A comparison of hospital-based and private outpatient physical therapy practices.

6

A paucity of data exists about the delivery of contemporary adult outpatient physical therapy services in the United States. Although many debates about this issue have taken place within the physical threrapy Profession, assumptions about practice patterns have generally not been studied. This article reports some of the firstyear findings from a 3-year survey research effort begun in September 1988 by Mathematica Policy Research Inc for the American Physical Therapy Association. Analyses are based on survey data from national probability samples of physical therapy facilities and discharged patients from 155 hospital-based and 166 private outpatient practices. The primary goal was to compare outpatient phsical therapy practice patterns in the hospital-based and private practice settings. Findings reveal considerable homogeneity in private and hospital-based outpatient physical therapy practices. Hospital-based and private practices serve predominantly young, white adults who have private health insurance and a prevalence of orthopedic impairments. In both settings, direct access to outpatient physical therapy services was the exception and not the rule. Even in states in which direct access without a physician's referral was permitted by law, the majority of outpatient physical threapy was provided with a physician's referral Most respondents in direct-access states indicated that reimbursement requirements were the major reason for needing a physician's referral to provide outpatient physical threapy. [Jette AM, Davis KO. A comparison of hospital-based and private outpatient physical threapy practices. Phys Ther. 199-1;71..366-381.1 Key Words: Ambulatory care,- Delivery of health care,, outpatients; Physical An unprecedented escalation of health care costs in the decades of the 1970s and 1980s has led to an increased focus on health care cost containment as a priority policy issue at both the federal and state levels of government as well as by the private sector.1-3 Although most attention has been directed toward containing hospital and physician costs, the profession of physical therapy has not been immune from external pressure regarding cost containment.4

In 1988, for example, the Health Care Financing Administration HCFA) of the US Department of Health and Human Services introduced a pre-reimbursement review of claims for outpatient physical therapy services provided to Medicare beneficiaries. Among other elements, these HCFA edits" (also known as "screens") are based on primary diagnosis, number of visits, and duration of care for specific episodes of illness. One of the major goals of these Medicare edits is the containment of physical therapy costs. in the development of these edits, the HCFA has attempted to establish a threshold for each specific diagnostic group that would theoretically screen and identify for secondary medical review 50% of all submitted claims. These diagnostic groups were constructed by the HCFA using similar diagnoses from the International Classification of Diseases Clinical Modification, 9th rev ed. Policy changes affecting physical therapy, such as the introduction of Medicare edits, have occurred despite the paucity of information about the delivery of outpatient physical therapy as currently practiced in the United States. The lack of available information on outpatient physical therapy has hampered the physical therapy profession's ability to contribute effectively to policy debates in the current politically charged health regulatory environment in the United States. Lack of information has also hampered the profession's ability to evaluate the impact of regulatory restrictions imposed on the profession from outside (eg, Medicare edits) as well as policy changes sought by the profession (eg, direct access) that may affect the quality of care provided to patients.5 Existing data sources on physical therapy services have been developed for other purposes and do not meet the profession's needs. The National Ambulatory Medical Care Survey (NAMCS), for example, provides some national data on provision of physical therapy services.6 It collects information, however, only on physical therapy provided by or ordered during a specific medical outpatient visit. The HCFA does possess data reported by its fiscal intermediaries on numbers of visits and numbers of patients receiving physical therapy (personal discussions with HCFA representatives, January 1989). These HCFA data, however, pertain only to Medicare beneficiaries, and no documentation is available on who provided the physical therapy services, how these data were collected by fiscal intermediaries, or how the sampling was done. This paucity of information severely limits the usefulness of these data to researchers. The American Physical Therapy Association (APTA) periodically collects systematic information about its members.7,8 This database provides information on individual therapists, but not on practices, and is restricted to members of the professional association. The growing demand for outpatient physical therapy and the shifting focus of the profession toward direct access to physical therapy services has been accompanied by a growing number of professional issues related to practice patterns, reimbursement policy, and referral relationships. These professional issues have generated debates within and outside of the physical therapy community and have increased the awareness that the profession needs adequate evidence to substantiate assumptions about outpatient physical therapy practice patterns.

