Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Key Patient Management Topics for Effectiveness Research in Hip Fracture PRELIMINARY DISCUSSION AND SELECTION OF MAJOR TOPICS The high priority patient management topics nominated by the commit- tee- primely prevention, treatment selection, and rehabilitation remained essentially unchanged from a homework exercise conducted before the workshop (see the appendix). Specific study topics within each group, how- ever, were quite broad in scope. In addition, prevention and management of secondary complications of hip fractures were noted by a few committee members as patient management topics suitable for effectiveness research. Issues of cost, cost-effectiveness, and cost-efficiency were noted with some frequency across all categories. In the second round of voting, a consen- sus by the committee to focus on prevention of hip fracture, selection of treatment alternatives, and rehabilitation was virtually unanimous. Bible N2 of the appendix summarizes the topics listed by the commit- tee in the preworkshop homework exercise. Methodologic and data issues related to the study of these topics are discussed in the next section of this report. SUMMARY OF RECOMMENDATIONS The workshop discussion did not propose testable hypotheses or spe- cific research questions. Rather, it was meant to arrive at a broader set of issues whose exploration would depend critically on the data that might be available through the Medicare files or that could be collected by indepen- dent effectiveness research projects. Even with this limitation, many useful subjects might be addressed, and the major questions that the committee thought belonged in an effectiveness research agenda are discussed here. 25
26 EFFECTIVENESS INITIATIVE The committee made recommendations that were specific to HCFAs existing capabilities as well as those related to the larger, evolving DHHS program of effectiveness and outcomes research. The administrative data- bases, by themselves, are limited in the type of effectiveness research ques- tions that they can address, but they provide some unique opportunities when combined with other data sources and research efforts. The committee recommended that explicit attention be given to four methodologic issues in hip fracture effectiveness research:) 1. health and functional status assessment; 2. definitions of outcomes and comparability of outcome measures; 3. the need for longitudinal data; and 4. risk stratification and classification by comorbidity. A methodologic issue that was not addressed by the hip fracture committee explicitly but was judged by the core committee to be integral to effectiveness research was: 5. patterns of care and variations in those patterns. The committee also recommends three clinical or patient management topics for initial study in an effectiveness research program: 1. prevention of hip fracture; 2. treatment options for hip fracture, particularly as a function of age, frailty, comorbidities, and provider characteristics; and 3. rehabilitation, particularly as a function of type and intensity of service, site of care, and provider characteristics. Finally, the committee noted that age cuts across all effectiveness and outcomes research affecting elderly people. In hip fracture, as with many other clinical conditions, age and the degree of frailty have profound implications for treatment choice and outcome. These issues are discussed below and mentioned throughout the report. METHODS ISSUES Health and Functional Status Assessment Recent work in measuring health status and functional capabilities is quite extensive. There is little consensus, however, as to formats that should be adopted by a government agency or by private researchers to measure iAll four methodologic issues noted in this report pertain as well to breast cancer and acute myocardial infarction research strategies, and they are discussed in the companion reports with modifications pertinent to the diseases under consideration.
HIP FRACTURE 27 either prognostic indicators or outcomes in effectiveness research. Many current measures are too long and are not specific to the clinical condition under consideration. ~ move forward in effectiveness research, the recent work on health and functional status assessment must be synthesized; the committee cau- tioned against "reinventing the wheel" in this area. The use of "generic" health status measures (e.g., the Sickness Impact Profile, the Visual Ana- logue Pain rating scales, or the Index of Independence in Activities of Daily Living), augmented by selected "disease-specific" measures (e.g., the Arthritis Impact Measurement Scale), is increasingly recognized as an ap- propriate research strategy (see the citations in the bibliography). The committee noted, for example, that the concept of frailty must be incor- porated into health and functional status assessments of patients with hip fractures. The committee supports funding of methodologic research that would construct a geriatric assessment instrument that includes the concept of frailty. Effectiveness research on hip fracture will require that baseline and follow-up data on health status be collected for several purposes: prognos- tic risk stratification, physician and patient decision making about treatment options, defining and understanding short- and long-term outcomes, and appropriate characterization of the patient populations being studied. Stan- dard patient-based measures of acceptable reliability and validity should, in the long run, become an integral part of the effectiveness research program. Regarding data sources, much health and functional status information will have to be obtained directly from patients (or their proxies); existing HCFA databases lack measures of functional status.2 The committee rec- ommends that HCFA develop health and functional status measures that can be linked longitudinally to the Medicare Part A and Part B files and to medical record data. Health and functional status information is needed on a broad population base and should be assessed periodically. Studies must be funded to determine the most effective way to implement such questionnaires to obtain consistent, reliable, and valid data. Eventually, this database could serve as a source of information for predicting future events (i.e., determining risk) and for measuring response to interventions. The committee further recommends that HCFA consider introducing a health assessment mechanism or questionnaire to patients on their entry to the Medicare system before any illness even for a limited sample of 2At the time of the workshop, HCFA was attempting to add a physician-based functional sta- tus measure to the Medicare insurance claims forms. Depending on the adequacy of any such measure, appropriate information for hip fracture (or for the other conditions recommended for the Effectiveness Initiative) may still need to be obtained directly. Presumably the reliability, validity, and cost-effectiveness of adding functional status information to the claims forms will be evaluated, but these issues were not addressed by the committee.
