Initial health status of patients at outpatient physical therapy clinics.

[Mossberg KA, McFarland C Initial health status of patients at outpatient physical therapy clinics. Phys Ther. 1995; 75:1043-1054.]

Key Words: Hospital practice, Private practice, Outcomes, SF-36, Survey.

Patient-perceived health status is becoming more important due to demands of the public and third-party payers to improve the quality of health care and decrease the cost. The primary driving force is the demand for increased accountability in this era of health care reform.[1] As a result, assessment of patient-oriented health status has recently received increased attention from health care providers in all disciplines.[2]

Patient-oriented health status questionnaires provide patients and health care professionals with a common language for assessing the effectiveness of an intervention.[3] Furthermore, these assessment tools can be used to screen for functional limitations and improve treatment strategies.[4] Recently, Jette[5] has argued for the application of health-related quality-of-life measures and the importance of obtaining these types of data for physical therapy practice.

Application of quality-of-life or heigth status measures in daily clinical practice necessitates a variety of considerations. Besides validity. and reliability' one of the biggest concerns is the burden that completing the instrument places on the clinician or clinical staff as well as the patient. In addition, one must consider using either a generic instrument that is applicable to a variety of disease conditions or a condition-specific questionnaire that applies only to specific impairments or disabilities such as low back disorders or arthritis. The advantage of a generic instrument is that it has broad application and comparisons can be made across disabilities. An advantage of a condition-specific instrument is that it may be more sensitive to the patient's problem and to the degree of change that may take place over time.[6] Several generic instruments have been designed to assess health; status in reliable and valid ways. Examples are the sickness impact profile (SIP),[7] the quality of well being scale,[8] and a shortened version of the Medical Outcomes Study (MOS)[9] referred to as the SF-36.[10] Each of these instruments encompasses the domains of physical maintenance and self-maintenance, social function, and role function.

Others have reported on the health status of individuals that could be seen in physical therapy practice using the SF-36. Kantz and co-workers[11] assessed surgical outcome in patients with total knee replacements. Patients With total hip replacements and those with low back pain have also born assessed with the SF-36.[12] These studies were initiated through the physician's office, and the patients play or may not have received physical therapy. Recently, Jette and Downing[13] assessed; the health status of patients entering hospital-based cardiac rehabilitation programs. No work has been reported on individuals seeking general outpatient physical therapy in which a variety of diagnoses are treated. In addition, there has been an increased demand for outpatient services and greater competition among different settings in which these services can be provided. The purpose of this investigation, therefore, was to characterize the patient-oriented health status of patient groups seeking treatment at hospital-based and privately owned physical therapy clinics using the SF-36 health status questionnaire. To gain a greater understanding of condition severity, we also assessed whether differences in health status existed between our patients and previously reported normative data: and patients with other medical conditions.

Method

Subjects

Six outpatient clinics participated as a sample of convenience, and the clinic directors gave written consent to have their future patients participate. Three hospital-based clinics and three clinics privately owned by physical therapists participated. The director of each clinical site was asked to recruit a minimum of 20 patients into the study.

Patients excluded from the study were (1) an those referred for postsurgical rehabilitation, (2) those referred for more than one problem (eg, motor vehicle accident with cervical and lumbar strain), (3) those referred for rehabilitation of central nervous system disorders or diseases (eg, stroke, traumatic brain injury), and (4) those with reading or English-language difficulties. Patients referred postsurgically and those with comorbidities were excluded so that future physical therapy outcomes could be examined rather than outcomes due to other interventions or postsurgical healing.

Patients included in the study fit within one of five diagnostic categories, including upper-quarter musculoskeletal involvement of less than 3 months' duration, lower-quarter musculoskeletal involvement of less than 3 months' duration, upper-quarter musculoskeletal involvement of greater than 3 months' duration, arid lower-quarter musculoskeletal involvement of greater than 3 months' duration. For musculoskeletal disorders, we defined less than 3; months' duration as "acute" and greater than 3 months' duration as chronic." Patients with neck or low back pain as a chief complaint and with minimal or no radiating symptoms were also included in either upper- or lower-quarter groups. The fifth diagnostic category was neurologic, which included patients with peripheral nerve injury or radiculopathies on the cervical or lumbar region of less than 6 months' duration.