In September 1988, the APTA undertook a 3-year research effort with Mathematical Policy Research Inc to systematically collect data at the facility and the patient level to describe outpatient physical therapy practice across a variety of settings. The primary research goals were (1) to describe the facilities and settings at which outpatient physical therapy is rendered with respect to allocation of personnel, staff productivity, employment relationships, utilization, fees, methods of charging for services, reimbursement sources, and policy on direct access; and (2) to describe a sample of discharged patients from the sampled practices with respect to their episodes of care, physical therapy services provided, goals and outcomes of care, and referral and reimbursement characteristics.

The goals of the first project year were to refine procedures and the overall study design and to provide an estimation of outpatient physical therapy practice patterns across practice settings. Plans for the second and third project years begun in late 1989) included the collection of in-depth patient-level data from a smaller range of practice settings, including detailed data on facility staff utilization, patient diagnoses, and reimbursement. The purpose of this article is to highlight the main results of the analysis of the facility and patient data collected from hospital-based and private outpatient practices in the project's first year. The primary goal was to compare outpatient physical therapy practice patterns in the hospital-based and private practice settings. Method

In the first year of the project, a multistage, stratified sample design was used to select a sample of hospital-based and private practices that provided outpatient physical therapy and to select discharged clients treated by those practices. By necessity, the practice settings and patient samples are linked, as it was not feasible to sample clients directly, without first sampling the facility that provided the service. The sampling plan consisted of three stages. First, preliminary samples of institutions or providers were selected from the American Hospital Association's 1987 list of general acute care hospitals and classified telephone directory listings of private practices. Second, one physical therapy practice within each selected institution was chosen, and its director or a designated representative was asked to complete a mail survey about the facility/ practice. Third, patient discharge records were selected from each practice. Within each setting, sample sizes were based on preliminary estimates of the number of facilities and the number of clients served and on analytic considerations. Details of the three-stage sampling method are presented in the Appendix.

A telephone screening was conducted to verify facility eligibility, to stratify sampled organizations, and to select a facility for the in-depth mail survey. A survey questionnaire was sent to each eligible facility to obtain information about the facility and up to 10 discharged clients. Follow-up mail and telephone contact was implemented to maximize response rates.

The resulting sample used in the firstyear analyses consisted of 155 general adult hospital practices and 166 private practices, which yielded 1,337 discharged clients (a 65% facility response rate). The private practices comprised 110 practices from nondirect-access states, which yielded 768 discharged clients (a 50% facility response rate), and 56 practices from direct-access states, which yielded 405 discharged clients (a 46% facility response rate). For the purpose of this study, a direct-access state was defined as a state that currently allows, by law, physical therapy evaluation and treatment without a physician's referral. Data Analysis Two separate sets of analyses were conducted. First, hospitals were compared with all private practices. Second, within the private practice sample, private practices in non-direct-access states were compared with private practices in direct-access states.

Two-tailed Student's t tests were computed for all comparisons. A probability level of .05 was used to determine statistical significance. Caution should be used in interpreting statistical significance from multiple t tests, because a least 1 of every 20 tests undertaken will achieve statistical significance by chance alone. Reported standard errors incorporate the design effects from the facility- and discharged patient-level data of the study and allow the reader to see the precision of estimated values. Reported means are weighted, based on the sampling design effects Appendix). Results

Given practice nonresponse rates ranging from 35% to 54% in this survey, the representativeness of the facility sample data was assessed by comparing current diagnoses from the telephone survey, between patients from facilities that returned the data-collection forms and those that did not. This comparison was possible because 91% of the eligible facilities completed the telephone screening survey, which included questions on the diagnoses assigned to the last five outpatients served by each practice. Current diagnostic data were coded into the HCFA diagnostic groupings. The percentage of diagnoses assigned to each group that had 40 or more clients was compared between responding and nonresponding facilities. The diagnostic distributions across facilities were very similar; only 2 of 15 comparisons differed at the .05 level of statistical significance. One of these two diagnostic groups included a number of neurological conditions such as multiple sclerosis, meningitis, encephalitis, and longstanding hemiplegia. The second diagnostic group included derangement of joints and other joint disorders. Staffing Size

Each participating outpatient facility reported on the total number of physical therapists (Pts), physical therapist assistants PTAs), and other staff who provided care to patients in their practice. Data on staffing distributions reflected the total number of staff in a practice rather than only the full-time equivalents. The information on number of staff in hospital practices included inpatient as well as outpatient components of the practice, thus limiting cross-practice comparisons. The data presented in Table 1 contrast the mean number of staff in hospital practices compared with those in all private practices and the mean number of staff in private practices in direct-access states compared with those in non-direct-access states.

