held, for-profit, physician practice management company with a focus on guideline-driven quality improvement. The company owns and manages medical and radiation oncology offices and clinics in 11 states. The 28 physician practices that are affiliated with OnCare have replaced paper medical records with an electronic charting system that includes embedded on-line decision support for all aspects of care (e.g., choice of initial therapy, dosage of chemotherapy, supportive care). Information about patient care is captured, analyzed, and fed back to the system's more than 100 physicians to assess their compliance to the guidelines. To date, the system includes information on more than 15,000 patients. OnCare clinicians enter patient data in a flow chart format as the patient is receiving care. Clinical options are displayed allowing providers to consider guideline recommendations as decisions about care are being made (e.g., selection of chemotherapy). Furthermore, providers can examine the processes of care and outcomes of patients within the OnCare system who have characteristics similar to the patient being evaluated. The OnCare electronic medical chart is called KnowChart® and was developed by a software vendor, KnowMed Systems. The electronic medical record is integrated with laboratory systems, patient scheduling, and billing. OnCare guidelines have been developed for most cancer sites/types and are integrated into the information system. The guidelines were developed according to a review of the literature, advice of experts, and input from local OnCare providers. If providers wish to deviate from the guidelines, they must document their rationale for doing so. Reasons for changing chemotherapy doses are also documented in the system. OnCare provides its practices with information on clinical and financial performance, as well as marketing support. OnCare is expanding. It has, for example, entered into a partnership with the not-for-profit M.D. Anderson Cancer Network to develop and market a managed-care plan SOURCES: G. Swanson, physician, OnCare, personal communication, October26, 1999; K. Bergstrom, vice president, disease management, OnCare,personal communication, October 26, 1999; R. Shiffman, physician,OnCare, personal communication, October 26, 1999; www.oncare.com; www.knowmed.com. Adherence to other cancer care guidelines has also been assessed using On Care's clinical information system. In 1996, for example, OnCare providers assessed their compliance to the 1994 American Society of Clinical Oncology's guidelines on the use of hematopoietic growth factors. Information relevant to end-of-life care is being tracked to assess the success of a physician training program in end-of-life care (K. Bergstrom, vice president, disease management, OnCare, personal communication, October 26, 1999). Case Study 3 provides another example of the centralization of quality improvement activities for a particular group of providers. The American College

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Case Study 3: Quality Monitoring of Radiation Oncology Care, American College of Radiology ACR has, since 1973, monitored the quality of U.S. radiation oncology care with support from NCI. ACR's “Patterns of Care Study” (PCS) process begins with the development of a consensus guideline that summarizes what is determined to be the “best current management” for a particular cancer. Panels are convened to review available evidence and develop decision trees for patient management. In developing the guideline, panelists may assess treatment modality, type of equipment, dose ranges, and treatment areas for radiation therapy. If there is disagreement among the panel members, a formal process is used to reach consensus (i.e., modified Delphi process). Panels also develop questions for a national survey of providers to assess how care is actually being delivered to patients and whether there is deviation from the guidelines. The next step of the PCS process is obtaining data on processes of care, and sometimes outcomes of care through a nationwide survey of radiation oncology providers. Independent surveyors visit selected radiation facilities, abstract information from a sample of their medical charts, and enter the data into laptop computers. Typically, 75 facilities and 10 cases per facility are represented for each cancer-specific study. Information on outcomes may be collected at the same time that processes of care data are gathered, or in the case of 5-year survival, years following the original survey. ACR staff researchers aggregate and analyze the survey data to ascertain patterns of care, the consistency of practice to guidelines, and the nature of interactions of structures of care, processes of care, and outcomes. The ability to generalize survey findings to the nation is possible because samples of radiation oncology providers are carefully drawn from a comprehensive list of facilities maintained by ACR. All radiation oncology departments in the United States are included in the listing, along with information about resources at each site (e.g., size, equipment, personnel). For the PCS survey, a sample of facilities is selected to ensure that all sizes and types of facilities are appropriately represented; then a sample of patients from each of these facilities is randomly selected from those eligible for the study. This process minimizes bias that could occur if facilities were to select their own cases for inclusion. Sometimes, certain facilities are sampled at higher rates so that statistically valid subgroup comparisons can be made (e.g., for studies of care of minority populations). A total of 29 PCS studies have been conducted to date (see Table 3.2), providing insights into the quality of U.S. radiation oncology care. SOURCES: J. Owen, director, Patterns of Care Study, American Collegeof Radiology, personal communication, November 12, 1999; Coia LR,Owen JB, Hanks GE, 1997, “Introduction” Seminars in Radiation Oncology 7(2):95–96; Hanks GE, Coia LR, Curry J, 1997, “Patterns of Care Studies: Past, Present, and Future,” Seminars in Radiation Oncology 7(2):97–100.