Survey Instrument

The SF-36 was derived by Ware and colleagues from the longer MOS health status survey[9,10] and consists of 36 items, with 35 items related to one of eight health concepts. The health concepts measured were limitations in physical functioning such as walking and climbing, stairs (10 items); limitations; in role functioning such as duties in the home or at work due to physical problems (role physical) (4 items); bodily pain (2 items); vitality and energy level (4 items); social functioning and interactions with others (2 items); limitations in role functioning due to emotional problems (role emotional) (3 items); mental health, depression, and mood states (5 items); and general health (5 items). One item assesses health transition over the past week, but because it does not constitute any part of the eight health concepts, it was not included in our data analysis.

The validity and reliability of the SF-36 have been reported previously.[14] We used the optically scanned version of the SF-36 developed and scored by a commercial scoring service.(*) Responses to each item were recorded, weighted, and converted to a scale score of 0 to 100.[15] Low scores indicated poor health, and higher scores indicated better health.

Two versions of the SF-36 are available. The standard version asks patients to relate their activities and behaviors to those carried out over the past 4 weeks, whereas the acute version asks patients to relate their activities and behaviors to those carried out over the mst week. We used the acute version of the questionnaire because of future studies aimed at posttreatment follow-up. Because physical therapy is encouraged to be as brief as possible, in many cases 2 to 3 weeks or less, we chose an instrument that had the greatest potential to reflect changes over a short period of time.

Procedure

Within each clinical site, a clinical research coordinator (CRC) was designated. The CRC was any responsible individual employed in the clinic, including therapists, aides, and administrative staff. The CRC was. not the clinician involved in the treatment of the patients in this study. Upon initial referral, patients were identified by the CRC and asked to participate based on the inclusion and exclusion criteria. The CRC then presented a cover letter explaining the study to each patient.

After giving verbal consent, the patient was given the health status questionnaire to complete. Completion of the questionnaire constituted written consent. The CRC was available to answer any questions the patients may have had while they responded to the questionnaire. The CRC also developed a patient profile that included demographic information, diagnosis, and a basic outline of treatment.

An attempt was made to prevent the therapist treating a patient from gaining knowledge of the patient's participation in the study. In many cases, the physical layout and patient flow of the smaller clinics did allow the treating therapist to gain knowledge of the patient's participation. In all cases, the treating therapist had no, knowledge of the patient's responses to the items in the questionnaire. Because patients were instructed to place the completed questionnaire in a self-addressed, stamped envelope and seal it, the CRC did not usually have access to the item responses either. In many cases, however, the patients simply handed the completed questionnaire back to the CRC without regard to confidentiality of responses.

Data Analysis

Chi-square analyses were performed to examine distributions across age, gender, clinical setting, and diagnoses. Due to the nonnormal distribution of health status scores, comparisons between groups were. performed with nonparametric statistical procedures.[16] Pair-wise comparisons for each of the eight health concepts were performed using the Mann-Whitney U test. We compared patients referred to hospital-based clinics and those referred to private practice settings, male and female patients, patients with upper-quarter musculoskeletal involvement and those with lower-quarter musculoskeletal involvement, and patients with acute musculoskeletal disorders and those with chronic musculoskeletal disorders. We also compared patients with acute musculoskeletal disorders to those with chronic musculoskeletal disorders to those with neurologic disorders using the Kruskal-Wallis one-way analysis of variance. Comparisons were made with normative data on the US population by using single-sample t tests and Bonferroni's adjustment to preserve the alpha level. The parametric single-sample t test was necessary because of the lack of an analogous nonparametric procedure for comparing our sample data to the normative means. Values for the US population were obtained from previously published literature.[15] All values reported are means ([+ or - ]SE). All statistical analyses were performed using a PC-based statistical program.([dagger])

