The relationship between duration of physical therapy services in the acute care setting and change in functional status in patients with lower-extremity orthopedic problems.

Key Words: Functional status; Lower extremity, general; Orthopedics, general; Treatment outcomes.

Physical therapists have defended their role in the acute care setting by asserting the belief that the services they provide improve the physical function of the patients receiving them. Unfortunately, very little direct clinical evidence has been published demonstrating the impact of physical therapy on patient outcomes in the acute care setting.

In recent years, efforts at cost containment have produced shorter hospital stays, resulting in less time for acute care physical therapy. We can begin to infer something about the impact of physical therapy in the acute care setting by examining the results of shorter hospitalizations. Studies examining the effect of decreasing the length of acute care hospital stays on patient Function provide some evidence that when patients receive less physical therapy in the acute care setting, they are less functional at the time of discharge from the hospital. A study of patients with hip fractures in the United States showed that, after initiation of the prospective payment system (PPS), the average length of hospital stay decreased from 21.9 to 12.6 days and the average number of inpatient physical therapy sessions decreased from 7.6 to 6.3.[1] This same study showed that the proportion of patients who were ambulating and the distance ambulated at discharge decreased after initiation of PPS. Furthermore, the number of patients discharged to nursing homes increased from 38% to 60%.(1)

A study of patients with hip fractures in Finland and Sweden demonstrated a positive relationship between the length of stay in an acute care setting and the proportion of patients who were discharged to their own homes. For hospitalizations of 13, 17, and 19 days, 14%, 35%, and 49% of the subjects, respectively, returned home.[2] Unfortunately, both of these studies[1,2] failed to distinguish the effect of a decreased amount of inpatient physical therapy from that of a shorter hospitalization. These studies did not directly examine the effect of inpatient physical therapy.

Previous attempts to describe the outcomes of physical therapy using nonstandardized medical records have failed because too little information was available in the medical record to document improvement in function.[3] Thus, little physical therapy outcome research has been published.

The purpose of our study was to examine the relationship between the number of minutes of physical therapy provided to patients with lower-extremity orthopedic problems and their functional status at discharge. Unlike previous studies, the therapists in our study measured each patient's functional status at regular intervals using a standardized instrument. Furthermore, data were available on the amount of physical therapy received, the duration of the acute care hospitalization, demographics, and the nature and number of diagnoses. Therefore, we were able to examine the effectiveness of acute care physical therapy after adjusting for the impact of other factors. Because the impact of physical therapy services may differ depending on diagnosis and because orthopedic injuries and diseases are among the most common pathologies treated in acute care physical therapy, we have focused our research on this patient group.

Method

Design and Subjects

This was a retrospective cohort study of patients with orthopedic problems seen by physical therapists during an acute care hospitalization. Because this study involved a review of existing medical records, it was exempted from informed consent by the institutional review board. Data were collected from the quality assurance records, the general medical record, the physical therapy treatment log sheets, and the patient billing records on 173 patients admitted to an acute care hospital (and referred for physical therapy) between August 1, 1993, and December 31, 1993. Patients were included in the study if they (1) were seen by a physical therapist as an inpatient, (2) had a primary diagnosis of a lower-extremity orthopedic problem, and (3) were over 18 years of age. Only data from a first admission during the study period were used in the analysis.

Experimental Procedure

Functional assessment. The physical therapy team, consisting of 18 physical therapists, 10 physical therapist assistants, and 2 physical therapist students, used the Acute Care Index of Function (ACIF)[4,5] to measure physical function. The ACIF was developed as a measure of physical function for use in an acute care setting. The interrater reliability and concurrent and predictive validity of the ACIF have been established in previous studies.[4,5]

The original 20-item ACIF had four subscales: mental status, bed mobility, transfers, and mobility. As very few of these patients used wheelchairs as their primary means of mobility, the wheelchair mobility items were seldom rated. Because most south Florida residents do not need to climb stairs to enter their homes and because easily accessible test stairs were difficult to locate when patients were seen at bedside, the stair-climbing items also were seldom rated. This produced a mobility subscale with only two items. As interrater reliability had been established at the item level rather than at the subscale level, we combined the two remaining mobility subscale items with the six transfer subscale items to form a new transfer/mobility subscale. Combining these items allowed us to examine changes in mobility and transfers separately from changes in bed mobility and mental status. Because the transfer and mobility subscales were weighted equally in the original calculation of the ACIF total score, a total ACIF score calculated using a two-item mobility subscale and a six-item transfer subscale was no different than a total ACIF score calculated using the new transfer/mobility subscale.

