Key Patient Management Topics for Effectiveness Research in Breast Cancer

PRELIMINARY DISCUSSION AND SELECTION OF MAJOR TOPICS

Using the results of the first round of a homework exercise and the issues raised by the presentation of preliminary HCFA data analyses (see Appendix A), the committee discussed the range and clinical depth of breast cancer studies that might be mounted through an effectiveness research program. This session made clear the enormous number and complexity of unanswered questions about the appropriate management of women who are at risk for, who have, and who have had breast cancer. Many of these questions still lie in the realm of efficacy rather than effectiveness—that is, whether a test, a procedure, or a drug works at all, rather than how well it works in the average practice of medicine. To the extent that is true, the issues belong more in the arena of clinical and biomedical research of the sort conducted by the National Institutes of Health than in the area of effectiveness research.

We tried to narrow our task to that of clarifying what set of questions could be satisfactorily addressed in one of three ways: (1) with existing HCFA administrative data of the sort on which the preliminary analyses presented at the workshop were based; (2) with existing HCFA data augmented by clinical data from medical records abstracted by PROs or from specially conducted patient surveys and follow-up studies; and (3) by carefully planned longitudinal studies that combine all necessary inpatient and ambulatory information, such as stage of disease, type of surgical and/or adjuvant therapy, and quality-of-life outcomes.

To this end, we generated numerous topics on which the opening discussions focused. Some topics cut across several issues, such as the need to examine patterns of care by geographic and rural/urban area, by hospital



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 24
Breast Cancer: Setting Priorities for Effectiveness Research Key Patient Management Topics for Effectiveness Research in Breast Cancer PRELIMINARY DISCUSSION AND SELECTION OF MAJOR TOPICS Using the results of the first round of a homework exercise and the issues raised by the presentation of preliminary HCFA data analyses (see Appendix A), the committee discussed the range and clinical depth of breast cancer studies that might be mounted through an effectiveness research program. This session made clear the enormous number and complexity of unanswered questions about the appropriate management of women who are at risk for, who have, and who have had breast cancer. Many of these questions still lie in the realm of efficacy rather than effectiveness—that is, whether a test, a procedure, or a drug works at all, rather than how well it works in the average practice of medicine. To the extent that is true, the issues belong more in the arena of clinical and biomedical research of the sort conducted by the National Institutes of Health than in the area of effectiveness research. We tried to narrow our task to that of clarifying what set of questions could be satisfactorily addressed in one of three ways: (1) with existing HCFA administrative data of the sort on which the preliminary analyses presented at the workshop were based; (2) with existing HCFA data augmented by clinical data from medical records abstracted by PROs or from specially conducted patient surveys and follow-up studies; and (3) by carefully planned longitudinal studies that combine all necessary inpatient and ambulatory information, such as stage of disease, type of surgical and/or adjuvant therapy, and quality-of-life outcomes. To this end, we generated numerous topics on which the opening discussions focused. Some topics cut across several issues, such as the need to examine patterns of care by geographic and rural/urban area, by hospital

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research and type of practitioner, and by type of breast cancer case (e.g., identified solely by screening or as presented clinically). Other topics were based on specific classes of patients, such as those patients with node-negative disease. Throughout the discussion were expressions of concern about the relative lack in the HCFA database of necessary clinical and therapeutic data and how those deficiencies might be overcome through effectiveness research. Eighteen specific study topics were advanced by committee members and recorded during the early part of the discussion; they are shown, in somewhat condensed form, in Table 4. Some relate to clinical issues involving the use of diagnostic or therapeutic modalities (e.g., mammography, surgical options); others relate more to research methods (e.g., outcomes measurement) or cross-cutting topics (e.g., age, node-negative disease). After discussion, the committee members voted for five topics that should receive the highest priority attention in effectiveness research on breast cancer. These were tabulated and the highest ranking topics formed the basis for the remaining part of the discussion. The following section elaborates this discussion and our recommendations. SUMMARY OF RECOMMENDATIONS This workshop discussion did not propose testable hypotheses or specific research questions. Rather, it was meant to arrive at a broader set of issues, the exploration of which would depend critically on the data that might be available in the near term (e.g., through the Medicare files) or that could be collected relatively efficiently by independent effectiveness research projects. Precision about topics is limited by the information that is available or could reasonably be seen to be become available. Even with this caveat, it became clear that a large number of useful questions might be addressed, and we discuss here the major groups of questions that we concluded an effectiveness research agenda might include. First, we recommend separate attention to four nonclinical or methods topics: continued improvement and expansion of the database to serve the goals of effectiveness research; patterns of care and variations in those patterns; definitions and patterns of outcomes, with particular attention to dimensions not addressed yet by HCFA data or analysis, such as psychosocial, functional, and quality-of-life parameters; and other general issues of methodology, including indices of severity, case mix, and staging.