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care TABLE 3.2 Patterns of Care Studies, 1973–1999 Disease Site Years of Treatment Cervix 1973, 1978, 1983, 1988–1989, 1992–1994 Prostate 1973–1975, 1978, 1983, 1989, 1994 Hodgkin's disease 1973, 1983, 1988–1989 Breast 1973, 1983, 1989, 1993–1994 Larynx 1973–1975, 1978 Bladder 1973–1975 Corpus uteri 1973 Nasopharynx 1973–1975 Testicular seminoma 1973, 1992–1994 Tongue 1973–1975 Palliation 1983 Rectum 1988–1989, 1992–1994 Esophagus 1992–1994 Tonsil 1976–1985 SOURCE: J. Owen, director, Patterns of Care Study, American Collegeof Radiology, personal communication, November 12, 1999. of Radiology (ACR) has developed a system to monitor the quality of radiation oncology care throughout the United States. Since 1973, ACR has identified best practices through guideline development, then conducted extensive targeted surveys of practitioners to evaluate adherence to the guidelines. The ACR effort is laudable because of its adherence to good sampling and data collection techniques, allowing it to provide accurate information on the quality of radiation oncology care for the nation. The program was not designed to provide information about quality to individual practices or institutions. Instead, quality improvement is promoted through ACR's educational activities (e.g., dissemination of patterns of care information in journals and professional meetings) and standards and practice accreditation programs. The program has long-standing financial support from the National Cancer Institute (NCI). In 1999, ACR received over $700,000 to support their quality program (F. Mahoney, Grants Office, NCI, personal communication, November 30, 1999). Quality systems described thus far have been limited to the care of certain providers, either oncologists or radiation oncologists. Cancer care is multidisciplinary and, within larger systems of care, quality improvement programs need to address the range of services needed by individuals with cancer. Case Study 4 illustrates how such cross-cutting quality measures can be applied within a large integrated system of hospitals and medical groups. Sutter Health has developed a quality improvement program for breast cancer and has put in place an information system to record compliance to quality algorithms. Sources of quality

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care information include the hospital cancer registry, administrative and clinical records, and patient surveys. With 2 years of data collected thus far, there is some evidence of success—there has been improvement for six of the eight measures. Significant variation remains, however, among the system's hospitals and medical groups (Katterhagen, 1999). Case Study 5 illustrates how Providence Health System, a large integrated delivery system, has adopted measures originally developed by the Foundation Case Study 4: Sutter Health Breast Cancer Quality Project Sutter Health is an integrated system of 23 acute care hospitals and 8 allied medical groups in Northern California. Since 1997, Sutter Health has monitored the quality and costs of breast cancer care throughout its system to improve clinical outcomes, reduce costs, and increase market share for hospitals, medical groups, and physicians. The quality monitoring system was initially developed in 1994 in selected facilities within the system (i.e., the Mills-Peninsula Health Services). Data from the hospital cancer registry, State cancer registry, hospital administrative systems, breast center clinical systems, and patient surveys are used to track the following measures (targets are for 1999): size or stage at diagnosis (with appropriate levels of screening, should see early stage disease, for example, ductal carcinoma in situ [DCIS] rate should exceed 21%); needle biopsy rates (should be high relative to surgical biopsy, exceeding 50%); axillary dissection rates for DCIS (should be 2% or lower); surgical breast conservation rates (should exceed 65% for early stage disease); radiation therapy as a component of surgical conservation (should exceed 85%, but be reduced for selected patient groups); adjuvant chemotherapy for Stage II, node positive cases (should exceed 95%); patient satisfaction (no targets established); and patient quality of life (use the Functional Assessment of Cancer Therapy-Breast quality-of-life instrument [FACT-B]) (no targets established). The performance of individual hospitals and medical groups relative to practice algorithms is reported by name on a quarterly basis, along with group norms and targets, and is disseminated at tumor board and departmental meetings, via newsletters, and for “outliers ” on an individual basis. Local senior management and medical directors are held accountable for performance. SOURCE: G. Katterhagen, personal communication, October 4, 1999.