Results

One hundred thirty-four patients volunteered to participate. Twenty-five patients had incomplete health status questionnaires or patient profile forms; therefore, data for a total of 109 patients were included in the final data analysis. Patients ranged in age from 17 to 82 years (43.2[+ or -]1.4 [[chi bar] [+ or -]SE]). Our sample comprised 31 male patients and 78 female patients, with 44 patients at hospital-based clinics and 65 patients at private practice clinics.

First, we compared subjects' ages and diagnoses across clinical settings. Table 1 shows the age distribution for the sample overall and compares hospital-based and private practice clinical settings. There was no difference in age distribution across clinical Settings ([chi square]=7.106, P=.311). We also found no difference in the distribution of male and female subjects among the different age groups ([chi square]= 3.007, P=.808) or across clinical sites ([chi bar]=0.429, P= .512).

[TABULAR DATA 1 OMITTED]

Figure 1 shows the diagnostic categories overall and the distribution relative to clinical setting. There was no difference in the distribution of the diagnostic categories across clinical settings ([chi bar]= 2.884, P=.577). The age distribution across the different diagnostic categories was not different ([chi square]= 22.313, P=.561), nor was gender by diagnostic category ([chi square]= 3.061, P=.548). In the aggregate, 40 patients had upper-quarter involvement and 37 patients had lower-quarter involvement. Of these patients, 40 patients had acute (<3 months duration) upper- or lower-quarter involvement and 37 patients had chronic (>3 months: duration) upper- or lower-quarter involvement.

A comparison of patient-perceived health status according to gender is presented in Table 2, Age-matched normative data from the general population of the United States are also included in Table 2. The normative data for men (n=384),and women (n=457) are the means of the two age groups most closely represented by our patient sample (35-44 years and 45-54 years). The trend in the normative data for men to consequently rate their own health better than women rate their own health was not demonstrated by our patient sample. The Mann-Whitney U test revealed that all health concepts were similar in our patient group, except male subjects score themselves higher than did female subjects for the concept of mental health (U=1,502, P<.05). Male subjects also perceived their vitality to be higher than did female subjects, but this difference was not statistically significant (U=1,474, P=.074).

[TABULAR DATA 2 OMITTED]

Figure 2 illustrates the similarity of health status between hospital-based and private practice clinical sites, as well as the health status scores of age-matched asymptomatic individuals from. The general population of the United States. When the data were analyzed with a parametric single-sample t test, all health concepts were lower (P<.01), except general health in the private practice group of patients. MOK but not all, health concepts, tended to be slightly higher in the patients coming to private; practice clinics as compared with the patients coming to hospital-based outpatient clinics. We found no difference, however, between patients when comparing clinical settings.

We also examined the differences in health status of patients according to diagnostic category. For analysis, we combined the patients with. acute upper-quarter and lower-quarter involvement into an acute musculoskeletal category. We shoarly combined the parents with chronic upper-quarter and lower-quarter conditions into a chronic musculoskeletal category. We also included, the patients with neurologic involvement. In the analysis. These data are shown in Figure 3. We found no difference in health concept scores for 4ny of the three diagnostic groups. There was a tendency, for the role emotional concept to@be highest in the acute musculoskeletal group and lowest in the neurologic group, but the difference was not significant (Kruskal-Wallis statistic=3.927, P=.14).

We also combined the patients with musculoskeletal diagnoses into upper-quarter and lower-quarter groups, disregarding the acute or chronic nature of their disorders. These results are summarized in Figure 4. We found lower physical functioning in those individuals with lower-quarter diagnoses (Mann-Whitney U statistic=470.5, P=.006).