28 EFFECTIVENESS INITL~4T~E beneficiaries.3 As implied above, this additional information should include physiologic parameters and risk factors as well as the health and functional concepts noted above. This registry, if implemented over time, could reveal information about change in health status in the Medicare population as people age and the prevalence of chronic disease increases. Ultimately, a component of this health status assessment would de- scribe very high risk individuals. These individuals are called frail for the purposes of hip fracture, but other categories might emerge from population-based health status assessments. Such a data set would be valu- able to effectiveness and outcomes research across many clinical conditions important in the Medicare population. It may also provide insights into clinical issues that emerge in late middle age and that ultimately impinge heavily on the Medicare program. ~ A ~ It was beyond the charge of this committee to specify components of a geriatric health and functional status assessment instrument suitable for hip fracture research. In general, however, the committee believed that this assessment instrument should include components of what has classi- cally been called functional capabilities (including Activities of Daily Living [AD Ls] and Instrumental ADLs [IADLs]), cognitive functioning' sociode- mographic variables, physiologic parameters, and prognostic indicators. A comprehensive assessment incorporating these various perspectives might aid in defining a category of people considered to be frail. This frailty index could be factored into treatment decision making. Incorporating a frailty dimension into assessment of elderly people is one aspect of recognizing the heterogeneity of the elderly population for purposes of research and patient care. Because hip fracture seems to be so intimately tied to the development of mobility and cognitive problems that come under the rubric of frail~, this clinical condition can potentially be used to understand and perhaps separate disease processes from the process of aging. The interest, energy, and financial resources that are focused on the study of hip fracture will therefore carry over to effectiveness research of other clinical conditions. Outcomes The committee recommends that HCFA focus on two outcomes issues, (1) a more comprehensive definition of outcomes, in line with recent developments in health status assessment and quality-of-life measurement, 3At the time of the workshop, very preliminary plans were being discussed at HCFA about de- veloping a Medicare "registry" perhaps involving a 5 percent sample of newly enrolled Medicare beneficiaries. Ibis recommendation was meant to support further exploration of this idea.