Physical therapists were the primary care providers in both hospital-based and private practices (Tab. 1). On average, hospital practices reported almost twice the number of Pts X= 6.1) and PTAs -X=3.0) than their private practice counterparts. Approximately 36% of the private practices reported having PTAs, on their staff as compared with 55% of hospital-based practices. No significant staffing differences were observed between private practices with direct access and private practices without direct access.

Facilities reported on the total number of visits per PT and per PTA per week within their practice. The number of visits per PT per week was significally higher in private practices as compared with that in hospital-based practices. Physical therapists in hospital-based practices averaged 45.8 visits per week compared with a mean of 63.3 visits per PT per week in private practices (Fig. 1). On average, private practices reported longer treatment visit times X=54 minutes) compared with hospital-based practices X-=44 minutes). Hospital-based practices also reported a lower mean number of visits per PTA 7X=33.2) as compared with private practices X-=43.7). Private practices in states without direct access reported a significantly higher mean number of outpatient visits per PT per week R=66.2) compared with the mean number of outpatient visits for private practices in direct-access states X=53.8). Outpatient Profiles

Few dramatic cross-practice differences were observed in the demographic and clinical characteristics of the discharge sample of patients included in the study. Outpatients were evenly divided between men and women; over three quarters of the discharged patients were white adults between the ages of 18 and 64 years. Patients 65 years of age and older constituted 15% of the private-practice clients compared with 24% of the hospital-based clientele. Less than 10% of all the discharge samples represented patients 75 years of age or older. Hospital-based and private outpatient physical therapy practices treated primarily patients with orthopedic impairments. As the data in Table 2 reveal, back injuries were the most prevalent impairment in the patients sampled from hospital and private practices. Patterns of Outpatient Care

Outpatients who received short-term episodes of care (up to 60 days) were prevalent in hospital-based and private practice settings. Short-term care accounted for 80.2% of the outpatients in hospital-based practices compared with 66.8% of outpatients in private practices (Fig. 2). Episodes of care of long duration (7 months or more) accounted for 10% or less of the outpatients seen in each setting. The outpatient physical therapy practices in this study reported a variety of methods of charging for their services and accepted a diverse mix of payment sources. The data in Figure 3 illustrate that hospital-based and private practices both relied predominantly on charging payers by the procedure" (70.3% and 60.2%, respectively). Private practices were significantly more likely than hospitals, however, to charge patients by the visit." There were significant differences in method of charging for services between private practices in direct-access states and those in nondirect-access states. Almost half of the private practices in direct-access states (41.0%) used the "relative value unit" method of billing for most of their services compared with 6.8% of the private practices in non-direct-access states. A relative value unit is a nonmonetary unit that is based on relative resource costs required to perform a specified service or procedure. Approximately half of all the practices in this study reported private insurance companies such as Blue Cross/ Blue Shield as the primary source of payment for their services (Fig. 4). Medicaid was a negligible source of payment to all types of practices. Hospital-based practices were more likely than private practices to be reimbursed by Medicare as well as by Medicaid. The typical charges for evaluation and treatment visits were slightly higher in hospitals compared with private practice settings; however, these differences were not statistically significant at the .05 level. The average charge per evaluation visit was $70 for hospital practices and $66 for private practices. The average charge for treatment visits was $58 for hospital practices and $54 for private practices. Referral Policy and Practice Survey respondents were asked to report their facility's policy on outpatient practice without referral and to estimate the percentage of their outpatients who generally received services without a physician's referral. Hospital-based physical therapy practices were significantly less likely than private practices to report a policy of evaluating outpatients without a physician's referral (Fig. 5). Only 9.7% of the hospital-based practices reported a policy of treatment without a physician's referral compared with 18.5% of the private practices. Practices in direct-access states were more likely (65.8%) than practices without direct-access provisions (11.2%) to report their policies allowed for evaluation and treatment of outpatients without a physician's referral. In direct-access states, however, 33.3% of the private practices reported they did not allow evaluation without a physician's referral, 45.3% reported they did not allow treatment without a physician's referral, and 34.2% reported that they allowed neither evaluation nor treatment without referral. The patient discharge data in Figure 6 reveal that the actual behavior in outpatient physical therapy practices did not always conform to the facility's reported policy. When we examined the referral status of the discharge sample, 80% to 90% of the outpatients discharged from all practices were patients who had a physician's referral for physical therapy services. In actual practice, hospital-based outpatient practices were significantly more likely than private practices to actually evaluate patients without referral (18.0% of the hospital discharge sample compared with 9.1% of the private practice discharge sample).