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care for Accountability (FACCT) for breast cancer. These measures are quite comprehensive and include indicators of the processes of care, patient satisfaction, and outcomes (Table 3.3). The case studies described so far have relied almost exclusively on clinical data recorded in the chart, data available from the hospital cancer registry, and hospital administrative records. Case Study 6 illustrates an alternative mechanism, the development of a very expensive clinical information system. Seventeen of the nation's premier cancer centers have joined together to form the National Comprehensive Cancer Network (NCCN) in an effort to measure and monitor the quality of care within the member institutions (Figure 3.1). In collaboration with one of the NCCN member institutions, Roswell Park Cancer Institute, a coalition of managed care organizations and insurers has banded together to assess the quality of care for the entire insured population of western New York State. This effort is described in Case Study 7. From a data perspective, the approach taken is quite unique. Cases are first identified through insurance claims; then medical records are abstracted to obtain information on stage of illness. Analyses of the supplemented claims data provide information on the quality of breast cancer care. Case Study 5: Providence Health System Providence Health System is an integrated delivery system with over 50 medical centers in Oregon, Washington, and Alaska. Since 1993, the Providence Health Plan of Oregon has monitored several clinical measures on all breast cancer patients diagnosed and/or treated in the health system: stage at diagnosis, rate of breast-conserving surgery (BCS), radiation treatment after BCS, chemotherapy for node positive disease, Tamoxifen for ER positive disease, and overall survival, These clinical measures are derived from reports from the hospital cancer registry reporting system and the medical chart. A sample of patients is contacted by mail to complete a survey regarding their satisfaction and experience with care. Providers within the system are apprised of the results of monitoring through an annual cancer program report, cancer committee meetings, tumor boards, and various regional quality improvement teams. No attempt has been made to give providers information on their individual performance. Since 1993 there have been increases in the share of patients diagnosed with early stage disease, the rate of BCS for early stage disease (from about 33 to 60% from 1993 to 1999), and use of radiation following surgery (from 76 to 84% from 1994 to 1998). The chemotherapy and tamoxifen measures were just introduced in 1999, and there are as yet insufficient data to report. SOURCE: L. Skokan, senior scientist, Providence Health System, personalcommunication, November 23, 1999.

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care FIGURE 3.1 National Comprehensive Cancer Network member institutions. SOURCE: National Comprehensive Cancer Network. 1999. Member Institutions. http://www.nccn.org/network_content.htm. None of the case studies described thus far have provided quality-of-care information for all cancer patients residing within a specific geographic area. Case Study 8 comes close to achieving this by providing information on some aspects of the quality of breast and colon cancer care for the elderly residents of Colorado. This case study illustrates the potential for collaborative efforts involving cancer registries, universities, and Peer Review Organizations (PROs) to improve cancer care quality. Each state has a PRO funded by the Health Care Financing Administration (HCFA). The PROs evaluate whether care given to Medicare patients is reasonable, necessary, and provided in the most appropriate setting. The most recent contract with HCFA requires the PROs to conduct local quality improvement projects focused on six clinical prioritized areas, one of them breast cancer (Jencks, 1999). Case Study 9 provides an example of employers joining together to form a large coalition to push for uniform quality assessment among the health plans covering their employees. Purchasers may use quality information to identify high-value plans, to steer employees into higher-performing plans, or as leverage when establishing rates for premiums (Darby, 1998). The last case study represents perhaps the largest single effort to monitor cancer care in the United States. The National Cancer Data Base (NCDB) includes quality-related information on more than three-quarters of newly diagnosed cases of cancer. The effort is jointly sponsored by the American College of Surgeons' Commission on Cancer (ACoS-CoC) and the American Cancer Society.