Discussion and Conclusions

We believe this is the first study to examine health status with the SF-36 in a group of patients referred to general outpatient physical therapy. We found health status was similar when comparing hospital-based and privately owned physical therapy clinics. Similarly, we found no differences in the health status of female or male patients. We found a difference in physical functioning when comparing patients with upper-quarter involvement with those with lower-quarter involvement. The aggregate was found to be less healthy in all health concepts except general health when compared with normative data.

In order to focus on the physical therapy outcome, we felt it necessary to eliminate those patients for whom the outcome might be due to (1) other interventions aimed at treating the comorbid condition or (2) the almost certain healing in the postsurgical period. In addition, we did not include patients with potential communication difficulties (eg, central nervous system disorders). Future studies will be required to examine the use of proxies in these patient groups. All of these factors necessitated using a subgroup of patients for this initial study. The results, therefore, can only be applied to individuals who fall into diagnostic categories similar to the five categories described in this report.

Because of the variety of patients seen in physical therapy practice, we chose the generic SF-36 health status questionnaire. A major factor in the decision was the concern of each of the clinics regarding the addition of a task to an already busy clinic schedule. We found that if the CRC was an administrative staff person who was well trained, overburdened clinicians did not need to be involved. The SF-36 was self-administered, with little to no assistance required by clinical staff. The questionnaire was easily completed by most patients within 15 minutes.

Equally important in our decision to use the SF-36 was the fact that this health status questionnaire has been shown to meet the minimal criteria of being a valid and reliable instrument.14 Reliability estimates for internal consistency Cronbach's alpha) have been calculated previously on samples of apparently asymptomatic individuals, with coefficients ranging from .63 to 94.[17,18] Both studies reported that social functioning was least consistent and physical functioning was most consistent (the mean reliability coefficient for all eight health concepts, .83).[17,18]

Numerous studies have investigated the reliability of the SF-36 items in a variety of patient populations.[11,13,19-22] In patients with chronic conditions, including hypertension, congestive heart failure (CHF), myocardial infarction, and clinical depression, coefficients ranged from lows of .78 for general health[22] and .80 for social functioning[21] to highs of .92[21] and .93[22] for physical functioning. Patients with total knee replacements had coefficients ranging from .77 (social functioning) to .90 (role physical)," and patients undergoing hemodialysis had coefficients ranging from .62 (vitality) to .90 (physical functioning).[19] Overall, these reports have demonstrated that physical functioning most often provided the most consistent scores, whereas social functioning most often provided the least consistent scores.

Our study provides the opportunity to examine differences and similarities between patients seeking treatment at hospital-based clinics and those seeking treatment at privately owned clinics. We observed no difference in health status when comparing patients going to hospital-based versus private practice clinics. In addition, we found similarities in terms of ages and diagnoses of individuals referred for treatment in these facilities. These data indicate that the patients' perceived level of condition severity, diagnosis, or age was not a factor in whether the patients received treatment at larger hospital-based clinics or at smaller privately owned clinics.

Although diagnostic categories were different between our study and a previous study,[23] we found a similar distribution of the five diagnostic categories between private and hospital-based clinics. In addition, Jette and Davis[23] found back problems to be the most prevalent impairment. Similarly, we found individuals with cervical and low back disorders were by far the greatest percentage of patients seeking treatment (25.7% and 34.9%, respectively).

Unlike Jette and Davis,[23] who reported an equal distribution of male and female patients, we had a predominance of female patients (72%) in our study. Other researchers[24] have documented a larger proportion of female patients. Our sample, however, was relatively small, and therefore generalizations cannot be made about the distribution. Furthermore, we found no difference between male and female patients for the health concept scores. These data suggest that male and female patients in the five diagnostic categories described perceive the severity of their health deficits as similar. When compared with the normative data,[15] we did not observe the same consistent trend for male patients to rate their own health higher than female patients rate their own health.