Mental status subscale items represent behaviors such as the ability to follow verbal commands, which are rated as present or absent. The four bed mobility subscale items and the eight transfer/mobility subscale items were rated on a three-point scale based on the degree of assistance required to perform an activity. Patients were rated "unable" if they could not physically assist in performing the activity, "dependent" if they assisted in performing the activity but required physical or verbal assistance to complete the activity, and "independent" if they performed the activity, meeting all stated criteria, without verbal or physical assistance.

Subscale scores were calculated by adding the points for all subscale items, dividing the sum by the total points possible for that subscale, and multiplying the resulting number by 100. The total score was calculated by taking a weighted average of the subscale scores.

Use of the ACIF by the acute care physical therapy staff was initiated as part of a quality assurance program. Prior to beginning use of the ACIF, all members of the physical therapy staff were trained in its administration and scoring using case studies.

The physical therapy staff used the ACIF to measure the functional status of each patient at the time of the first physical therapy visit, every 3 days thereafter during the hospitalization, and a final time within 48 hours before discharge.

Duration of physical therapy. The billing records for the physical therapy department contained information on the dates and duration of physical therapy for each patient. The duration of physical therapy sessions was recorded as units, with each unit representing 15 minutes of treatment. Duration of physical therapy was recorded for both morning and afternoon sessions.

Medical and demographic information. The patient's age, gender, hospital admission date, and primary and secondary diagnoses were recorded on the medical record face sheet. A copy of this face sheet was attached to the physical therapy department's quality assurance records. The quality assurance records also contained documentation of the patient's discharge destination, discharge date, impairments, and treatment plan. Length of hospitalization was calculated using hospital admission and discharge dates. All data were kept in the physical therapy department's quality assurance file.

Data collection. We developed a set of standardized data collection forms, operational definitions, and decision rules that were used to record and code the data from the quality assurance records and medical records. Prior to beginning data collection, all investigators were trained by a member of the research team in the use of the operational definitions and decision rules. All data collectors had to demonstrate 100% agreement with the trainer and with their own previous performance before beginning data collection.

We assigned an identification number to each patient. Because some patients had been admitted more than once during the study period, we also assigned an identification number to each admission. To maintain confidentiality, we did not record the patient's name or medical record number on any of the forms. We maintained a code list linking identification numbers to patients so that we could check the accuracy of data during data verification procedures.

Data Analysis

All data were managed and analyzed using SAS Version 6.08 statistical software(*) on a VAX mainframe computer.[6] Descriptive statistics were generated to characterize the patients. Paired t tests were calculated to compare initial and discharge ACIF total and subscale scores. Student's t tests were calculated to compare the characteristics of patients discharged to their home and those of patients discharged to other destinations. Multiple linear regression was used to explore the relationship between the duration of physical therapy and functional status at discharge. There are a number of factors other than the amount of physical therapy a patient received that could be expected to influence functional status at the time of discharge from the hospital. We suspected that functional status at the time of discharge from the hospital might be influenced by such things as functional status when first seen by the physical therapist, the length of the hospital stay, age, and general health status as indicated by the number of secondary diagnoses. Therefore, we included these variables in the initial regression models to examine and control for their effect.

Results

The patients had a mean age of 67.9 years (SD = 20.5, range = 18-101) (Tab. 1). The majority of the patients (66.5%) were female, 37.9 % were married, and 88.5% were admitted from their home, but only 37.0 % were discharged to their home (Tab. 2). Fracture of the hip or pelvis was the most common diagnosis (47.3%), followed by fracture of the tibia or fibula (13.3%) and other foot or ankle problems (9.8%) (Tab. 2).

Table 1.
Patient Characteristics (N = 173)

Characteristic                [bar] X   SD      Range

Age (y)                        67.9     20.5   18-101
Minutes of physical therapy   238.5    153.6   15-1,110
Days of physical therapy        3.5      2.8    0-17
Days in hospital                6.4      4.2    1-27
Days in hospital before
 physical therapy               2.9      3.2    0-27

Initial evaluation
  ACIF(a) total score          41.75    22.5    0-83.6
  Mental status score          77.62    30.7    0-100
  Bed mobility score           42.72    30.7    0-100
  Transfer/mobility score      30.27    21.87   0-78.18

Discharge evaluation
  ACIF total score             56.5    20.62   0-100
  Mental status score          83.92   26.93   0-100
  Bed mobility score           57.63   31.59   0-100
  Transfer/mobility score      46.97   20.41   0-100

(a) ACIF=Acute Care Index of Function.[4,5]

Table 2.
Patient Characteristics

                                Frequency   Percentage

Female                            115         66.5
Married                            64         37.9
Mental impaired                    22         12.7
Admitted from home                148         88.5
Discharged to home                 64         37.0

Primary diagnosis
  Fracture of hip or pelvis        82         47.3
  Total hip replacement             7          4.1
  Other hip problem                 9          5.2
  Total knee replacement           13          7.5
  Distal/mid femur fracture        10          5.8
  Other knee problem               12          7.0
  Tibia/fibula fracture            23         13.3
  Other foot or ankle problem      17          9.8

The patients were hospitalized for an average of 6.4 days, with the length of hospitalization ranging from 1 to 27 days (Tab. 1). On average, patients were seen for only 3.5 days of physical therapy, during which they received an average of 238.5 minutes of treatment.