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care TABLE 3.3 Foundation for Accountability (FACCT) Breast Cancer Quality Indicators Measure Performance Value Instrument or Data Source Steps to Good Care Mammography Proportion of women age 52–69 who have had a mammogram within 2 years Doctor's billing or claims records (NCQA's HEDIS® 3.0 breast cancer screening measure used) Early-stage detection Proportion of patients whose breast cancer was detected at Stage 0 or Stage I Patient records from cancer registry Information about radiation treatment options Proportion of Stage I and II patients who indicate that they had adequate information about their radiation treatment options before deciding about treatment One question in patient satisfaction survey completed 3–6 months after diagnosis Breast conserving surgery (BCS) Proportion of Stage I and II patients who undergo BCS Patient records from cancer registry or claims records Radiation therapy following BCS Proportion of BCS patients who receive radiation treatment after surgery Patient records from cancer registry or claims records

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Experience and Satisfaction Patient satisfaction with care Mean score for patients' level of satisfaction with breast cancer care, including the technical quality, interpersonal and communication skills of their cancer doctor, their involvement in treatment decisions, and the timeliness of receiving information and services 32-item patient satisfaction survey completed 3–6 months after diagnosis Results Experience of disease Mean score for patients on CARES-SF survey, which assesses patients ' quality of life and experience in living with breast cancer 59-item CARES-SF patient survey completed 12–15 months after diagnosis 5-year disease-free survival (cancer treatment center measure) Probability of disease-free survival for a group of patients, Stages I–IV, who were diagnosed during previous 5 years Patient records from cancer registry NOTE: CARES = Cancer Rehabilitation Evaluation System. SOURCE: Foundation for Accountability, 1998. FACCT Quality Measures—Breast Cancer. Available on line at: http://www.facct.org.

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Case Study 6: The National Comprehensive Cancer Network NCCN institutions cooperated to develop a set of clinical practice guidelines for the treatment of common cancers. Available sources of data were determined to be inadequate to monitor adherence to the guidelines, and an extensive uniform clinical data set was defined to guide data collection. Now detailed sociodemographic, clinical, treatment, outcome, and cost data are gathered on all patients with selected cancers treated in the NCCN network. Data are used to evaluate both the performance of providers relative to the guidelines and the effectiveness of treatments. The database is very extensive; for breast cancer alone, over 200 data elements are collected. Data managers from each cancer center report encrypted data via the internet to a central analytic office. Strict data security and confidentiality measures are in place (e.g., individual authentication of users, stripping personal identifiers from records), as are methods to ensure the quality of data (e.g., standardized coding schemes, audits). Performance results are posted on a Web page accessible to providers within the system. Data collection for the pilot condition—breast cancer—began in July 1997. As of October 1999, information on over 2,000 patients had been entered into the system. Preliminary results show high compliance to guidelines but some areas of practice variation. Among women treated with BCS, for example, 89% received radiation therapy overall, with a range across institutions from 80 to 100%. Plans are to expand to other cancer sites and to add information on complications and patient-reported quality of life. The NCCN would like to extend participation to selected outside institutions/practices and to establish new partnerships with pharmaceutical and biotechnology companies, insurers, and regulatory/accrediting bodies. SOURCE: Weeks, 1999. ACoS-CoC regularly conducts surveys of facilities that provide cancer treatment and approves those that meet its standards.* Approximately 1,500 programs have been approved, which are estimated to provide care for 80% of the nation's newly diagnosed patients (Morrow, 1999). Most facilities are community-based hospitals, but a few freestanding cancer programs are also included (Table 3.4) (Morrow, 1999). * The Joint Commission on Accreditation of Healthcare Organizations as of January 1999 accepts the American College of Surgeons' Commission on Cancer accreditation decisions for cancer treatment facilities or cancer hospitals that are affiliated with health plans and health systems (www.facs.org). This will effectively reduce the redundancy of site evaluations of cancer facilities.