Our results indicate that across the diagnostic categories, as we defined them, there was no difference in health concept scores when comparing acute and chronic conditions or musculoskeletal and neurological impairments. We did, however, find lower physical functioning scores in patients with lower-quarter involvement compared with those with upper-quarter involvement. This finding was not surprising given that nearly all the items in the physical functioning scale are in some way associated with the use of the lower extremities. A patient with a lower-quarter disorder would be limited in activities such as walking and running, bending and kneeling, and carrying groceries, whereas an individual with upper-quarter involvement may score low only on items such as carrying groceries, bathing, or dressing.

A major advantage to using a generic instrument such as the SF-36 is that it provides the clinician with a common measure of severity of impairment in patients with single or multiple problems. Indeed, Patrick and Deyo[6] have advocated that generic instruments can be utilized as screening tools. The ability to compare and contrast with normative data or other patient groups can provide a better understanding of a patient's overall health status and allow the clinician to be more sensitive to the overall needs of the patient. As Table 2 and Figure 2 indicate, we found that the patients in our sample scored lower in seven of the eight health concepts than a sample of apparently asymptomatic individuals. The areas with the most dramatic deficits were physical functioning, role limitations due to physical problems, and bodily pain. Surprisingly, the psychosocial domain was also lower, reinforcing the concept that the domains of physical functioning, social functioning, and mental health are interrelated. The clinician can use this information to understand the potential complexity of a seemingly straightforward single-area problem. in addition, each domain may serve as a guideline for progress during or after treatment.

The use of a generic health status measure also provides the clinician with a better understanding of his or her patients when placed in the context of other medical conditions and can give the clinician a sense of the level of severity. Figure 5 compares our sample with individuals referred for cardiac rehabilitation[13] and patients diagnosed with CHF,[25] type II diabetes mellitus,[25] or clinical depression.[26] Our patients stand out in terms of severity of bodily pain, whereas patients with clinical depression stand out in terms of diminished emotional and mental health.

Our patients had a single musculoskeletal or peripheral nerve involvement, with the most prevalent impairments located in the low back, neck, shoulder, or knee. We found physical functioning in our sample to be similar to that of patients with CHF, vitality to be equal to that of patients with CHF and clinical depression, and social functioning to be equivalent to that of patients with clinical depression. If the health concepts are examined more closely, the clinical meaningfulness of these comparisons becomes more evident. For example, low physical functioning scores mean that the patients were limited a great deal in their performance of physical activities such as walking, climbing stairs, bathing, and dressing. It is not surprising that the patients in our sample would have a deficit in this health concept. Nor should it be surprising that patients with CHF would have low physical functioning scores given that they typically have dyspnea on exertion, which can severely limit all physical activity. What was unexpected was that these two very different groups of patients would be so similar in physical functioning. Furthermore, our patients were on average younger than the sample of patients with CHF (mean age = 67 years) and based on age alone would be expected to have greater physical functioning.

When considering vitality, low scores mean that the individual feels tired and worn out all the time, whereas high scores mean that the individual feels full of pep and energy all of the time. it should not be surprising that patients with CHF or those with clinical depression score low on this concept. Again, somewhat unexpected was the fact that our patients scored equally low. Although patients may express subjective feelings with respect to vitality, it is not an aspect of healthy hat most clinicians inquire about directly. Similarly, direct inquiries about how physical or emotional problems interfere with normal social activities are not part of a standard physical therapy assessment. Our data suggest that the SF-36 or other similar instruments may provide additional information revealing clinically meaningful deficits in other aspects of health that therapists commonly do not assess. Having this information not only may provide a better understanding of the severity of a patient's condition but also may assist in the management of the patient.

As the database grows and the number of patients in diagnostic categories increases, comparisons with specific orthopedic and other diagnoses can be made. Lansky and co-workers[12] reported on the health status of patients receiving treatment in a spine center and found deficits similar to those of our patients. Unfortunately, comparisons are not possible between the patients in our study and the patients in Lansky and colleagues' study because of the lack of specific diagnostic categories, both in our study and in theirs. Lansky and colleagues did not consider anatomical location or duration of the problem, nor did they differentiate neurologic impairment from muscle strain or soft tissue involvement.