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research TABLE 4 Key Patient Management Topics: Round Two of Committee Ratings PREVENTION   Risk factors, e.g., body weight and diet in postmenopausal (not premenopausal) women Health profiles Methods PREDISPOSING DRUGS AND HORMONES   Use of medications common among the elderly, such as nonsteroidal anti-inflammatory agents, diuretics, and anti-depressive agents Use of post-menopausal hormones and their possible predisposing role MAMMOGRAPHY (Its role, significance, impact, and utility in the management of breast cancer)   Variations in use (e.g., by geographic region; provider) Variations in how results are used (e.g., by geographic region; provider) Accuracy in community practice in terms of true-positive, false-positive, and false-negative rates Interactions with biopsy (e.g., when does mammography detect an unsuspected abnormality and lead to biopsy and when is it a part of a standard workup for a palpable mass) Characteristics of test results (e.g., false-positive/true-positive rates, false-negative rates; predictive values) Breast self-examination and clinical examination in relation to mammography Outcomes of screening mammography in terms of disease detection, prevention of advanced cancer, prevention of death, improvement in quality of life, and cost-effectiveness THERAPEUTIC ALTERNATIVES: RADIATIONa THERAPEUTIC ALTERNATIVES: CHEMOTHERAPYa THERAPEUTIC ALTERNATIVES: SURGERY OPTIONSa THERAPEUTIC ALTERNATIVES: HORMONAL TREATMENTa RECONSTRUCTIVE (COSMETIC) SURGERY FOLLOW-UP CARE   Relationship of follow-up tests and age, stage of disease Role of bone scan Frequency of monitoring tests Psychological impact of follow-up PALLIATIVE CARE PATIENT EDUCATION   Example: evaluate mailing educational materials in Social Security checks PATTERNS OF CARE   Understanding patterns of use of the whole range of primary therapy alternatives (e.g., surgery options, radiation, chemotherapy)   Elements in analysis:       geographic area (e.g., is the common wisdom true that patients on the East and West coasts are more likely to receive breast conservation and reconstruction than are those in the interior of the country?)

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research       type of provider (e.g., hospital) and practitioner (e.g., specialty of physician) type of case as characterized by how identified (e.g., screening versus incident case [palpable mass]) type of patient in terms of age and in terms of risk factors OUTCOMES   Definition of “best outcome” (e.g., survival versus quality of life) Development of methodology applicable to HCFA data sets and/or to data that HCFA can collect Dimensions of outcomes include: mortality, morbidity, health status, and quality of life Use of self-report measures STAGING, CASE-MIX, AND SEVERITY OF ILLNESS   Comorbidity Pathologic type Receptor status AGE (As an independent, universal issue)   Methods of age-adjustment NODE-NEGATIVE CANCER (As a factor that cuts across staging and treatment issues) aQuestions relating to patient characteristics and a broad range of outcomes were raised in all of these topics. All four issues are relevant for many of the research strategies that might be adopted to address the three clinical topics that are recommended next. Points relating to the improvement of HCFA data files and collection of study-specific data are made throughout this report, and they are not discussed further below. The last three topics (patterns of use, outcomes, and staging) are discussed further in the subsection on “Methods Issues,” below. Second, we recommend three clinical or patient management topics for initial attention in the effectiveness research program targeted on breast cancer: screening and mammography; therapeutic alternatives for primary cancer, particularly as a function of age and stage of disease and of provider characteristics; and approaches to follow-up monitoring and care (including therapy for recurrent disease). Other clinical or management topics receiving support from the committee voting on priority topics included palliative care, patient education, and drugs. The three main topics selected for priority work are discussed more fully in the subsection on “Patient Management Topics.” Finally, we concluded that, for breast cancer, age is a major cross-cutting issue. This point is briefly discussed in the subsection entitled “Age.”

