5

Status of the Cancer Care Data System

The National Cancer Policy Board (board) has defined several characteristics of an ideal cancer care data system. The system should include meaningful indicators of care quality and provide quality benchmarks for use by systems of care (e.g., hospitals, provider groups, and managed care systems). The system should include recently diagnosed individuals with cancer in care settings representative of contemporary practice across the country, as well as draw from information sources with sufficient detail to allow appropriate comparisons. Chapter 2 described 10 attributes of an ideal system (which could include several distinct databases):

  1. a set of well-established quality-of-care measures,

  2. reliance on computer-based patient records for information on patient care and outcomes,

  3. standard reporting of cancer stage, comorbidity, and processes of care,

  4. national, population-based case selection,

  5. repeated cross sectional studies to monitor national trends,

  6. established benchmarks for quality improvement,

  7. data systems for local quality assurance purposes,

  8. public reporting of selected aggregate quality scores,

  9. adaptability, and

  10. protections to ensure privacy of health information.

This chapter evaluates how close we are to having such an ideal cancer care data system and discusses how progress could be made to achieve these attributes.



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 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care 5 Status of the Cancer Care Data System The National Cancer Policy Board (board) has defined several characteristics of an ideal cancer care data system. The system should include meaningful indicators of care quality and provide quality benchmarks for use by systems of care (e.g., hospitals, provider groups, and managed care systems). The system should include recently diagnosed individuals with cancer in care settings representative of contemporary practice across the country, as well as draw from information sources with sufficient detail to allow appropriate comparisons. Chapter 2 described 10 attributes of an ideal system (which could include several distinct databases): a set of well-established quality-of-care measures, reliance on computer-based patient records for information on patient care and outcomes, standard reporting of cancer stage, comorbidity, and processes of care, national, population-based case selection, repeated cross sectional studies to monitor national trends, established benchmarks for quality improvement, data systems for local quality assurance purposes, public reporting of selected aggregate quality scores, adaptability, and protections to ensure privacy of health information. This chapter evaluates how close we are to having such an ideal cancer care data system and discusses how progress could be made to achieve these attributes.

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care WELL-ESTABLISHED QUALITY-OF-CARE MEASURES In its April 1999 report, the board noted the absence of a core set of quality measures for cancer care and recommended that such a set be developed through a coordinated public–private effort. While there is no accepted set of measures, several “process” quality measures have been used to assess cancer care (Table 5.1). Process quality refers to what providers do for patients and how well they do it, both technically and interpersonally. Effective process measures are identified through evidence from appropriately designed studies (e.g., clinical trials) that show a link between a particular process of care and better outcomes. Once this link has been established, systems can be put into place to monitor adher- TABLE 5.1 Examples of Process Measures Used to Assess Quality of Cancer Care Cancer Site Process Measure Breast Quality of staging information (tumor size, estrogen receptors, lymph node dissection) Initial biopsy prior to total mastectomy Use of breast conserving surgery for local disease Local breast radiation following lumpectomy Adjuvant chemotherapy for premenopausal women with node-positive cancer Use of perioperative bone scan (not indicated) Use of perioperative abdominal CT scan (not indicated) Referral to medical oncologist to discuss adjuvant therapy If mastectomy, visit to plastic surgeon to discuss reconstructive surgery Follow-up mammography within first 18 months post-operatively Follow-up bone or CT scans for suspicious symptoms per year (not indicated) Colon Adjuvant chemotherapy for Stage III disease Rectum Adjuvant chemotherapy and radiation therapy for Stage II/III disease Prostate Presentation of treatment alternatives to patient Rates of surgical treatment among men aged 70 and older (should be low) Small-cell lung cancer (SCLC) Chemotherapy for limited SCLC Nonseminoma testicular cancer Chemotherapy Hodgkin's disease Chemotherapy for Stage IIB or IV disease Non-Hodgkin's lymphoma Chemotherapy for diffuse intermediate or high-grade disease SOURCE: IOM, 1999a.