Private practices in direct-access states reported that reimbursement requirements were the most common reason for having a physician's referral. Hospital practices also reported that institutional policy was their most common reason for having a physician's referral. Orthopedic surgeons and primary care specialists were the most common sources of patient referrals to physical therapy in both settings (Tab. 3). These analyses provide a look at some widely held assumptions and beliefs about contemporary practice of adult outpatient physical therapy in two major settings. One of the most interesting first-year findings is the difference between the opportunity for outpatient physical therapy practice without referral and actual clinical practice patterns observed for discharged patients. These data confirm what many have expected: Direct access to contemporary outpatient physical therapy services, regardless of practice setting and state law, is still the exception and not the rule. In 1988, the vast majority of outpatient physical therapy was provided with a physician's referral, even in states in which direct access was allowed by law and in practices in which the stated policy was to allow treatment without referral. Many explanations may account for this difference between stated policy and actual clinical practice. For many states, direct access is a relatively new phenomenon, and one would not expect a dramatic or an immediate change in practice patterns to occur. In 1988, on average, the 22 direct-access states studied had legalized direct-access practice for less than 3 years and half of the states had it in effect for 1 year or less. In addition, 82% of the therapists in direct-access states indicated that reimbursement requirements were the major reason for receiving a physician's referral for physical therapy. The second most frequent reason, given by 65% of the respondents, was that the therapist preferred to practice with a physician's referral. The reason for these physical therapists' preference for a referral is unknown. Further research is needed to determine why a large percentage of physical therapists indicate a preference for a physician referral when the opportunity for direct access is available and allowed by state law and practice policy.

The survey findings reveal some very interesting similarities and differences in the practice of outpatient physical therapy in hospital-based versus private practice settings. On average, hospital-based physical therapy practices are larger than private practices. Hospital practices employ or contract with twice the number of staff employed or contracted with by private practices. This difference, however, is probably due in large part to the staffing requirements and schedules associated with most hospital-based practices. It may also reflect the size of the institutions in which hospital practices are located. The staffing levels that were reported by the hospital-based practices in this investigation are similar to estimates from prior research.

In a 1986 APTA survey of current members, for example, the respondents who were employed in hospital-based practices reported that their facilities had an average of 305 beds, with an average PT-to-bed ratio of 1:44-an average of 7 Pts per hospital-based practice.9 The data suggest that utilization in hospital-based practices appears, on average, to be lower than the utilization achieved in private practice settings. The term "utilization" in this context refers to the number of outpatient visits per provider per week and does not take into account the quality of treatment or differential patterns of patient diagnoses between the two practice categories. Despite the smaller number of therapists and support staff, private practices reported slightly more visits per week compared with hospital-based practices. In addition, hospital-based practices achieved lower levels of utilization per therapist, averaging fewer visits per therapist per week, compared with private practices. Utilization differences across settings cannot be explained by differences in visit length, because, on average, private practices reported longer treatment visit times as compared with hospitalbased practices. Differences in utilization also are not attributable to the hours of staffing within each category of practice. Therapists in hospital practices, on average, reported working slightly more hours per week 41.2) than therapists in private practice settings (38.3). Lower utilization in hospital practices, compared with private practices, is most likely due to hospital-based therapists having to treat both inpatients and outpatients. The hospital-based therapists also serve a patient population that is somewhat older and more heterogeneous, with a higher prevalence of neurological conditions, than the patient population served by therapists in private practice. Hospital-based therapists, therefore, are possibly in need of more time to provide and coordinate interdisciplinary care. The difference could also be attributed to the non-patient-care factors and activities inherent in a hospital-based practice, such as unforeseen patient schedule conflicts resulting in higher cancellation rates; internal staff meetings, teaching, conferences, and scheduling requirements, which are not usually as time-consuming in private practices; and more burdensome documentation requirements related to accreditation.