HIP FRACTURE 29 and (2) techniques for acquiring data that cross the full spectrum of care delivery sites. Mortality rates are often used as outcome measures. Although the mortality rate at one year for hip fracture is substantial, mortality alone is not the only appropriate outcome measure for hip fracture. The 30~ay mortality rate for hip fracture, for instance, is 7 percent and does not reflect the often rapid decrease in functional and health status that follows injury. Various clinical and surgical outcomes are important, such as stability of the joint, healing of the fracture, and prevention of acetabular erosion. The committee wishes to emphasize that for hip fracture (more than for breast cancer or acute myocardial infarction), health and functional status assessment before the adverse clinical event is necessary to interpret these and other outcome measures appropriately. For all patients, measures of morbidity (including pain and other symp- toms), 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 important. Adequate ascertainment of these health and functional status levels should affect the choice of surgical procedure (or, in a subset of frail elderly people, the de- cision to treat the fracture nonsurgically) and may determine the type and site of rehabilitative services needed and offered. Thus, it is important to differentiate between outcomes and patient preferences for outcomes and to encourage the acquisition and use of information on patient preferences for different outcomes. The committee recommends that HCFA continue to seek expert as- sistance to develop outcome measures other than mortality, including guidance for selecting the instruments for general effectiveness use as well as for hip fracture. This committee draws special attention to the emerging evidence that the use of "generic" health status measures augmented by selected "disease-specific" measures is an appropriate, desirable, and prac- tical research strategy. The committee does not, however, advocate that a single outcome measure (or even a single set of measures) be mandated for effectiveness research; the goal is to obtain comparable information across studies by using measures whose elements can be mapped to one another. For hip fracture, a definition of poor outcome would include several variables: increased length of hospital stay (not explained by provision of comprehensive rehabilitation services), increased rates of institutional placement, continued pain, lower health and functional status along sev- eral dimensions (mobility and physical activity, mental distress, reduced social interaction, inability to conduct activities of daily living), and pre- mature mortality. Short- and long-term outcomes for patients with hip fractures (controlling for the type of surgical procedure and comorbidity at admission) must be determined.
30 EFFECTIVENESS INITLi4TIVE One outcomes assessment strategy, cited by the committee as a poten- tial model for a multifaceted assessment, is Ellwood's Outcomes Manage- ment approach. HCFA has funded Quality Quest to work with the PROs in three states to assess systematically outcomes of large groups of peo- ple using information in addition to billing data. The research instrument includes measures of functional status, patient satisfaction, and other indi- cators of health status. The committee recommends that HCFA continue to support this kind of effort. Longitudinal Follow~up The committee recommends that hip fracture outcomes be measured longitudinally, for instance, at the time of hospital discharge and at six weeks, six months, and one year after discharge. As noted earlier, pre- fracture health status should also be obtained. Outcomes as a function of living arrangements should, for example, be examined in detail to deter- mine the variables that permit individuals to return to the community (e.g., utilization of a skilled nursing facility, living with family or a paid caretaker, Medicare home health service). Short-, mid-, and long-term outcomes can be contrasted with prefecture status and correlated with type of treatment, length of stay, and type, site, and intensity of rehabilitation. Research currently being conducted on hip fracture outcomes by in- vestigators at the University of Minnesota School of Public Health was identified as an example of a follow-up study that will yield health and func- tional status information across various settings. In that study, prehospital functioning was assessed retrospectively; functional status was measured at the time of discharge and at two weeks, six weeks, six months, and one year after discharge. Patients are followed as they move from hospital to rehabilitation facility to nursing home to a community setting, sometimes all within the span of six weeks. Risk Stratification and Classification of Comorbidities The committee recommends that case mix indices for hip fracture be developed (including a measure of prefecture frailty, to predict, for example, cost, mortality, and ADL functioning sur months postfracture. Attention should be given to the minimum clinical data needed and the procedures for acquiring such data. The importance of good risk stratifi- cation is obvious; a vigorous 70-year-old tennis player who breaks a hip, for instance, is in a vein different category than a 70-year-old resident of a nursing home. Categorizing patients in terms of prognosis becomes very critical in un- derstanding the effectiveness of treatment. Risk stratification, for example,
HIP P~C=~ 31 may suggest that the intervention that will achieve the "best" outcome for some patients is nonsurgical management. This difficult ethical question can only be addressed satisfactorily with accurate risk stratification. Again, for HCFA, this type of analysis will depend on linking administrative data to primary data collection by the PROs or other research efforts. Information about the presence of comorbid conditions is also im- portant, such as kidney failure secondary to atherosclerotic cardiovascular disease, diabetes with severe peripheral vascular disease, or stroke with cognitive impairment or necrologic conditions. Knowing the medications the patient was taking before the hospitalization can also contribute signif- icantly to the risk stratification and classification process. A scale could be developed that would enable researchers to stratify (or grade) the patient population into various classes resecting their baseline functional status. If four or five different classes of patients were identified, more realistic outcome goals for each class of patient could be developed. Patterns of Care and Variations Analysis of the change of patterns of care over time