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care ence to the recommended processes. Current performance can be compared to a physician's or plan's prior performance, to the performance of other physicians and plans, or to benchmarks of performance. If performance falls below expected levels, educational or other interventions can be employed to change provider behavior. Although they are intuitively appealing, measures of patient outcomes have certain limitations in the context of quality assessment. Many factors other than health care can affect outcomes; consequently, a finding of a higher-than-expected 5-year mortality rate in one hospital relative to another could just as easily reflect differences in patients' comorbidity status as the actual quality of care received. Process measures are advantageous because they tend to be frequent, immediate, controllable, and rarely confounded by other factors (Eddy, 1997). Some process measures, however, may depend on compliance or patient preferences. For example, although chemotherapy is highly recommended after surgery for certain cancers, some patients might decline treatment because they do not wish to experience its associated toxicities. Therefore, 100% adherence may not be a reasonable target for an indicator specifying adjuvant chemotherapy for these patients. A set of “process” and “outcome” quality-of-care measures has been proposed for each phase of care for women with breast cancer (Mandelblatt et at., 1999) (Table 5.2). The authors of these proposed measures suggest that the measurement set be reviewed for its clinical relevance and feasibility, assessed for its validity and reliability, evaluated against established methodologic standards, and be subject to peer review. Since the publication of the board's Ensuring Quality Cancer Care report, the National Cancer Institute (NCI) has committed to developing a core set of outcome measures. For each major cancer site, NCI plans to adopt, adapt, or develop one or more outcome measures applicable to each stage of the cancer care continuum, from prevention through end-of-life care (NCI, 1999b). NCI has commissioned several literature reviews to support this effort. NCI will also collaborate with the Agency for Healthcare Research and Quality (AHRQ, formerly the Agency for Health Care Policy and Research) to support and fund research aimed at increasing the use of guideline information in practice, for example, developing Web-based decision support tools for guideline development. NCI-sponsored research will also support evaluations of how such tools increase the likelihood that appropriate care is delivered to patients (NCI, 1999b). Other efforts under way to define cancer care measures include the National Committee for Quality Assurance's (NCQA) appointment of an Oncology Measurement Advisory Panel to review measures relevant to cancer care (Oncology News, 1999; Winn, 1999). NCQA accredits managed care plans and has produced a widely used report card monitoring system called the Health Plan Employer Data and Information Set (HEDIS ®) (www.ncqa.org). The HEDIS® cancer quality indicators have thus far targeted early detection and diagnosis, not care received after cancer is diagnosed. NCQA has implemented treatment-related quality measures for other conditions (e.g., diabetes) and has experience evaluating the feasibility of instituting a measure taking into consideration statistical issues in

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care volving sampling, the reliability of available clinical sources of information, and ensuring that valid comparisons can be made between health plans. The American Society of Clinical Oncology (ASCO) is also planning to define a set of measures of quality of care to use in the development of a quality monitoring system for cancer patients (ASCO, 2000). Health services research is also advancing in this area. Cancer quality indicators have, for example, recently been identified for six cancer sites (breast, cervical, colorectal, lung, prostate, skin) as part of a comprehensive quality measurement system designed to assess quality within managed care plans (Malin et al., 2000). While first steps to identify and adopt cancer care measures are being taken within the cancer care community, a broadly focused response to issues of healthcare quality is taking shape at the federal level. In the wake of the influential report of the President's Advisory Commission on Consumer Protection and Quality in the Health Care Industry (1998), complementary bodies have been formed on healthcare quality—one in the public sector to promote interagency coordination among the Department of Health and Human Services (DHHS) and other federal agencies (Quality Interagency Coordination Task Force [QuIC]), and the other in the private sector to improve healthcare quality, measurement, and reporting (National Forum for Health Care Quality Measurement and Reporting [forum]). The aims and activities of both the QuIC and the forum are quite relevant to the quality cancer care agenda. The QuIC's goal is to ensure that all federal agencies involved in purchasing, providing, studying, or regulating healthcare services are working in a coordinated way toward the common goal of improving quality of care. The Secretary of Health and Human Services and the Secretary of Labor serve as the cochairs of the QuIC, and the Administrator of the Agency for Healthcare Research and Quality serves as the chairman for day-to-day operations. The QuIC seeks to (www.ahrq.gov/qual/quicfact.htm): provide people with information to assist them in making choices about their care, improve the care delivered by federal providers and purchased on behalf of federal beneficiaries, and develop the infrastructure need to improve the healthcare system. The QuIC has established work groups in five areas: Patient and consumer information—to address critical barriers to effective communication with patients about quality; Improving quality measurement—to develop a “tool box” of quality measures and risk adjustment methods used by federal agencies (the work group is developing an inventory of measures and risk adjustment methods for use within federal agencies);

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care TABLE 5.2 Potential Measures of the Quality of Breast Cancer Care Care Domain Potential Measure Process Outcome Data Source Screening Mammography rate (initial and return)   Administrative database, chart   Clinical examination rates   Chart, patient self-report   Stages of cancer   Chart, tumor registry Diagnosis Time from abnormal screen to diagnosis   Chart, administrative data   Estrogen/progesterone receptors   Chart, laboratory, pathology reports   Rate of true-positive biopsies   Chart, administrative data Treatment, local and systemic Documentation of choice for local and systemic treatment; documentation of patient participation   Chart, patient self-report   Pain and symptom control   Chart, patient self-report   Time from diagnosis to treatment   Administrative data   Rate of tamoxifen prescription or chemotherapy   Chart, pharmacy data   Offer of reconstruction/plastic surgery referral after mastectomy   Chart

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care   Functional status/Quality of life   Patient survey, chart   Satisfaction   Patient report   Disease-free survival, survival   Chart Rehabilitation Evaluation of psychosocial needs   Chart, administrative data   Rehabilitation evaluation   Chart, administrative data   Psychologic function   Patient survey, chart   Physical function   Patient survey, chart Surveillance Documentation of recurrence; rates of mammography and clinical breast exam   Chart Palliative care Pain control; offer of hospice; documentation of discussion of “Do Not Resuscitate” orders and living wills   Chart, patient report   Quality of dying experience   Patient/family report SOURCE: Mandelblatt et al., 1999.