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research METHODS ISSUES Patterns of Care Analysis of how patterns of care change over time and as a function of numerous epidemiologic, clinical, and health care system factors is a major dimension of effectiveness research. This is true partly because of the extraordinary attention being placed on evidence of three problems: wide variations in population-based rates of use of services, high levels of use of inappropriate services and procedures, and differences in practice styles that appear to reflect areas of professional confusion and disagreement. The committee fully supports the use of the HCFA databases, augmented as appropriate by information obtained from patient records, in this analytic effort. Many committee members argued that the first stage of this research strategy for breast cancer must be to analyze patterns of care, partly for the policy and epidemiologic information such work would provide and partly as the necessary background for designing good follow-on studies. Thus, we recommend that an effectiveness research program continue to support work on patterns of care, and that such projects be linked expressly to the three key clinical topics identified (screening and mammography, therapeutic alternatives, and posttreatment monitoring). With respect to breast cancer, the variables and factors discussed below should be a part of a comprehensive set of analyses of patterns of care and variations. Geographic region (even at the level of major census divisions) may prove to be a crucial element in understanding which women can receive alternative types of care for breast cancer. Note should also be taken of SEER registry areas, and coordination of data undertaken. Small-area analysis should also provide useful information on differences in patterns of care. Differences by institutional providers and by professional characteristics of physicians should also be studied, as they may shed light on differences in organizational factors or professional training and practice styles that could be used to explain or to change performance patterns. Institutional membership in cooperative clinical groups and involvement with a Community Cancer Oncology Program (CCOP) should be noted. Physician status as specialist (and type of specialist) or generalist should be explicitly included. There is a need to distinguish asymptomatic breast cancer cases identified on routine screening from those identified because the woman sought care for a problem directly related to or pathognomonic for breast cancer. Although we note this point here in relation to patterns of care, the possible importance of this distinction should be recognized for analyzing other breast cancer issues. For instance, mass screening efforts in their early years will

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research detect a considerable number of slow-growing cancers that have been present for some time. These have good prognosis, so treatments given during the first few years of any new mass screening effort will appear to be better than they really are. Once these “existing” prevalent cancers are dealt with and routine screening only detects absolutely new cases, the treatment has its real test. Outcomes We recommend that effectiveness research deal with three related issues concerning outcomes. The first is the need for more comprehensive definitions of outcomes in line with recent developments in the field of health status and quality-of-life measurement. The second calls for better definitions of alternative “best outcomes” (from the patient’s point of view), which take health status and quality-of-life dimensions more fully into account. Third is the development of better techniques for acquiring a full spectrum of outcomes data related to breast cancer. More generally, HCFA and other DHHS agencies should give considerable priority to understanding and developing these methods for use throughout the effectiveness research program, not only for breast cancer. Several committee members asked what constitutes a “best outcome.” They noted that the longest possible survival may not be regarded by some patients (or their physicians) as a best outcome; to the extent this is so, death or length of survival is an important, but incomplete, indicator of outcome. (In this context, mortality is to be understood as referring to death from breast cancer.) Furthermore, breast cancer can be an indolent disease; the absence of survival differences at one point in time does not guarantee similar results later. For all patients, measures of morbidity (including pain and other symptoms), functional status (including physical capacity and ability to function in daily life), psychological and emotional well-being, social functioning and support networks, and general outlook on health are all important. For many patients, these may be more important than survival itself. Thus, it is important to differentiate outcomes from patient preferences for outcomes and to encourage the acquisition and use of information on patient preferences, given different outcomes. The explicit trade-off of survival and other measures of outcome may be most pertinent in Stage IV disease, although it arises with respect to all stages of disease. Other clinical outcomes, such as disease-free survival, contralateral breast cancer, or other non-breast-cancer malignancy are additional, but incomplete, measures of outcome that apply to women with all stages of breast cancer. The functioning and quality-of-life measures, too, apply to women across the entire range of breast cancer. Apart from their