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Case Study 7: Implementing a Regional Cancer Quality Monitoring Program Using Private Insurance Claims Data A regional total quality management program for cancer care is being developed in western New York State through a collaboration of the Roswell Park Cancer Institute, the Independent Health Association (IHA), a managed care organization, and a board of community physicians. The first phase of the project involved a retrospective quality review of breast cancer care in 1995 and 1996. Under development is the second phase, a prospective quality improvement program that will involve interventions such as: reporting of aggregate and provider-specific data to specialty physicians, distributing guidelines and aggregate data to primary providers, case management through payers, and disseminating treatment guidelines to the public. As part of phase I, IHA's breast cancer-related claims from 1995 to 1996 were analyzed to assess compliance to guidelines of the NCCN. Claims that included ICD-9 diagnostic and CPT procedure codes indicative of a diagnosis of breast cancer were collected into a Microsoft Access database. For patients whose sequence of procedures and other care was consistent with a diagnosis of breast cancer, medical records (pathology and operative reports) were retrieved by IHA to confirm the diagnosis and to determine the cancer stage. Case identification generally occurs 6 to 9 months after diagnosis allowing for timely quality evaluation. Virtually all cancer cases were identified using claims data according to results of a validation study. Quality indicators included: Diagnosis Positive rate for mammographic directed breast biopsy. Use of needle biopsy for diagnosis. Work-up Rate of use of bone scan, computerized tomography, and serum tumor markers. Surgery Rate of reexcision by biopsy type with breast conserving surgery (BCS). Rate of BCS. Rate of immediate breast reconstruction. Guideline Compliance Rate of variation from NCCN guidelines for surgery, radiation, and adjuvant systemic therapy.

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care The study demonstrated a rate of BCS higher than the national average and overall high rates of compliance with NCCN guidelines. Provider-specific data analysis showed substantial variation between physicians in the use of biopsy techniques and the rate of positive biopsies. Treatment-related quality problems identified among the 379 women with Stage 0/I/II breast cancer identified in 1995–1996 included: 13% of women (5 of 38 women) with DCIS treated with BCS had axillary lymph node dissection (not indicated with DCIS), 10% of women (18 of 188 women) did not receive radiation therapy following breast conserving therapy when this was indicated, 8% of women (8 of 105 women) were treated with radiation therapy following mastectomy when the treatment was not indicated (i.e., they had fewer than 4 positive nodes and tumors less than 5 cm), and 68% of women (19 of 28 women) were not treated with radiation therapy following mastectomy when the treatment was indicated (i.e., they had 4 or more positive nodes). Designers of this program estimate that the initial data collection and analysis of claims data for breast cancer among an insured population of 1 million can be accomplished with a relatively small budget (under $150,000). Plans are to extend the program to monitor the quality of other common cancers (i.e., colon, lung, prostate, lymphoma). The current program covers an estimated 25% of the population in the region, but with the anticipated cooperation of other insurers (e.g., Blue Cross Blue Shield of western New York and Univera), a comprehensive quality assurance program could reach up to 90% of the insured population of western New York. The program will also define mechanisms to include nonmanaged Medicare and uninsured patients. SOURCE: S. B. Edge, chief, Breast Department, Division of SurgicalOncology, Roswell Park Cancer institute, personal communication,November 3, 1999. Case Study 8: Monitoring the Quality of Breast and Colon CancerTreatment Among Medicare Beneficiaries in Colorado The quality of cancer care for Medicare beneficiaries in Colorado was assessed through a collaborative effort of the University of Colorado Cancer Center, the Colorado Central Cancer Registry, and the Colorado Foundation for Medical Care (the state Peer Review Organization responsible for Medicare quality initiatives funded by the Health Care Financing Administration). Data from the state's central cancer registry were linked to Medicare administrative records (parts A and B claims) to assess the appropriate use of: (1) adjuvant radiotherapy following lumpectomy for women with Stage I and II breast cancer and (2) adjuvant chemotherapy following surgery for people with