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Developing the workforce—to determine how to expand and improve the current methods of ensuring the skills of the healthcare workforce, and equipping healthcare workers to improve the care they deliver; Key opportunities for improving clinical quality—to mount an effort to improve clinical quality of care in two areas, diabetes and depression. For diabetes, the work group is focusing its efforts on having all federal programs agree to use the Diabetes Quality Indicator Project measures of care and then to improve healthcare provider performance based on these indicators. For depression, the work group is developing an evidence-based guideline to improve the identification and treatment of depressed individuals served by federal healthcare programs; Improving information systems—to further efforts to develop a standardized language that will enable computerized comparisons of quality across federal agencies, and to examine the potential uses of telemedicine for helping to improve the quality of care. The complementary Quality Forum established with a private-sector base will attempt to (http://www.qualityforum.org): ensure system-wide capacity to evaluate and report on the quality of care, promote and inform consumer choice and further consumer understanding and use of quality measures, enable providers to use data to improve performance, allow meaningful quality comparisons of healthcare providers and plans, promote competition on the quality of healthcare services, use broad representation to marshal market forces for quality, and reduce the burdens on providers and health plans by enabling them to collect consistent data that avoids duplication. The board is encouraged that first steps toward the development of a core set of cancer care quality measures have been taken, and it hopes that its vision of a public/private collaboration on its creation will be realized. The mechanisms for such a collaboration appear to be in place. COMPUTER-BASED PATIENT RECORDS The revolution in information technology provides many opportunities to improve the quality and timeliness of quality-of-care studies. Recording patient data using computer-based patient records (CPRs), Internet (and Intranet) communications, and statistical software provides opportunities to rapidly turn raw data into meaningful reports. The healthcare industry, however, lags behind others in adopting information technology that could promote faster feedback to providers on the quality of care (IOM workshop on Healthcare Informatics,

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care 1999). For example, while all financial transactions are now electronic, relatively few health systems have CPR systems (www.himss.org/survey). A very sophisticated oncology patient record system has been developed for physicians practicing within the OnCare system (see Chapter 3). Here, physicians enter information into the CPR as they provide care. Embedded in the system are clinical practice guidelines, information on standard chemotherapy regimens, and summary data on the experience of other patients throughout the system available in near real time. Such systems are not yet in common use, and while promising, their impact on quality of care has not been fully evaluated. While consumers are navigating the Internet in record numbers searching for information about their health and care, only 37% of office-based physicians were estimated to be using the World Wide Web in 1999 (www.ama-assn.org/ad-com/releases/1999/991203b.htm). This trend may change as new Internet-based products designed to ease the burden of filing insurance claims are targeted to physicians, and as medical professionals become more aware of informational resources provided via Internet access. Until physicians increase their use, the potential for the Internet or Intranets (controlled-access versions of the Internet) to improve the timeliness of reporting between clinicians and cancer registries will not be met. A major barrier to the adoption of information technology in medicine is concern about protecting the privacy of confidential medical information (see discussion below). STANDARD REPORTING Quality assessments depend on the accurate recording of cancer stage and degree of comorbidity because what is considered appropriate treatment varies by these patient attributes. These factors must be carefully controlled when comparing the quality of care (e.g., by site of care, type of provider) so that apparent differences in quality can be correctly attributed to differences in attributes of care, rather than to differences in the degree of illness of the patients being compared. Cancer Stage Three major staging systems are used for cancer surveillance in the United States: summary stage (SS); extent of disease (EOD); and tumor, node, metastasis (TNM). In addition, a number of site-specific staging schemes (e.g., prostate and bladder cancer, melanoma) are available. Within a particular staging system, different assignments also can be made, depending on whether one considers information from the pathology report or information from the patient's clinical assessment. This variation further complicates the staging options. Cancer cases are often reported to multiple organizations, each with a preferred staging system. These varying expectations are burdensome for the person reporting the case. Furthermore, analysts who try to integrate these multiple

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care sources of information, or who want to make comparisons across data systems, may have difficulty collapsing staging categories in one system to match another. Even within states, hospitals vary on how they report stage. Summary stage is used by most population-based registries, but clinical researchers do not find this measure to be sufficiently precise to evaluate the quality of care. To address these issues, a Stage Task Force was established in 1997 to recommend a “best” staging system (Edge et al., 1999).1 However, organizations resisted switching to a single staging system because of their various orientations and needs (e.g., historical trend analyses, evaluations of appropriateness of care). Consequently, the Task Force has instead recommended a uniform set of data items from which SS, EOD, and TNM can be derived. This uniform data collection set would likely ease the burden of reporting because one system with a single set of rules would be applicable for all organizations. Organizations that wanted to report data using their preferred system could use a computer algorithm to translate from the new uniform system back to any one of the other systems (i.e., SS, EOD, TNM). Comorbidity The majority of cancer patients are over age 65 and often have ailments other than cancer. These conditions may render an overall prognosis so poor for the patient that an otherwise recommended treatment might be withheld. Furthermore, some comorbid conditions may affect patients ' ability to tolerate recommended treatment for cancer, negatively affecting their response to treatment. Incorrect conclusions could be reached about the quality of care without sufficient information about comorbidity (Greenfield et al., 1988). Lower than expected use of a recommended cancer treatment at a particular hospital, for example, could signal a patient population too ill to tolerate the treatment, rather than the receipt of poor quality care. Several instruments have been developed to classify different comorbid diseases and to quantify the severity of the overall comorbid condition (e.g., Charlson Comorbidity Index, Kaplan-Feinstein Index). None of the instruments were specifically designed to study comorbidity in cancer patients. Nevertheless, these instruments have been used to classify comorbidity in several types of cancers and have performed well (Piccirillo, 1999). Very accurate assessments of comorbidity can be made with information from medical records, and recommendations have been made to train cancer registrars to code comorbidity and to include this information as a required data element in cancer registries (Piccirillo, 1999). In a pilot study, trained registrars could accurately and quickly code cases using a comorbidity index while they 1   The Task Force is a collaboration of the American Joint Committee on Cancer (AJCC), the National Cancer Institute, the Centers for Disease Control and Prevention, and the North American Association of Central Cancer Registries (Edge et al., 1999).