The choice of diagnostic groups in our study was the result of consensus agreement of both hospital-based and private practice clinicians and resulted in a compromise between very specific and very general diagnostic categories. Our patient subgroup did not seek treatment for any secondary problems. Future consideration must be given to those individuals who are treated for more than one area and have comorbid conditions. Predictably, multiple physical problems have been shown to have further adverse effects on health status.[15,27] The question of assessing physical therapy outcomes in postoperative patients will be more difficult and will require carefully designed studies.

The use of patient-oriented surveys by health care providers has increased dramatically in the last few years. Brown and Adams[28] have shown that patients are reliable when assessing issues related to their health, and health status self-assessments have been shown to be accurate.[29-31] Self-assessments have the potential to increase active patient involvement in health care decisions, and greater involvement has been shown to improve outcome.[2] Physical therapists have been trained to include patient goals in the management of problems, but external forces are now dictating that we provide more easily understood evidence of our ability to manage functional problems effectively.[32,33] Given the "common currency"[3] that patient-based health status questionnaires provide to patients, health care providers, and payers, these assessment tools deserve closer attention by physical therapists.

[Figures 1 to 5 ILLUSTRATION OMITTED]

Acknowledgments

We thank the clinical research coordinators and therapists in Texas who participated in the data collection. We are also appreciative of the assistance provided by Dr Frank Underwood, Ms Caroline Jansen, and Ms Annie Pati.

(*) RT-2000, Response Technologies Inc, East Greenwich, RI 02818. ([dagger]) SYSTAT for Windows, Version 5, SYSTAT Inc, 1800 Sherman Avi, Evanston, IL 60201-3793.