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research use in studies that involve direct outcomes of therapy, such measures are needed to create decision-analytic models (which require transition probabilities between health states) to identify optimal management strategies. In general, the committee judged that these dimensions of health status should be viewed as independent of the need for or use of health care, although measures based on utilization could form part of more comprehensive set of outcome measures. In addition, considerable skepticism was expressed about the reliability, validity, and overall value of using the Karnofsky Index (a cancer-specifc, physician-rated measure of physical status and physical activities) in this capacity, despite its application over the years in carefully designed cancer studies. Thus, we recommend that agencies involved in effectiveness research solicit outside expert opinion to define and select an adequate, appropriate set of outcome measures (other than mortality) and instruments to be considered for use in this research effort generally and for use explicitly for breast cancer. In this regard, we draw attention to the emerging consensus that, for health status and quality-of-life measurement, use of reliable and valid generic measures of health status coupled with selected disease-specific measures is an appropriate, desirable, and practical research strategy. The existing HCFA databases lack any measures of functional status (although HCFA is attempting to add a functional status measure to its claims forms). In the short run at least (or in the event that the measure selected for the claims form is inadequate), this information would have to be obtained directly from patients (or, perhaps, proxies). Information about patient preferences is also critical to many analyses of breast cancer questions; almost certainly this type of information will have to be sought from patients. In sum, the application of patient-based measures of acceptable reliability and validity must be an integral part of the effectiveness research program over the longer term. The committee noted the importance of specific subtopics relating to outcomes. One is the lack of knowledge about complications of surgery, especially in the Medicare population, whether those complications vary by geographic region, by provider, or by type of surgery, and whether the use of chemotherapy or tamoxifen is related to complications in any way. Some of these analyses might be done immediately, such as studies focused on complications among Medicare patients during hospitalization or within the first 30 days of hospitalization or surgery; important disease-specific complications include wound infections, impaired wound healing or skin flap loss, and excessive arm edema. A second subtopic involving complications concerned toxicity of standard therapies (radiation and chemotherapy) in elderly women. These would not lend themselves as readily to analysis of inpatient data, but because of the relative lack of studies of toxicity (outside narrowly targeted RCTs), the

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research committee believed this would be an important adjunct to descriptive effectiveness research. A third subtopic concerned the importance of quality-of-life outcomes for women with recurrent breast disease. For patients with recurrent breast cancer, most therapeutic approaches will not influence survival time or rate very significantly, if at all. Consequently, the importance and impact on the patient of early detection of recurrent disease and the relatively greater importance of psychological support for this patient subgroup should be investigated. Staging, Case-Mix, and Disease Severity We recommend specific attention to the gathering of a minimum set of appropriate clinical data for a standard severity-of-illness or staging index that can be used to adjust for stage of breast cancer disease. Effectiveness studies will depend critically on clinical and pathological disease staging. Staging is important in prognosis, treatment options, and physician and patient decision making about therapy. It also plays a significant role in characterizing the patient populations being studied. We came to less agreement about the key pieces of information that belong in such an index. Among the elements on which we reached relative consensus were the following: comorbidity (e.g., coexisting diagnoses of the sort that can be derived from Medicare Part A files), pathological stage and type of disease, and hormone receptor status. Other possible components of a staging index were mentioned, but they were recognized as very new and perhaps controversial techniques not accessible for incorporation into HCFA databases or effectiveness studies. PATIENT MANAGEMENT TOPICS Screening and Mammography We recommend that agencies involved with effectiveness research begin preliminary analyses on mammography and screening while undertaking a detailed planning effort to design a comprehensive, prospective, longitudinal research program focused on screening mammography. In the workshop discussion, committee members raised numerous questions about mammography itself, about widespread use of mammography as a screening technique, and about the downstream consequences and outcomes of screening mammography. Thus, we identified mammography as an important screening and diagnostic maneuver that offers unique opportunities for prospective analysis. Related topics were the role of breast self-examination and physician examination (as a screening modality) and