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care abstracted information from the medical chart (Piccirillo et al., 1999). Somewhat less accurate and complete assessments of comorbidity can be made using claims data (e.g., Medicare hospital discharge summaries). In addition to having standardized reporting of important clinical determinants of treatments and outcomes, information has to be available within a relatively short period of time, allowing for timely assessments of care. In its review of the literature on the quality of cancer care in Ensuring Quality Cancer Care, the board noted that many recently published studies relied upon the experience of patients cared for in the 1980s. In part, this observation reflects the inclusion in some research of a 5-year mortality outcome that, by definition, slows the reporting of results. One of the advantages of quality assessments that rely on process of care measures is the more rapid turnaround time for results. NATIONAL, POPULATION-BASED CASE SELECTION Almost all of the case studies of quality care assessment described in Chapter 3 relied on convenience samples of patients. The initiative at Roswell Park Cancer Institute, for example, identified cancer cases through one insurance company's claims. This quality assessment was cosponsored by the insurer, and the intention of the study was to assess the quality of care delivered to that defined population. The convenience sampling employed was therefore entirely consistent and appropriate given the aims of the study. Indeed, quality studies are probably most appropriately conducted through health systems responsible for the care of the group of patients under study. Within these circumstances, results are more likely to be acted upon by administrators and providers. Broad-based assessments, however, are needed to determine the quality of care nationally and to understand if certain population subgroups are particularly subject to substandard care. Furthermore, the establishment of national benchmarks of care that can be used by local systems of care to evaluate their programs relative to others in their area, or to the population as a whole, depend on samples that are nationally, or at least regionally, representative. Notable in almost all of the case studies was the lack of comparison values on measures applied within the studies. Currently, no national benchmarks of the quality of cancer care exist. The source with the greatest promise for delivering representative samples of patients with which to measure the quality of cancer care is the state cancer registries, recently organized into the National Program of Cancer Registries (NPCR). Other potential sources include the Surveillance, Epidemiology, and End Results (SEER) program of the NCI, and the National Cancer Data Base (NCDB), cosponsored by the American College of Surgeons' Commission on Cancer (ACoS-CoC) and the American Cancer Society (ACS) (also described in Chapter 3).

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care National Program of Cancer Registries In all states, cancer is a reportable disease; that is, by law it must be reported to a state registry (J. Enders, acting section chief, Cancer Surveillance Branch, CDC, personal communication, March 30, 2000; www.cdc.gov/nccdphp).2 Registries are essential to understanding the burden of cancer and in evaluating the success of cancer prevention programs. They are also used to target resources to areas that may be underserved by screening or public health education programs. The central purpose of state cancer registries is cancer surveillance, monitoring the burden of cancer for a given population. When registration activities are organized to obtain complete counts of cancer for a given population (e.g., a state), the system is said to be “population-based” (Austin, 1994). Some registries focus on developing incident rates, the rate of new cases of cancer for their population. Such “incidence only” registries must have accurate and complete case counts and the data necessary to categorize patients and tumors. The need for, and the success of, cancer control programs can be assessed with these registries. Other population-based registries have a broader charge and may follow up on cases identified to collect data on outcomes (e.g., recurrence, death) and treatment. These registries are able to provide information on survival rates among those diagnosed with cancer and to address issues regarding access to certain cancer care services. Cancer registries strive to identify at least 90% of new cases of cancer within 2 years after the diagnosis year and maintain a high degree of accuracy as determined by the North American Association of Centralized Cancer Registries (NAACCR) certification process (Tucker et al., 1999). In 1999, NAACCR certified 19 U.S. population-based state registries as having achieved this level of quality (Appendix D). The draft Healthy People 2010 objectives for cancer surveillance include “Increas[ing] the number of states that have a statewide population-based cancer registry that captures case information on at least 95% of the expected number of reportable cancers” (M. Kaiser, Office of Program and Policy Information, NCCDPHP, CDC, personal communication, January 19, 2000). The cancer registration process is very labor intensive, and registries may not report on cases ascertained in a given year for 2 to 3 years. It can take this long to complete data collection on identified cases, for example, those residing in one state but diagnosed and treated in another. Extensive time may also be needed to obtain information on those reported to the registry by private pathology laboratories. Private laboratories may not provide necessary identifying information on the patient for whom the diagnosis was made (i.e., the registry may have to contact the physician who referred the specimen for analysis). 2   While all states have some form of mandated cancer reporting, the degree to which reporting is required varies. Most states require reporting from all acute care hospitals, but others require more extensive reporting (e.g., from outpatient surgery/treatment centers, pathology labs, or physician offices) (J. Enders, acting section chief, Cancer Surveillance Branch, CDC, personal communication, March 30, 2000).