and as a function of epidemiologic, clinical, and health care factors has emerged as a major dimension of effectiveness research. This is true because evidence has revealed (1) wide variations in rates of use of services (in hip fracture, pri- marily surgery and rehabilitation), (2) variations in use and site of services (especially for rehabilitative care), and (3) differences in practice styles that appear to reflect professional disagreement. The committee supports the use of the HCFA databases, augmented as appropriate by information from patient records, to analyze patterns of care for hip fracture. This could be done partly for policy and epidemiologic purposes and partly to help design good follow-up studies of groups of hip fracture patients. Patterns-of-care studies should be linked expressly to issues of preven- tion, treatment options, and rehabilitation. Several variables 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 who experiences hip fractures and what type of treatment option (e.g., surgical or nonsurgical; type of surgical interventions is chosen. Small-area analysis should also yield use- ful information on different patterns of care. Differences by institutional providers (e.g., types of hospitals; rehabilitation settings) and by professional characteristics of physicians and various types of therapists should also be studied, as they may shed light on differences in organizational factors or professional training and practice styles that might be used to explain or to change performance patterns. Other variables (e.g., prepaid versus
32 EFFECTIVENESS INITL24TIVE fee-for-service practice) may also prove to be important, especially for the frail elderly population. PATIENT MANAGEMENT ISSUES Prevention Prevention of hip fracture was identified as a priority patient manage- ment issue because of the high social costs and grim personal consequences for elderly victims in terms of pain, disability, functional limitations, and displacement from community living. Prevention is a very promising area of study. For the person 65 years of age, the incidence of hip fracture doubles every five to seven years, with the risk increasing 8-fold to 16-fold between the ages of 65 and 85. Even from the time that a person enters the Medicare system, certain preventive interventions may modify the risk. If the onset of a hip fracture could be delayed five years, the incidence of hip fracture could be reduced by as much as 50 percent, with considerable savings in patient suffering and costs. The committee identified osteoporosis and falls as priority targets for prevention research, because they are major risk factors for hip fracture. Prevention research for falls should include preventing falls per se and reducing fracture risk for falls that do occur. Studies on medications should also be pursued as corollary to understanding prevention issues. Osteoporosis Because of the intricate relationship of osteoporosis, bone strength, falls, and hip fracture, the committee recommends that osteoporosis be made a focus of prevention research. Several prevention strategies for this risk factor are probably most effective if they are implemented before the person enters the Medicare system (i.e., reaches age 65~. Consideration should be given to the impact that third party payer denial of preventive services before age 65 has on shifting costs of these sequelae after age 65 onto the public sector. One example of work that might be done in this area involves medi- cations that are believed to have a protective effect. Observational studies suggest that thiazide diuretics reduce the risk of hip fractures if they are used for a period of several years; this is consistent with their known effects on reducing calcium excretion. Replacement estrogens are protective for newly postmenopausal women; it is unknown if starting these medications for women age 65 and older will delay osteoporosis. Calcium supplements, calcitonin, biphosphonates, and coherence therapy are all potential drug treatments for osteoporosis whose efficacy for preventing hip fractures is
HIP FRACTURE 33 unknown.4 Life-style changes such as an increase in the level of physical activity may also preserve bone mass. There is an urgent need both for efficacy and effectiveness research to clarify the clinical utilibr of these potential preventive measures. Thus, because this topic is clinically, epidemiologically, and adminis- tratively important, we advise that those planning the effectiveness research programs reach beyond the Medicare eligible population and ask what studies can be done now for those 65 years of age and older and what can be initiated for those under 65. Falls Falls are the leading cause of death from injury in the United States for people age 65 and older. As noted earlier, approximately 260,000 hip fractures occur among the elderly each year, with an associated medical cost estimated at over $6 billion, and falls are clearly a major factor in this picture.5 The committee recommends that effectiveness research efforts ad- dress pathophysiologic processes, the primary aging processes, and behav- ioral and environmental factors associated with falls. Outcomes and the cost-effectiveness of various prevention strategies must be evaluated. The mechanisms of falls and their role in hip fracture etiology should also be addressed. Administrative databases can make only limited contributions to these questions, but certain uses should not be overlooked. For instance, such data might be used to evaluate programs to prevent falls, such as a demon- stration project comparing an intervention in one community (a city- or state-wide program in which benefits include assessment of the home en- vironment for hazards such as throw rugs on polished floors, inadequate lighting, and lack of handrails on stairs) and no intervention in another. Another potential contribution that the HCFA data might make Is to develop a risk profile for fall-related fractures. Differences in fracture related to race, body weight, and the like can be examined. The validity of this kind of analysis requires that the presence or absence of comorbidity be explicitly recognized. Although this type of analysis is not currently possible, diagnosis-related group (DRG) data will become much more useful if the codes account for the presence or absence of complications or comorbidities. This change, if implemented by HCFA, would help 4 Coherence therapy is a cyclic combination of therapies stimulating the rate of bone remodeling and depressing resorption. 5A relatively small proportion of falls in old people result in fracture (4-6 percent), but nearly all hip fractures are consequent to falls.