Because the hospital practices surveyed included inpatient as well as outpatient components, the crosspractices comparison of utilization and interpretation of findings arc limitations of the study. Based on the utilization findings, future research is needed to investigate the various reasons for the different levels of utilization, their relationship to productivity, and the impact of these differences on the quality of patient care. Hospital-based physical therapy practices are similar to private practices in the types of patients served. Outpatient practices in both settings appear to serve predominantly young, white adults who have private health insurance. Although hospital-based practice did serve a greater percentage of clients over 65 years of age (24%) compared with the private practice clientele 15%), less than 10% of the discharge samples represented patient 75 years of age or older, the age cohort with the highest levels of impairment and disability.10 Reimbursement sources across settings reveal a greater diversity of patients in the hospital set ting than in the private practice setting Hospital practices, as compared with private practices, serve a significantly higher percentage of Medicare and Medicaid beneficiaries. Nonetheless, Medicaid beneficiaries represent less than 4% of the discharged patients from hospital-based practices compared with under 1% of private practices. Research is needed to identify whether barriers exist that limit access to outpatient physical therapy services for aged and disadvantaged patients. Orthopedic impairments represent the primary reasons for receiving physical therapy among the discharge sample from both settings. Stroke patients were more frequently reported in the hospital-based practices than in the private practices. Referral patterns mirror the diagnostic distribution in both settings, with the highest proportion of outpatients having been referred from an orthopedist. Practices also received almost a third of their referrals from primary care physicians. Few referrals came from a physiatrist, reflecting the increasing divergence of outpatient physical therapy from historical referral relationships."

The average number of visits per episode for hospital-based and private practices appeared to fall within most of the stipulated HCFA edits currently in use for selecting bills for level 11 medical review. The limited sample of patients 65 years of age or older precluded an examination of this issue specifically for patients in the Medicare program. Conclusion

This article describes some of the key findings from the first year of the APTA's survey of outpatient physical therapy. Although these findings are the beginning of a 3-year effort, they show that considerable homogeneity exists in the manner in which outpatient physical therapy is provided in both hospital based and private practices. They also show that there has been little or no change in the manner by which clients gain access to physical therapy services in states with direct access compared with states in which direct access is not legal.

Future research undertaken as a part of the second year of this project will compare outpatient physical therapy as practiced in rehabilitation and general acute care hospitals. Comparisons will also be made regarding physical therapy in physicians' offices, physician-owned physical therapy practices, and private physical therapy practices. The research will be analyzed by specific diagnostic categories to evaluate possible differences that might exist in treatment patterns among the various practices, including duration of episodes of care, number of visits, procedures, and costs. Finally, the study will seek to compare the actual episodes of care with the Medicare diagnostic edits used for claims review.

The use of epidemiologic health services research in future physical therapy practice is essential to increase the profession's understanding of the nature, the costs, and ultimately the effectiveness of care provided, to influence the formulation of public policy, and to differentiate professional reality from rhetoric. Acknowledgments

We acknowledge Richard Strouse and john Hall of Mathematical Policy Research Inc for designing and conducting the survey and analyzing the survey data and Laurie Hack for her role as project consultant; John Mirone of the APTA's Department of Practice for technical assistance in development of the manuscript; the APTA Committee on Physical Therapy Practice for ongoing advice provided to the project team; Dan Dyrek, Liz Gaynor, and Steve Haley for their thoughtful comments on an earlier version of the manuscript; and all respondents for their cooperation in completing the survey questionnaires. 1. Facility Sampling Frame There were three stages of sample selection. First, preliminary samples of institutions or providers were selected. Second, one facility within each institution was selected. Third, patient discharge records were selected. Within each setting, sample sizes were based on preliminary estimates of numbers of facilities, numbers of clients served, and analytic considerations of minimum numbers of facilities that would permit analysis of settings of particular interest. The sample selection method used required that weights be constructed to produce unbiased estimates and dictated that estimates of sampling error (confidence intervals, tolerances) take account of these departures from simple random sampling. Sampling errors are affected by weighting, by stratification, and, for the examination of discharge records, by clustering.

Facilities were sampled in two stages. In the first stage, preliminary sample lists were obtained or constructed for selected settings that provide physical therapy services, and organizations were sampled within each setting. In the second stage, these organizations were screened to verify that they were correctly classified by type of setting and that outpatient physical therapy services were provided, and one facility within the organization was selected to be surveyed. Sampled facilities were asked to specify, in the facility questionnaire, the settings in which they provide physical therapy services. Although all facilities were sampled in only one setting, 17 facilities classified themselves in settings other than those in which they were sampled. Misclassified facilities were reassigned to appropriate strata, but their initial probabilities of selection were retained. Some facilities, while providing services in the setting in which they were sampled, characterized their facilities as providing service in other settings. In these cases, the facility was not reassigned to a different stratum. A. Hospital-based Practice