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care National Cancer Data Base American College of Surgeons/American Cancer Society Designed to monitor quality of cancer care In 1997, an estimated 69% of incident cancer cases were reported by 1,629 of 2,000 hospitals with tumor registries (81% of hospitals with registries). A high proportion of incident cancers included nationally Recurrence and survival outcomes tracked Data available for analysis relatively soon after collection Individuals not diagnosed or treated in hospitals are not represented (e.g., cancer, such as melanoma, treated in outpatient settings). Relatively high response rate among selected hospitals, but there is potential for bias.

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care applied to the determination of a target for population-based mammography screening rates; this target has been proposed as a realistic, data-driven goal for adoption in Healthy People 2010 (Allison, 1999). Once agreement on “process” measures for cancer care has been reached, this model developed for AMI could be tested among cancer care providers. DATA SYSTEMS FOR LOCAL QUALITY ASSURANCE PURPOSES Once benchmarks are established, hospitals, health plans, and provider groups should be able to assess their care relative to national or regional norms and identify ways that care could be improved. Individual providers who are working within large systems, for example, an integrated delivery system, or who are caring for employees of big companies or subscribers of large insurers, may be subject to quality improvement programs. Experience suggests that once credible performance data are presented to providers, better clinical behavior and improved care will follow (Lazar and Desch, 1998; Newcomer, 1997). An accountability framework is developing within the private sector, which incorporates performance measurement of health plans and other healthcare organizations. Inclusion of cancer care measures into these systems could provide valuable information about cancer care to consumers, purchasers, and providers of care:* Health plans report quality of care data to the National Committee on Quality Assurance (NCQA), an accrediting body for managed care plans. Hospitals and healthcare organizations are surveyed and accredited according to standards established by the Joint Commission on the Accreditation of Healthcare Organizations (JCAHO) and, since 1997, have been required to participate in a quality performance system. In 1998, NCQA, JCAHO, and the American Medical Association established the Performance Measurement Coordinating Council, a 15-member group that will work to coordinate performance measurement activities across the entire healthcare system (www.jcaho.org). National Committee for Quality Assurance NCQA accredits managed care plans, Health Maintenance Organizations (HMOs), and Preferred Provider Organizations (PPOs) (www.ncqa.org). NCQA has also produced a widely used report card monitoring system called HEDIS®. HEDIS® measures were initially designed to provide information to large pur- * Cancer care quality assurance programs are described in detail in Chapter 6 of Ensuring Quality Cancer Care (IOM, 1999a).

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care chasers about the quality of care offered to employees. More recently, the audience for results from HEDIS® has broadened, and HEDIS® indicators are often reported in consumer-oriented report cards. HEDIS® is a performance measurement tool designed to assist purchasers and consumers in evaluating managed care plans and holding plans accountable for the quality of their services. Because HEDIS® has standard measures and uniform data reporting requirements, comparisons can be made across various health plans and their organizational structures (e.g., staff-model HMOs, point-of-service plans). The most recent iteration, HEDIS® 2000, assesses plans in eight domains (www.ncqa.org): effectiveness of care, accessibility and availability of care, satisfaction with the experience of care, stability of the health plan, use of services, cost of care, informed healthcare choices, and descriptive information about the plan. HEDIS® 2000 measures relevant to cancer care are shown in Box 5.1. BOX 5.1 Selected Cancer-Specific (or cancer-relevant) HEDIS® 2000 Measures Effectiveness of Care Advising smokers to quit Cervical cancer screening Breast cancer screening Access to or Availability of Care Adults' access to preventive ambulatory health services Availability of language interpretation services Satisfaction with the Experience of Care Member satisfaction Health Plan Stability Disenrollment Practitioner turnover Health Plan Descriptive Information Provider board certification or residency completion Practitioner compensation Arrangements with public health, educational, and social service organizations SOURCE: www.ncqa.org, Accessed February 17, 2000.