34 EFFECTIVENESS INlTLATIVE researchers obtain a clean sample of patients, that is, one with similar complications or comorbidities; this risk stratification would then permit useful comparisons. Medication Usage Effectiveness research should include work on medications thought to be helpful in preventing osteoporosis (and, hence, falls and fractures) and on those believed to be harmful.6 One example of the former thiazide diuretics- was noted earlier. Possibly more important, however, is improve- ment in the understanding of which medications (including combinations of medications) may foster osteoporosis (and indirectly increase the risk of fractures) and which may affect balance or cognition and, thus, more directly increase the risk of falls and fractures. Psychotropic medications are a class of drugs that constitute a signif- icant risk factor for falls and fractures. Certain psychotropic medications (e.g., hypnotic-an~olytics with long elimination half-lives, tricyclic antide- pressants, and antipsychotics) produce about a twofold increase in the risk for hip fracture. An estimated 20 percent of Medicare beneficiaries are cur- rently being prescribed psychotropic medications (although the Medicare program will not provide reimbursement for them when they are prescribed on an outpatient basis). The potential for effective prevention intervention by changing practice patterns and refining indications for drugs is great. Thus, clarification of how widely and how appropriately these agents are used and their role in falls or fractures is badly needed. Treatment Options The committee recommends that three questions be addressed con- cerning choices of treatment: (1) electiveness of selected surgical in- terventions, (2) appropriateness of nonsurgical (medical and supportive) management, and (3) other treatments that pose high costs of care. 6 Medicare does not cover outpatient prescription drugs at present. The Medicare Catastrophic Coverage Act, which would have instituted an outpatient drug provision, was repealed in late 1989. The committee noted, however, that the absence of outpatient medication information will limit effectiveness research that otherwise might be conducted using the Medicare insur- ance claims files. That is, HCFA databases currently offer little assistance in this area because Medicare does not cover outpatient prescription drugs. One possible way around this might be to link Medicare data to Medicaid data for those states that provide ambulatory medication in- formation for the subsample of Medicare patients dually covered by Medicaid (i.e., the poor and disabled elderly population). By and large, however, effectiveness research directed at issues involving medications, which for hip fracture may be fairly important, will have to be done based on data collected through means other than the HCFA files.
HIP FRACTURE Effectiveness of Surgical Interventions 35 The committee recommends that electiveness research on hip fracture treatment alternatives develop scientifically based indications to determine which patients should have the following: (1) open or closed reduction with internal fixation, (2) prosthetic replacement of the femoral head, or (3) primary total hip replacement. Nonsurgical Management Many- questions remain unanswered in the treatment of hip fracture for the oldest old and the frail elderly populations. Anecdotal experience suggests that a category of patients exists for whom standard care (i.e., surgical treatment) may not be indicated. Those who were minimally functional before the fracture often remain so afterward. They undergo a major operation at a very high cost (in physical, mental, and perhaps financial terms), yet they return to their previous functional status and face an extremely high probability of death within three to six months. Research is needed to determine whether there is a subset of patients who sustain hip fractures who are so disabled that the usual operative intervention offers little if any functional gain. Recommendations for the treatment of the frail elderly population, therefore, must be developed consistent with the purposes of treatment. These include return to previous function, control of pain, ease of nursing management at the site to which the patient will be discharged, and the speed with which mobilization can be accomplished. Current predictors of hip fracture treatment by nonsurgical means ap- pear to include: inability to walk three months before the fracture, low ADL status before admission, an impacted (i.e., compressed) fracture, and cancer of the hip.7 It is possible now to analyze the HCFA administrative databases (mainly in terms of mortality and some simple morbidity vari- ables) to generate hypotheses regarding this ethical question. Information not available from these existing files might be obtained through medical record abstraction or other activities of the Medicare PROB. In the future, health and functional status information should be matched with Medicare Part A and Part B data. ~ These factors for predicting selection of nonsurgical treatment options are based on preliminary results of a study being conducted by investigator at Lee RAND Corporation. Approximately 2,800 patients with hip fractures were selected from the Medicare files and their medical records were abstracted to obtain information to clarify treatment selection factors. Approximately 5 percent of the patients with hip fracture were managed nonsurgically.