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Stage III colon cancer. These therapies have been proven to improve 5-year outcomes (i.e., reduce local recurrence, increase survival) and are appropriate for individuals with a life expectancy of at least 5 years. Life expectancy in Colorado is more than 10 years for those having reached age 75, and so underuse of adjuvant therapies among the elderly can serve as a population-based indicator of poor care. Cancer registry data for patients diagnosed in 1994 and 1995 with breast or colon cancer were matched with the Medicare master beneficiary file for Colorado. After matching, information from the Medicare Part A inpatient billing file and the Medicare Part B outpatient billing file was added to the analytic file. Thereafter, all personal identifiers were removed from the data set. Analyses were conducted on a statewide basis, without regard to the specific identities of either the communities, healthcare providers, or hospitals. During the study period, roughly one-quarter of Medicare beneficiaries were enrolled in HMOs. HMOs do not submit diagnosis and treatment data to Medicare; therefore, information on use of adjuvant therapies among HMO members was limited to data from the cancer registry. The registry captured information on radiotherapy quite well, but it missed a sizable share of chemotherapy treatments. The medical charts of patients who did not receive adjuvant therapies were reviewed to verify that they had not been used (83% of records were obtained for review). Only 72% of women aged 65 and older received adjuvant radiotherapy following lumpectomy for Stage I and II breast cancer, and only 52% of individuals treated surgically for Stage III colon cancer received adjuvant chemotherapy. Investigators assessed the role of a number of factors potentially associated with underuse of adjuvant therapy, including age, gender, race, place of residence (metro vs. nonmetro area), type of medical insurance (fee-for-service vs. HMO), tumor size and grade, and comorbidity (as captured from Medicare Part A hospital records). The principal determinant of use of adjuvant therapies was age. The increasing prevalence of comorbidities with advancing age had a minimal role in the less frequent use of adjuvant therapy after age 65. Reviews of medical charts indicated that patient refusal was rarely the reason therapies were not used. To disseminate the study findings and improve care, a short (17 minute) video was made and distributed to hospital tumor boards and registrars around the state, findings were published in a newsletter mailed to all physicians in the state, and presentations were given at professional conferences. Response on the part of providers has been generally positive. To better understand why adjuvant therapies are being underused among the elderly in Colorado, a new study has been launched of physician –patient interactions in decision-making about adjuvant chemotherapy for colorectal cancer. The Medicare-registry matching study will be redone in 2000 on cases diagnosed from 1996 to 1998 to assess progress. SOURCE: T. Byers, principal investigator, University of ColoradoCancer Center, personal communication, October 7, 1999.