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care The HEDIS® cancer quality indicators have targeted early detection and diagnosis, not care received after cancer is diagnosed. Treatment-related indicators are being evaluated, for example, assessment of the effect of breast cancer therapy on a woman's ability to function and patients' satisfaction with breast cancer treatment. NCQA has halted further work on the indicator related to the stage at which breast cancer is detected because the incidence of breast cancer cases in most health plans is too low to make meaningful comparisons of stage at diagnosis across health plans (Schuster et al., 1998a). NCQA has appointed an Oncology Measurement Advisory Panel to review quality measures relevant to cancer care. HEDIS® is a voluntary system, although managed care plans are finding it increasingly necessary to participate to compete for patients. More than 90% of HMOs report HEDIS® measures, and about 65% of Fortune 500 employers use NCQA accreditation and HEDIS® measures to evaluate the managed care plans with which they contract (O'Kane, 2000). NCQA produces Quality Compass, a CD-ROM-based system that makes it possible for consumers to obtain comparative HEDIS ® ratings for HMOs in communities throughout the United States. A subset of Quality Compass measures appears on the World Wide Web. A health plan can refuse to disclose its HEDIS® profile to the public. In 1997, less than half of plans (45%) permitted public reporting of the data (Bodenheimer, 1999). Joint Commission on the Accreditation of Healthcare Organizations The nonprofit JCAHO, the oldest and largest standard-setting and accrediting body in health care, has broadened its institutional coverage from solely hospitals to a wide array of delivery systems, including health plans, integrated delivery networks, PPOs, home care organizations, nursing homes and other long-term care facilities, behavioral healthcare organizations, ambulatory care providers, and clinical laboratories. JCAHO evaluates and accredits more than 19,000 healthcare organizations in the United States (www.jcaho.org). About 80% of U.S. hospitals participate, representing about 96% of all inpatient admissions. For accreditation, JCAHO conducts an on-site quality assessment every 3 years. It covers such topics as patient rights, patient care, patient education, continuity of care, ongoing efforts to improve quality, safety plans, information management, and infection control. Although JCAHO (and other accrediting organizations) has traditionally focused on structural measures of quality—such as whether a hospital has appropriate capacity for the covered patient population—it now incorporates process and outcomes measures into its accreditation criteria. JCAHO relies on ACoS-CoC survey findings for cancer programs within JCAHO-accredited organizations (www.facs.org/about_college/acsdept/cancer_dept/cocjcaho.html). JCAHO instituted the ORYX system in 1997, which requires organizations seeking JCAHO accreditation to select from among 60 performance measure

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care ment systems and two specific indicators on which they will report their care. Hospitals and long-term care facilities began reporting with these indicators during early 1999. With institutions choosing their own indicators, making comparisons across institutions will be challenging. It should allow for comparisons with prior years within the same institution, benchmarks, and goals. One of the accepted indicator systems is the MEDSTAT Group's Indicator Measurement System (IMSystem®), which has specifications for 42 quality-of-care indicators (including 5 for cancer care, Table 5.4). About 20–25 hospitals currently use the IMSystem® oncology measures (L. Homra, clinical consultant, MEDSTAT Group, personal communication, February 17, 2000). Public Reporting of Aggregate Quality Scores The board has recommended that “Cancer care quality measures should be disseminated widely and communicated to purchasers, providers, consumer organizations, individuals with cancer, policy makers, and health services researchers, in a form that is relevant and useful for healthcare decision-making” (IOM, 1999a). Quality measures enable consumers and purchasers to judge the quality of a system of care by its performance relative to evidence-based standards. Many opportunities are available for exerting leverage on the healthcare system to improve quality: Large employer groups are holding managed care plans accountable for quality performance goals. HCFA requires Medicare and Medicaid health plans to produce standard quality reports. State Medicaid programs are beginning to include quality provisions in their contracts with plans and providers. Six of 10 new cancer cases occur among people aged 65 and older and, consequently, Medicare is the principal payer for cancer care. The number of evaluations of the quality of care among Medicare beneficiaries will increase because breast cancer has been identified as one of six conditions that state-based PROs must now target in their efforts to improve the quality of care (Jencks, 1999). Information about quality cancer care is becoming more available to individuals with cancer (or at risk for cancer), but it is not yet easily accessible or understandable by consumers. By the time a diagnosis of cancer is made and individuals have a clear reason to seek quality care, it is often too late to switch health plans. Also, even if they wanted to switch, most people do not have access to alternative plans. Individuals may use available quality indicators to choose doctors and hospitals within their plans, and perhaps to choose alternative courses of treatment, but evidence suggests that individual consumers can exert only a modest “market” pressure for quality improvement through access

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care TABLE 5.4 IMSystem® Oncology Indicators   Data Staging Breast Cancer Lung Cancer Colon or Rectum Cancer Focus Availability of data for diagnosis and staging Use of staging by managing physicians Use of tests critical for prognosis and clinical management of female breast cancer Effectiveness of preoperative diagnosis and staging Comprehensiveness of diagnostic workup Numerator Patients undergoing resection for primary cancer of the lung, colon or rectum, or female breast for whom a surgical pathology consultation report is present in the medical record Patients undergoing resection for primary cancer of the lung, colon or rectum, or female breast with stage of tumor designated by a managing physician Female patients with Stage I or greater primary breast cancer who, after initial biopsy or resection, have estrogen receptor analysis results in the medical record Patients with non-small-cell primary lung cancer undergoing thoracotomy with complete surgical resection of tumor Patients undergoing resection for primary cancer of the colon or rectum whose preoperative evaluation by a managing physician included examination of the entire colon Denominator Patients undergoing resection for primary cancer of the lung, colon or rectum, or female breast Patients undergoing resection for primary cancer of the lung, colon or rectum, or female breast Female patients with Stage I or higher primary breast cancer undergoing initial biopsy or resection Patients with nonsmall-cell primary lung cancer undergoing thoracotomy Patients undergoing resection for primary cancer of the colon or rectum SOURCE: IMSystem,® 1997.