36 Other Treatment Issues EFFECTIVENESS INlTLt4TWE The committee recommends that HCFA address four other clinical treatment effectiveness issues: impact of comorbidity, patient predictors of excessively long lengths of stay or high costs of care, better clinical data sets, and payment incentives affecting clinical decision making. First is the question of comorbidity and its observed effects on treatment choices and outcomes. As noted elsewhere in this report, work is needed on patient classification and risk stratification problems. One instrument cited by the committee that might be used in effectiveness research to monitor the impact of comorbidity is the Gonnella disease staging system. This severity- of-illness classification system, which is based chiefly on the presence of comorbid conditions, uses data that already appear in the HCFA Medicare Part A files; thus, it might be useful now for risk stratification purposes. Stratification by comorbidities may enable researchers to pose more precise questions regarding timing and intensity of rehabilitation. Second, work should be done to identify the clinical and demographic variables that predict cost and utilization outliers, that is, those patients who greatly exceed the average costs or institutional lengths of stay for treatment of their hip fractures. Such research may provide important clues to treatment effectiveness. Third, the committee recommends that HCFA continue its efforts to establish a valid and reliable clinical data sets for use in effectiveness research. Additional mechanisms should be developed by which all data sets (i.e., the administrative claims files and the files containing inpatient clinical information) might be linked. Survey data, containing functional and health status information, when linked to inpatient and outpatient data, may ultimately provide an additional powerful input into treatment effectiveness and outcomes research. The barriers that exist to linking person-specific data bases from different sources are very great, however, and issues of confidentiality and privacy must be addressed. In the near future, therefore, the likely "linkage" will be on a population basis. Finally, many committee members expressed concern that the payment system may drive decision making for treatment choice. Monitoring utiliza- tion trends can highlight odd or perverse patterns of use and direct policy attention in such a way that treatment decisions will in future be made on the basis of desired outcomes and not on the level of reimbursement. Cone need for a minimal data set shared by all effectiveness research is implicit throughout these recommendations. The appendix contains additional information on the Uniform Clinical Data Set.
HIP FRACTURE i] 37 Rehabilitation The committee recommends that three rehabilitation questions receive priority consideration in effectiveness research because of their importance to health status and quality of life: 1. What are the appropriate rehabilitation programs for different kinds of hip fractures? 2. What rehabilitation programs are most appropriate for different kinds of patients? 3. Which sites for delivery of rehabilitation services are most effective n outcomes and cost? Professional uncertainty exists in matching rehabilitation programs to type of fracture. There is no firm consensus on extensiveness, for example, intensity of physical or occupational therapy or emotional counseling as a function of type of fracture. Rehabilitation after hip fracture is highly sensitive to cognitive, phys- ical, and functional status before the fracture. Some elderly people who fracture a hip are healthy and have no ADL or IADL impairments; others have numerous problems that predispose them to falls and fractures. The increased prevalence of chronic conditions in those most at risk for hip fracture make the issues of data reliability, validity, and comprehensiveness particularly important. Certain coexisting ailments (e.g., dementia, depres- sion, and chronic infections) may also inhibit the rehabilitation process. Assessment must not only determine the presence and severity of these conditions but also monitor their impact on the outcome of hip fracture treatment. The proper management of coexisting illnesses is potentially as important in rehabilitation as the type of fracture or the surgical procedure elected. Again, information on cognitive and physical function is generally not available from the administrative data set and often not from the medical record. The committee noted that a long-range goal for HCFA should be to incorporate this information into the medical record and claims data. In the short run, primary data collection will be required. Rehabilitation outcomes may also depend on the availability and uti- lization of appropriate services following acute care. The density of avail- able rehabilitation beds per population, the availability of appropriate personnel in home health agencies, and access to outpatient physical and occupational therapy are all important. The committee recommends that HCFA concentrate its effectiveness research for rehabilitation on four areas: (1) minimal data set, (2) timing and intensity, (3) longitudinal follow-up across settings of care, and (4) hypothesis generation.