References

[1] Relman AS. Assessment and accountability: the third revolution in medical care. N Engl J Med. 1988;319:1220-1222. [2] Greenfield S, Kaplan S, Ware JE. Expanding patient involvement in care; effects on patient outcomes. Ann Intern Med. 1985;102:520-528. [3] Schipper H, Clinch J, Powell V. Definitions and conceptual issues. In: Spilker B, ed. Quality of Life Assessments in Clinical Trials. New York: Raven Press; 1990:11-24. [4] Deyo RA, Carter WB. Strategies for improving and expanding the application of health status measures in clinical settings. Med Care, 1992;30:MS176-MS186. [5] Jette AM. Using health-related quality of life measures in physical therapy outcomes research. Phys Ther. 1993;73:528-537. [6] Patrick DL, Deyo RA, Generic and disease-specific measures in assessing health status and quality of life. Med Care. 1989;27(suppl): S217-,S232. [7] Bergner M, Bobbitt RA, Carter WB, Gilson BS. The sickness impact profile: development and final revision of a health status measure. Med Care. 1981;19:787-805. [8] Kaplan RM, Anderson JP, Wu AW, et al. The quality of well-being scale. Med Care. 1989;27:S27-S43. [9] Tarlov AR, Ware JE, Greenfield S, et al. The Medical Outcomes Study: an application of methods for monitoring the results of medical care. JAMA. 1989; 262:92 5-930. [10] Ware JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). Med Care. 1992;30:473-483. [11] Kantz ME, Harris WJ, Levitsky K, et al. Methods for assessing condition-specific and generic functional status outcomes after total knee replacement. Med Care. 1992;30:MS240-MS252. [12] Lansky D, Butler JBV, Waller FT. Using health status measures in the hospital setting: from acute care to "outcomes management." Med Care. 1992;30(suppl): MS57-MS73. [13] Jette DU, Downing J. Health status of individuals entering a cardiac rehabilitation program as measured by the medical outcomes study 36-item short-form survey (SF-36). Pbys Ther. 1994;74:521-527. [14] McHorney CA, Ware JE, Rogers W, et al. The validity and relative precision of MOS short- and long-form health status scales and Dartmouth COOP charts. Med Care. 1992;30: MS253-MS265. [15] Ware JE. SF-36 Health Survey: Manual and Interpretation Guide. Boston, Mass: The Health Institute, New England Medical Center; 1993. [16] Portney LG, Watkins MP. Foundations Of Clinical Research: Applications to Practice. East Norwalk, Conn: Appleton and Lange; 1993. [17] Jenkinson C, Coulter A, Wright L. The SF-36 health status survey questionnaire: normative data from a large random sample of working age adults. BMJ. 1993;306:1436-1440. [18] McHorney CA, Kosinski M, Ware JE. Comparisons of the costs and quality of norms for the SF-36 health survey collected by mail versus telephone interview: results from a national survey. Med Care. 1994;32:551-567. [19] Kurtin PS, Davies AR, Meyer KB, et al. Patient-based health status measures in outpatient dialysis. Med Care. 1992;30:MS136-MS149. [20] Nerenz DR, Repasky DP, Whitehouse FW, Kahkonen DM. Ongoing assessment of health status in patients with diabetes mellitus. Med Care. 1992;30(suppl):MS112-MS124. [21] Garratt AM, Ruta DA, Abdalla MI, et al. The SF-36 health survey questionnaire: an outcome measure suitable for routine use within the NHS. BMJ. 1993;306:1440-1444. [22] McHorney CA, Ware JE, Lu JFR, Sherbourne CD. The MOS 36-item short-form health survey (SF-36), III: tests of data quality, scaling assumptions,. and reliability across diverse patient groups, Med Care. 1994;32:40-66. [23] Jette Aim, Davis KD. A comparison of hospital-based and private outpatient physical therapy practices. Phys Ther. 1991;71:366-381. [24] Liang MH, Fossel AH, Larson MG. Comparisons of five health status instruments for orthopedic evaluation. Med Care. 1990;28: 632-642. [25] Stewart AL, Greenfield S, Hays RD, et al. Functional status and well-being of patients with chronic conditions. JAMA. 1989;262:907-913. [26] Wells KB, Stewart AL, Hays RD, et al. The functioning and well-being of depressed patients. JAMA. 1989;262:914-919. [27] Verbrugge LM, Lepkowski JM, Imanaka Y. Comorbidity and its impact on disability. Milibank Q. 1989;67:450-484. [28] Brown JB, Adams ME. Patients as reliable reporters of medical care process. Med Care. 1992;30:400-411. [29] Kaufert JM, Green S, Dunt DR, et al. Assessing functional status among elderly patients: a comparison of questionnaire and service provider ratings. Med Care. 1979;17: 807-817. [30] Rubenstein LV, Calkins DR, Greenfield S, et al. Functional status assessment for elderly patients: report of the Society of General Medicine Task Force on Health Assessment. J Am Geriatr Soc. 1989;37:562-569. [31] Burns RB, Moskowitz MA, Ash A, et al. Self-report versus medical record functional status. Med Care. 1992;30(suppl): MS85-MS95. [32] Golden WE. Health status measurement. Med Care. 1992;30(suppl):MS187-MS195. [33] Thier SO. Forces motivating the use of health status assessment measures in clinical settings and related clinical research. Med Care. 1992;30(suppl):MS15-MS22.

KA Mossberg, PhD, PT, is Associate Professor, Department of Physical Therapy, School of Allied Health Sciences, University of Texas Medical Branch, Galveston, TX 77555-1028 (USA) (kmossbero/[e-mail]). Address all correspondence to Dr Mossberg.

C McFarland, PT, OCS, is Owner, McFarland Physical Therapy, Tyler, TX 75702.

The Institutional Review Board of the University of Texas Medical Branch reviewed and approved the methods of data collection in this study.

This work was funded in part by a grant from the Texas Physical Therapy Association and was presented in part at the annual conference of the Texas Physical Therapy Association, 1994.

This article was submitted December 22, 1994, and was accepted August 15, 1995.