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care to better information about the quality of cancer care (IOM, 1999a). Large purchasers, such as employers, are likely to exert more leverage and to have designated staff to assess alternative plans. A few purchasing coalitions are using information on cancer care quality to improve care (see Case Study 9 in Chapter 3). The Pacific Business Group on Health (PBGH), for example, is a nonprofit coalition of large healthcare purchasers in California and Arizona representing, as of 1996, 2.5 million insured individuals. PBGH collects and analyzes health plan performance data to produce report cards for consumers; promotes shared treatment decision making between providers and consumers; and collects, analyzes, and reports plan-level consumer satisfaction ratings (Castles et al., 1999; President 's Advisory Commission, 1998). PBGH is also developing several disease-specific quality assessment programs, including one for breast cancer. PBGH was the first purchasing coalition to impose a condition on contracting plans whereby it would withhold 2% of the premium until the plans achieved specific goals for improving customer satisfaction and quality of care. NCI is cofunding with AHRQ a Request for Proposals entitled “Making Quality Count.” One of the areas of study that will be encouraged is the use of report cards and other tools for communicating quality of care information to payers, providers, and patients. Innovation and research in the presentation of this information will also be sought (NCI, 1999b). NCI also plans to conduct a national cancer communications survey, which will have a quality of cancer care module to assess how people obtain quality of care information, and what their preferences are for this type of information. NCI will also develop and issue an RFA for fiscal year 2001 to create cancer communications “centers of excellence.” Prospective applicants will be encouraged to focus, in part, on quality-of-care issues. Protections to Ensure Privacy of Health Information Legal protections and data security systems must be in place to ensure that data collected and stored about an individual's diagnosis and treatment of cancer are used only for legitimate public health purposes. No legal action regarding a breach of confidentiality from cancer registries has been reported, but concerns about privacy1 and confidentiality2 are increasing as data are being transmitted electronically and cancer registry data are being merged with other sources (Coleman et al., 1992; Newcombe, 1995). Personal identifiers, including name, birth date, and social security number are required data elements in the NPCR and NCDB because effective cancer registration involves the collection and 1   Health information privacy is an individual's claim to control the circumstances in which personally identifiable information is collected, used, and transmitted (Hodge, 1999). 2   Confidentiality is privacy interests arising out of a specific relationship with the person about whom information is gathered (Hodge, 1999).

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care linkage of data about individuals with cancer, often from several different sources. Such linkages, for example, are necessary to eliminate duplicate reports of a case from different healthcare providers, and to identify deaths that have occurred among cases reported to the registry so that survival statistics can be calculated (linkage to death certificate data). Furthermore, data sharing must occur between states to identify residents diagnosed out-of-state. Even though the NPCR and NCDB require that social security number and other identifying information be reported when they are available, these data elements are not forwarded to a central entity at either the CDC or NCDB. Having personal identifiers available centrally could facilitate national or regional assessments of the quality of cancer care, but such use of personal identifying information raises a host of legal and ethical issues. The CDC has provided model legislation for states to assist them in complying with provisions of the Cancer Registries Amendment Act of 1992 (the Act that established the NPCR). According to this act, states must promulgate regulations providing “for the protection of the confidentiality of all cancer case data reported to the statewide cancer registry, including a prohibition on disclosure to any person of information reported to the statewide cancer registry that identifies, or could lead to the identification of an individual cancer patient, except for disclosure to other State cancer registries and local and State health officers” (P.L. 102-515). The regulations must also provide “for a means by which confidential case data may in accordance with State law be disclosed to cancer researchers for the purposes of cancer prevention, control and research.” Almost all of the NPCR programs have established all of the regulations specified in the Cancer Registries Amendment Act on the use of registry information for cancer prevention and control. A number of federal and state approaches to protect privacy of health information (not specific to cancer) are described in Box 5.2. BOX 5.2 Federal and State Approaches to the Protection of Privacy of Health Information Federal Approaches U.S. Constitution—While the federal Constitution does not expressly provide individuals with privacy rights, the Supreme Court has recognized a limited right to health informational privacy as a liberty interest within the Fifth and Fourteenth Amendments. Federal Statutes—The federal government has enacted several statutes and regulations to protect privacy of health information: Privacy Act of 1974 requires federal agencies to use fair information practices with regard to the collection, use, or dissemination of systematized records.