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research the utility of other imaging modalities as a substitute for biopsy following a positive mammogram. Effectiveness of Mammography Medicare will begin reimbursement for screening mammography in October 1990 (assuming this part of the Medicare Catastrophic Coverage Act is not repealed).11 Before that time, a prospective plan could be developed for evaluating the effectiveness of mammography as a screening as well as a diagnostic tool. That plan could be initiated before October 1990 with analyses of patterns of use of mammography generally and some preliminary patient follow-up studies. These preliminary studies would be used to provide background data to highlight the key questions to be addressed in later studies (perhaps including the generation of testable hypotheses), and to develop and pretest instruments and methods for medical record abstraction, for patient follow-up, and for outcome and quality-of-life studies. In the meantime, a comprehensive request for proposals or grant solicitation should be developed. For design of patient outcomes analysis, a full range of reliable and valid health status and quality-of-life measures must be considered; outside expert opinion should be sought about good candidate measures for this purpose. In addition, this plan should incorporate information and methods available from the SEER registries and the National Cancer Institute CCOPs. Community Versus Medical Center Practices Analogous to the issues of efficacy and effectiveness are questions about how well mammograms are conducted and interpreted as the sites and practitioners performing them proliferate (e.g., from radiology departments in tertiary hospitals to the offices of private physicians).12 Can reliability and other standards of the test and its interpretation be maintained at the levels in academic medical centers as the examination disseminates into community practices or clinics? This will be an especially important question with the advent of screening mammography as a Medicare benefit, because poor 11   As noted earlier, the act was repealed in November 1989. 12   The interpretation of mammograms is related not only to the quality of the films produced but also to the expertise of the diagnostic radiologist reading the films. The American College of Radiology has established a voluntary accreditation program for mammography; it gives a comprehensive assessment of mammographic services, including personnel qualifications, image quality, and radiation dosage. This may be one way of indirectly assessing the reliability and validity of the mammograms.

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research mammography on a wide scale could be an economic, clinical, and personal disaster. Interpreting Mammograms Criteria for interpreting mammograms and reaching diagnostic or therapeutic decisions are very important, although attacking this topic lies outside the direct reach of effectiveness research. Nevertheless, a strategy for investigating the use of screening and diagnostic mammography requires that the pertinent clinical controversies be recognized. Several relate to the interpretation of mammograms and the relationship of mammograms to biopsy. First, screening mammograms are alleged to have a high false-positive rate. To the degree that this is true and that screening examination leads to biopsy, questions arise as to the costs and benefits of those biopsies. Moreover, very little is known about the predictive value of various mammographic abnormalities in patients given various sets of clinical features or about the predictive value of mammograms given various combinations of mammographic features. A second controversy relates to biopsy of “precancerous lesions,” perhaps secondary to a false-positive mammogram. That is, a mammogram may not have accurately detected malignant disease, but it may have detected atypical or proliferative lesions that may, in time, became malignant tumors. Some clinicians believe that these lesions warrant excision; others disagree. Although many of these lesions may indeed be “premalignant,” the biopsy often samples only a small portion of a condition that may be widespread in the breast; therefore, excision of only the mammographic abnormality may do little to prevent subsequent development of breast cancer in that woman. The biopsy results may be used, however, to determine the risk of the patient for breast cancer; for example, severe atypical hyperplasia at biopsy increases a woman’s risk of later breast cancer. Third, the clinical circumstances of mammogram and biopsy are important. One distinction is whether the mammogram is part of a standard workup in a symptomatic woman or has been done in a routine screening program. These factors should be clear before conclusions are drawn about the temporal relationship of a mammogram and a biopsy. Fourth, appropriate biopsy techniques following a positive mammogram warrant attention; sorting out the alternative techniques for biopsy (such as fine-needle aspiration biopsy) will become an issue in and of itself. Finally, if diagnostic criteria for breast cancer are changing over time, it will be important to know what impact those changes have on use of biopsy (and related disease staging issues), use of mammography, and ultimately choice of therapy.