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Case Study 9: Central Florida Health Care Coalition The Central Florida Health Care Coalition (CFHCC) represents area businesses, including Walt Disney World, Universal Studios, Lockheed Martin, as well as public employers, that together insure nearly 900,000 individuals. Employees of coalition members are covered by eight health insurance plans. CFHCC has for 8 years used Atlas® data to monitor inpatient care. Patient satisfaction with care is assessed using the Consumer Assessment of Health Plans Survey (CAHPS), developed by the Agency for Healthcare Research and Quality. CFHCC has analyzed data from the state cancer registry regarding breast cancer treatment, but difficulties with data from cancer registries, such as inaccurate information on stage of diagnosis, led the coalition to pool other sources of data to assess cancer care quality. CFHCC has recently turned its attention to the quality of outpatient care and has contracted with a private firm, ProtoCare Sciences, to develop an extensive claims database that will include claims for outpatient care for three of the Coalition employers. Attention will initially be focused on 10 noncancer diagnosis-related groups (DRGs) (e.g., cholesterol control, depression) and 5 physician groups (i.e., family practice, internal medicine, gastroenterology, cardiology, obstetrics-gynecology). Profiles of care will be developed for each physician with a case load of at least 30 for the given indication. Plans are to expand the database to include other conditions such as breast cancer and to profile cancer care providers such as surgeons and oncologists. SOURCE: Becky J. Cherney, President/CEO, CFHCC, personal communication,November 23, 1999. SUMMARY As the case studies illustrate, available data resources provide a number of ways to implement quality improvement programs. Some programs depend entirely on retrospective reviews of medical charts or hospital cancer registries while others rely on multiple sources, for example, administrative claims data linked to cancer registry data. The case studies are, in fact, a testament to creativity—data intended for other purposes have in several instances been manipulated to monitor quality care, and sometimes appear to have been used within programs to effect improvements in care. The case studies demonstrate several barriers to systematic quality monitoring. First, because of a lack of recognized measures of quality, provider groups or organizations have themselves frequently assumed the task of reviewing evidence, developing guidelines, and identifying measures. These activities are very costly undertakings, require considerable expertise, and need to be continually reviewed in light of new evidence. Many of the systems could have benefited from an established set of quality measures. The development of

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care TABLE 3.4 Facilities Approved by American College of Surgeons' Commission on Cancer, 1999 Category No. Percent Total 1,479 100% Community Hospital Cancer Program 631 3 Community Hospital Comprehensive Program (>300 analytic cases/year) 465 31 Teaching Hospital Cancer Program 314 21 NCI-designated programs 23 2 All others (e.g., freestanding cancer program offering two of three treatment modalities; single modality centers) 46 3 SOURCE: Morrow, 1999. Case Study 10: The National Cancer Data Base A major initiative of the Commission on Cancer, in collaboration with the American Cancer Society, is NCDB, which collects data from 1,600 hospitals and other facilities in all 50 states. Since 1996, all approved facilities have been required to report all of their cases for 38 cancer sites to the program. Data are also collected from facilities that are not approved. In 1997, 835,000 cases were reported to NCDB. Each participating hospital is given back its own data in summary form, which it can use to compare with national data. The national data allow problem areas to be pinpointed (e.g., widespread use of an inappropriate treatment for a particular type of cancer) and trends to be observed over time in such characteristics as stage at diagnosis, percentage of patients who have complete staging information, and type of treatment given. Results of NCDB analyses are published regularly in professional journals (Bland et al., 1998; Sener et al., 1999). In addition to routine data collection, each year two Patient Care Evaluation (PCE) studies are carried out, focusing on specific cancer types or general treatment issues. Since 1977, 33 PCE studies have been conducted. Recent PCE studies have focused on colorectal cancer and Non-Hodgkin's lymphoma. More extensive data are collected for these special studies, allowing a more detailed analysis of how patients are treated, with the data again fed back to hospitals for comparison in the national spectrum. SOURCE: Morrow, 1999. guidelines through NCCN seems to have spurred quality monitoring activities beyond the cancer centers that developed them (e.g., the Roswell Park Cancer Institute initiative), and the development of breast cancer quality indicators by FACCT seems to have also promoted measurement activities (e.g., Providence

OCR for page 18
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Health System). If an available set of core cancer care measures were available, it would likely be adopted. Likewise, virtually no benchmarks exist with which to gauge success. Systems sometimes establish internal benchmarks based on practice norms, but they often have no way to know whether their performance is better or worse than that of providers outside their practice system. In addition, some statistical issues, if not taken into account, may skew benchmarks (e.g., providers with small numbers of cases unduly affecting norms). Again, these case studies may not represent well the scope of quality assessment activities in cancer care, but they would suggest that relatively little attention is being paid to the full spectrum of cancer care, for example, the quality of pain management and end-of-life care.