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Freedom of Information Act of 1966 (FOIA) requires the federal government to provide various information but exempts from governmental disclosure several categories of records which include health information. Electronic Communications Privacy Act of 1986 protects electronic communications during transmission or while in storage against unauthorized interceptions and improper uses, although it likely does not protect interceptions of nonencrypted information over radio frequencies. Federal regulations require privacy protections in relation to the administration of human subject research. Health Insurance Portability and Accountability Act (HIPAA) seeks to reduce the administrative and financial burden of health care by standardizing the electronic transmission of health-related data. HIPAA requires DHHS to set uniform standards for the transmission of health insurance information, including recommendations for security measures to protect private medical information. The DHHS recommendations focus on five key principles: Boundaries: Healthcare information should be disclosed for health purposes only, with limited exceptions. Security: Health information should not be distributed unless the patient authorizes it or there is a clear legal basis for doing so. Those who receive such information must safeguard it. Consumer control: persons are entitled to know of and correct information in their health records and the purposes in which it is being used. Accountability: Those who improperly hold, distribute, or use health information should be criminally punished, especially when such actions are for monetary gain. Those individuals affected by such actions should have civil recourse. Public responsibility: Privacy interests of individuals must not override national priorities of public health, medical research, health services research, healthcare fraud and abuse, and law enforcement in general. State Approaches State constitutions—More than a dozen states have adopted constitutional amendments designed to protect a variety of privacy interests, including limitations on access to personal information. Most only protect against breaches of privacy by government. State statutes—States have enacted health information privacy protection in many forms, including laws similar to the federal Privacy Act and FOIA. State common law—State case law imposes duties of confidentiality on certain healthcare professionals not to disclose health information concerning patients. SOURCE: Hodge, 1999.

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Population-based registries have developed a number of policies and procedures to ensure secure handling and processing during data collection, storage, and analysis of confidential data including (NCI, 1999a): employees of registries sign pledges to maintain and protect confidential information; paper and electronic files are locked in secure areas with restricted access; institutional research review committees govern access to confidential information by persons external to the registry; written agreements outline the responsibilities of investigators requesting registry data and requirements for maintenance of confidentiality; information that could potentially identify an individual, institution, or health-care provider is excluded from public-use data tapes, and analyses of groups are restricted to a sufficiently large size so that personal or institutional identities are obscured; and researchers given permission to contact individuals identified in registries must do so through the patient's physician, who seeks permission from the patient to be contacted directly by the researcher. HIPAA (see description in Box 5.2) requires that national patient identifiers be created, not only for patients, but for employers, providers, and payers. Such identifiers would facilitate data linkages and health services research. The Congress, however, passed a measure that prevents DHHS from using any federal funds to create regulations requiring national patient identifiers, or from assigning numbers (P.L. 105-277, P.L. 106-113; Ziegler, 1999). In the absence of a unique patient identifier, data linkages currently depend on personal identifiers reported to registries, for example, social security number, name, and birth date. Some individuals do not have social security numbers (e.g., illegal immigrants), and in other cases the social security number may not be available in the medical record (e.g., if not required for reimbursement purposes). Computer programs are available to match records based on all available identifying information (the application of such a program is described in Chapter 4). When personal identifiers are reported to either hospital or state registries, they may be held there and not forwarded to central data repositories. For example, even though NPCR and NCDB require reporting of personal identifying information, the data elements are not forwarded to a central entity at either the CDC or NCDB. The CDC plans to pool state cancer registry data into a Cancer Surveillance System (CSS) to provide national estimates of cancer incidence and to facilitate epidemiologic and health services research. The CDC will not, however, directly receive patient identifiers. It is unclear whether state laws will prohibit the release of confidential data to a third party. 3 Within states, 3   According to the HHS Office for Protection from Research Risks, the NPCR does not need CDC/IRB approval for conducting cancer data collection activities because they are considered public health practice surveillance. However, CDC/IRB approval for a call for data through the NPCR-CSS has been applied for and obtained. Such IRB approval is needed because the CSS involves subsequent development of information supplied by NPCR programs into public-use data sets, which could be used for research. In addition to this IRB approval, NPCR has applied for a CDC Assurance of Confidentiality as an extra level of protection for data that will be submitted by NPCR state programs to the CDC (K. Brady, assistant branch chief, NCCDPHP, CDC, personal communication, December 9, 1999).

OCR for page 46
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care record linkages are often performed by in-house registry staff with Internal Review Board (IRB) approval and are protected by state registry law. While these issues will need to be resolved, there are examples of successful data aggregations and linkages—the NCI has pooled cancer data from its participating registries and linked them to Medicare files while adhering to strict privacy and confidentiality rules (see description of the Medicare-SEER linkage studies in Chapter 4). SUMMARY Relative to the ideal described in Chapter 2, current data systems applicable to cancer care quality assessments have serious shortcomings. No well-established set of quality of care measures exists; consequently, quality assessment initiatives have faced the task of defining such measures for themselves. Quality improvement initiatives have been impeded both by the absence of good cancer care measures and limitations of available data systems. Standards are lacking for reporting factors that are needed in the measurement of the quality of cancer care, for example, stage and comorbidity. Providers do not yet use computer-based patient records, and abstraction of quality information from medical records is time consuming, expensive, and labor intensive. Currently, no data systems are in place with which to make national inferences about the quality of cancer care, and providers do not have benchmarks or targets for gauging their performance relative to others. Despite these shortcomings, there is great potential for enhancing current systems to provide better information on the quality of cancer care.