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research Mammography with No Subsequent Care Related to Breast Cancer Several committee members raised the question of follow-up of screening mammograms—specifically, how to interpret the information that a woman had a mammogram (presumably reimbursed) but no further, immediate workup. Hence, in correctly interpreting claims data on screening mammograms, it will be important to determine the numbers and outcomes of patients in three categories. First are mammograms read as negative that were truly negative. Second are mammograms read as negative that were truly positive. Third are mammograms read as positive with no further workup or treatment performed because (a) the patient refused, (b) the patient’s condition precluded further evaluation or treatment, (c) the patient and/or physician were never notified of results, (d) the physician did not follow up the results, or (e) the patient was lost to follow-up. Each of these possibilities has different implications for interpreting effectiveness studies. The committee recognized the difficulty of pursuing that line of questioning with the HCFA databases or even with targeted research, but it believed the issue was important enough to warrant some research effort by HCFA or others. Treatment Alternatives We recommend that effectiveness research be conducted in several categories of primary care. Categories of primary treatments include those listed just below. Mastectomy—simple (total), modified radical (extended simple), radical, and extended radical; Lumpectomy (wide excision) with or without axillary node dissection; Lumpectomy (wide excision) with or without axillary node dissection and radiation; Post-mastectomy radiation; Adjuvant chemotherapy in node-positive or node-negative disease; Adjuvant hormonal therapy; and Primary treatment of intraductal (noninvasive) breast cancer. The focus should be on patterns of use and outcomes, as discussed above. We advise further that HCFA and other federal agencies use the HCFA databases and other resources to track the emergence of other therapies or combinations of therapies that might warrant closer attention in the future. As with screening mammography, some issues can be addressed immediately with a combination of existing HCFA databases and medical record

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research abstraction procedures (of both inpatient and outpatient records). For instance: What are the patterns of use of different surgical options by geographic area, by type of provider (e.g., type of hospital), and by type of patient (especially by age and, if medical record data are available, by stage)? What are trends over time in use of different surgical options (again by these types of variables)? What proportions of patients receiving lumpectomy also receive axillary node dissection? Also receive radiation? What approaches are being used for the care of intraductal carcinoma? What proportions of women treated for breast cancer also receive care for problems related to emotional health (e.g., formal mental health services for depression)? Other questions call for descriptive, case-control, or longitudinal studies. These should incorporate a wide range of patient outcomes and examine practice behaviors of physicians and other providers more closely. We specifically mean to include health status and quality-of-life measures in this concept, even if they require that information be obtained directly from patients in addition to that obtained indirectly through information in medical records. For instance: What are the short-term (e.g., posthospitalization) and long-term (e.g., one year; five years) outcomes of women whose breast cancer is managed in different ways (e.g., by different surgical therapies, with/without radiation, with/without other forms of adjuvant therapy)? What are the same types of outcomes for women whose breast cancer is detected in different ways (e.g., by screening versus as a “clinical incident” case)? Does the recommendation of therapy differ among providers according to expected outcomes for patients of different characteristics (e.g., stage of disease, age, preferences and attitudes toward quantity and quality of life)? Would provision of information about expected outcomes, perhaps for different categories of patients, change (or improve) the recommendations that physicians make about therapy? Does the choice of therapy differ among women according to personal, sociodemographic, or clinical characteristics? Two other points should be noted. First, the questions above by no means exhaust the set that might, over the years, be addressed through a