38 Establishment of a Minimal Data Set EFFECTIVENESS INITLfT[VE Section 9305(h) of the Omnibus Budget Reconciliation Act (OBRA) of 1986 mandated the development (but not the implementation) of a uniform needs assessment instrument by the Secretary of DHHS. Its content will include measures of functional capacities, nursing care requirements, and the social and family supports that are available. The instrument would be used to evaluate the needs of patients for posthospital extended-care ser- vices, home health services, and long-term-care sentences of a health-related or supportive nature, and the data would be used by hospital discharge planners, home health care agencies, other health care providers, and Medicare fiscal intermediaries and carriers. It might also be used to deter- mine whether payment for long term care should be approved. A different requirement item from OBRA 1987 is for a uniform minimum data set for nursing homes.9 The committee strongly supports these developmental efforts and further field testing. Timing and Intensity Issues in Rehabilitation Medicare currently covers 100 days of rehabilitation services and nurs- ing home care when a potential for improvement can be demonstrated. Electiveness research could help define appropriate services and time frames for benefits that are effective in terms of both health status and cost. Sites appropriate for analysis include acute care hospitals, reha- bilitation hospitals, extended-care facilities, home settings when the care is rendered through home health agencies, and other settings delivering relevant outpatient care. Medicaid files might also be used to capture intermediate care facility utilization. Longitudinal Follow-up A major deficit in the knowledge of hip fracture rehabilitation is outcomes over time. Detailed health status information before fracture is not usually available. The current Medicare benefit structure permits tracking of patients across only a limited number of settings, for example, in acute and rehabilitation hospitals but not skilled nursing facilities after 100 days. 9An advisory panel appointed by the Secretary of DHHS has developed a draft of the uniform needs assessment instrument that was reviewed by interested organizations, associations, and provided. As of November 1989, a final instrument had been prepared but not yet transmitted to the U.S. Congress by the Secretary of DHHS. Both this effort and the uniform minimum data set for nursing homes are being coordinated by the Office of Survey and Certification, Health Standards and Quality Bureau, HCF~
HIP FRACTURE 39 Nevertheless, the results of rehabilitative care should be examined in different settings to determine whether Medicare program operations and coverage benefits limit those outcomes. Such outcome information obtained over time is critical if appropriate and effective alternatives for rehabilitation are to be identified. Put another way, this information is needed to decide what services should be made available to Medicare beneficiaries, and to avoid policies that may make outcomes worse. Governmental behavior, as well as provider and patient behavior, need to be critically evaluated in this regard. Hypothesis Generation The possibility that the HCFA databases can be used to generate hypotheses should not be overlooked. Often, a hip fracture signals a rapid decline in health status and functional capabilities. Through the HCFA data, variations within geographic areas and among different practitioners might be examined to generate hypotheses regarding type, site, and intensity of rehabilitation services; for instance, about the relative merits of providing rehabilitation in the acute setting, a rehabilitation hospital, or at home. These hypotheses could then be examined by linking HCFA data files to other data bases or by funding experimental and quasi-experimental studies. Hip Fracture in the Context of Aging Few data are directed specifically to understanding what a hip fracture actually represents biologically, that is, in terms of aging per se. If hip fracture represents the accumulation of risk factors and therefore is a clinical manifestation of decline, then it may be inappropriate (if the rate of decline is moderate or rapid) to expect many individuals to return to prefecture levels of functioning. If, on the other hand, the fracture is an acute injury that itself can precipitate a major decline in an otherwise intact individual, then treatment strategies must be developed to minimize side effects or inadequate rehabilitation. Some experts suspect that, for some people, hip fracture is a precipitating event and that, for others, it signals a decline that is well in progress. Research is needed to determine how to classify individual patients and how to apportion resources so that the maximum recovery level for each person suffering a hip fracture can be achieved at a reasonable cost.