A "hospital-based outpatient department" was defined as an outpatient facility located in the same building with a hospital inpatient physical therapy department or, if there was no facility in the same building, as a separate facility that was staffed by the inpatient physical therapy department. They were sampled f rom the American Hospital Association Abridged Guide to the Health Care Field Diskette (1987 Edition), which contains data from the 1986 AHA Annual Survey of Hospitals conducted by the Hospital Data Center of the American Hospital Association (AHA, 1987). The hospitals were stratified according to three categories, separating the pediatric and rehabilitation/orthopedic hospitals from the general/adult category. Preliminary samples for these strata were chosen by selecting hospitals in proportion to the number of inpatient admissions reported in 1986. If the number of admissions was missing, the hospital was excluded from the sample. Within the hospital strata, selection at the first stage was made using systematic selection after a random start, with probability proportional to size, using the number of yearly admissions as a measure of size. It was assumed that there was only one outpatient department providing physical therapy services. B. Private Practices

The preliminary sample for private physical therapy offices and clinics was drawn from Survey Sampling Inc's (SSI's) database of all organizations listed under the heading "Physical Therapy" in Yellow Pages directories across the country. For each organization, the SSI database documents duplicate listings, but excludes duplicate listings from different directories. Therefore, for each organization in the preliminary sample for this setting, it was known whether it was also listed under physicians and surgeons, rehabilitation services, and so on. This setting includes offices and clinics in which clients are seen, but excludes private practices that staff only facilities owned or managed by organizations whose clients could be sampled through other practice settings. For example, a private practice that staffs a hospital physical therapy department, but sees no patients elsewhere, would be excluded. One exception of this general rule was made in hopes of increasing the number of physician-owned physical therapy services (POPTS). Private practices that have no facilities of their own, but that do staff physician-owned off ices or clinics, were considered eligible for this sample. The preliminary sample for the private physical therapy offices and clinics was stratified to oversample the 22 states that treat patients following direct access rather than referral. 2. Sample Verification and Facility Mail Survey At the second stage of sample selection, sampled institutions were screened by telephone to verify eligibility and stratum assignment and to select the facility to be surveyed by mail. Because the screening interview was completed prior to the facility questionnaire (which was self-administered), some stratification assignments were incorrect. Seventeen cases were sampled for the wrong stratum and were reclassified. 3. Sample of Client Discharge Forms In principle, information about outpatient physical therapy at the client level can be collected following discharge from treatment, while treatment is in progress, when the client begins treatment, or when treatment is first prescribed. This study relied on retrospective collection of client data at the point of discharge or as close to it as possible.

Most of the client information gathered in this study concerned the entire episode of care. Information concerning the total length of the episode, the total number of visits, and the total charges can be given accurately only following discharge. Total reimbursement should be reasonably well-estimated at discharge, even though actual reimbursement References

1 Waldo DR, Levit JR, Lazenby H. National health expenditures, 1985. Health Care Financing Review. 1986;8(l):1-21.

2 Prospective Payment Assessment Commission., Medicare Prospective Payment and the American Health Care system. Report to the US Congress; Washington, DC; February 1987.

3 Herzlinger RE, Schwartz J. How companies tackle health care costs: part 1. Harvard Business Review. 1985;63(4):68-81.

4 Brown GD. Changing health care environmentsimplications for physical therapy research, education, and practice: a special communication. Phys Ther. 1986;66:1242-1245.

5 Singleton MC. Independent practice-on the horns of a dilemma: a special communication. Phys Ther. 1987;67:54-57.

6 Nelson C. Physiotherapy Office Visits.. National Ambulatory medical Care surveyUnited States, 1980-81. Washington, DC; National Center for Health Statistics, US Dept of Health and Human Services; july 11, 1986; NCHS Advance Data #120.

7 1981 Active Membership Profile Survey. Washington, DC: American Physical Therapy Association; 1981.

8 1987 Active Membership Profile Survey. Alexandria, Va: American Physical Therapy Association; 1987.

9 The Impact of the Prospective Payment System on the Delivery of physical threapy Services. Alexandria, Va: Department of Practice, American Physical Therapy Association; 1986.

10 Fulton J, Katz S, jack S, Hendershot G. Physical Functioning of the Aged: United States 1984. Washington, DC: National Center for Health Statistics, March 1989. US Dept of Health and Human Services publication no. 167 (series 10).

11 Lott D. Physical therapists and physicians: autonomy and control. Rehabilitation Management. 1989;2(l):44-49.