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research program of effectiveness research; they are solely illustrative. Second, all comparative studies in these categories—for example, those comparing rates of use across geographic areas or outcomes across patient subgroups and drawing health care policy decisions from them—will require adequate control of or adjustment for case mix and stage of disease. The committee emphasized the need for adequate data on the ambulatory management of breast cancer before any longitudinal study of alternative therapies can be undertaken. The difficulties of analyzing effectiveness of alternative therapies using only hospital records cannot be overstated. Chemotherapy and radiation, as well as use of tamoxifen, may well be initiated after hospital discharge. Adequate appreciation of adjuvant and drug therapy requires some knowledge of dosage, drug combinations, and the like; such information will not be in claims-based files or even in hospital records. Furthermore, in the case of tamoxifen administered in the ambulatory setting, HCFA claims data would not be available until 1991 (assuming an outpatient drug benefit mandated by the Medicare Catastrophic Coverage Act goes into effect).13 Thus, we believe that any outcome or effectiveness studies lacking accurate data of this sort will be flawed. For this reason, we believe that adequate effectiveness analyses on treatment options, using the basic Medicare claims files and other parts of the Medicare Decision Support System, will necessitate database development to meets three requirements. First, the data files should be able to identify specific drugs used for treatment of each patient (or this information must be obtained in some other way). Second, an adequate strategy should be devised to use medical record information to stratify the risk of breast cancer, including at a minimum comorbidity, pathological stage and pathological type, and hormone receptor status. Third, review of physician records in the ambulatory setting might be done by Medicare PROs or others (i.e., supplementary to whatever abstracting is done of inpatient records for the Uniform Clinical Data Set). This is in addition to investigations that collect project-specific information on patients or providers. Studies to compare outcomes of therapeutic alternatives should not be carried out until these conditions are satisfied. Although we recognize this will probably preclude their implementation before the end of calendar 1990 (for instance, until data on medications can reliably be collected), we believe that the intervening period can be used profitably in two ways. First, researchers can begin to investigate patterns of care for those therapies that can presently be adequately tracked by the existing HCFA data systems. Second, research agencies can mount a planning activity that will result in appropriate research designs for studies to be started as the means for acquiring the ambulatory data mentioned above come into being. 13   As noted elsewhere, this act was repealed as this report was being prepared.

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research Follow-Up Care and Protocols We recommend that planning begin for a series of research studies that would address several key questions about follow-up care and protocols for women initially treated successfully for primary disease. Continued monitoring of women treated for breast cancer is a major area of concern because of clinical controversy over what tests and other interventions should be used in standard follow-up protocols, with what frequency, and over what time period. Studies on this topic also provide a introduction into issues relating to the diagnosis and treatment of patients with recurrent breast cancer. Among the questions warranting earliest attention are: What is the incidence and prevalence of primary breast cancer and what pool of Medicare beneficiaries does that yield as a potential universe of patients for a follow-up study? What are appropriate elements of a follow-up protocol (e.g., laboratory work, chest X-rays, repeat receptor studies, contralateral mammographic studies, bone scans), particularly the role of bone scans and mammography? Does early detection of recurrence improve treatment and outcome? Do different stages of primary disease warrant different treatment and follow-up protocols, and if so, how? Do different types and combinations of treatment warrant different follow-up protocols, and if so, how? How do the results of previous monitoring tests affect the intensity of future follow-up; that is, do all negative results lead to less intense follow-up in the future? What is the cost-effectiveness of different approaches to follow-up? Such projects could be started with existing databases (e.g., those relating to prevalence of primary disease) and expanded with primary data collection as needed (e.g., for studies concerning early detection of recurrence). AGE We have repeatedly stated the importance of analyzing information in terms of the patient’s age. This issue deserves emphasis because of the emerging evidence that older women receive different types of therapy than younger women with no apparent epidemiologic, clinical, or demographic justification. Women who are 65 or older may have relatively more advanced stages of breast cancer when they first appear for treatment, and

OCR for page 24
Breast Cancer: Setting Priorities for Effectiveness Research they may not be treated according to the usual and customary alternatives provided to the rest of the population. Certain women may receive no definitive treatment at all, but whether this is related to age is unknown. We believe innovative ways to investigate this issue should be developed, perhaps by supporting research efforts that would use insurance claims files for non-Medicare populations in comparison with those for Medicare.