On average, the patients demonstrated improvement in the ACIF total score and in all three subscale scores. The largest improvement was in the transfer/mobility subscale, with a mean improvement of 18.42 points (P = .0001). The smallest improvement was in the mental status subscale, which improved, on average, only 6.27 points (P = .0002) (Tab. 3).

Table 3. Initial and Discharge Acute Care Index of Function (ACIF)[4,5] Total Scale and Subscale Scores

                          Initial   Final    Mean
Subscale                   Score    Score   Change   P(a)

Mental status score        77.62    83.9    6.27     .0002
Bed mobility score         42.72    57.63   16.54    .0001
Transfer/mobility score    30.27    46.97   18.42    .0001
ACIF total score           41.75    56.46   15.42    .0001

(a) Paired t test.

Only 8.2% of the patients were independent in ambulation with an assistive device at the time of their first physical therapy visit, although 20.8% of the patients were independent at discharge (P = .0057). Only 19.4% of the patients were unable to ambulate with an assistive device by the time they were discharged, as compared with 50.6% initially (P = .0001).

There are a number of factors other than the amount of physical therapy a patient received that could influence the patient's functional status at the time of discharge from the hospital. Therefore, we used multiple linear regression to examine the relationship between functional status at discharge and the amount of physical therapy received by a patient, while controlling for age, length of hospital stay, and number of diagnoses. Functional status at discharge was the dependent variable, amount of physical therapy (in minutes) received by a patient was the independent variable, and age, number of days hospitalized, number of diagnoses, and initial functional status were the possible confounding variables. The amount of physical therapy that a patient received explained much of the variance in this full regression model, as did age and initial functional status (Tab. 4). Although the number of diagnoses explained almost 1% of the variance in the regression model, it failed to achieve statistical significance. Not only did the number of days hospitalized not contribute importantly to the model, it explained none of the variance.

Table 4. Relationship Between Minutes of Physical Therapy and Functional Status at Discharge(a)

                      Standardized            Partial
Variable              Coefficient      P     Correlation

Intercept                0.00        .0001
Initial ACIF(b)          0.56        .0001   .386
Minutes of physical
 therapy                 0.15        .0489   .039
Days hospitalized        0.01        .8742   .000
Age                     -0.31        .0001   .159
No. of diagnoses        -0.10        .1466   .009

(a) The amount of variance explained by the regression model was [R.sup.2] = .5960.

(b) ACIF = Acute Care Index of Function.[1,5]

A model consisting of those variables achieving statistical significance (Tab. 5) predicted 59% of the variance in functional status at discharge ([R.sup.2] =.59). In this model, the amount of physical therapy that a patient received was an important predictor of functional status at the time of discharge (P = .0173) and by itself predicted almost 6% of the variance in functional status at discharge (Tab. 5).

Table 5. Relationship Between Minutes of Physical Therapy and Functional Status at Discharge(a)

Variable              Coefficient     P       Partial
                                            Correlation

Intercept               49.92       .0001
Initial ACIF(b)          0.58       .0001   .385
Minutes of physical
 therapy                 0.03       .0173   .055
Age                     -0.35       .0001   .184

(a) The amount of variance explained by the regression model was [R.sup.2] = .5872.

(b) ACIF = Acute Ore Index of Function.[4.1]

Discussion

The patients in our study were predominantly elderly. with a mean age of 67.9 years. This finding is not surprising given that a much larger proportion of individuals aged 65 years or older are hospitalized at any point in time than are younger individuals.[7]

The most common diagnosis seen was fracture of the hip or pelvis, which could be expected based on the age of our subjects (Tab. 2). The risk of hip fracture increases with increasing age.[8]

The length of stay in acute care hospitals has been decreasing for more than a decade. After the initiation of the PPS, the length of stay in acute care hospitals deceased 24%,[9] with an average stay for patients with hip fractures of 12.6 days.[2] We found a length of stay in acute care settings of 6.4 days for patients with orthopedic problems, which was substantially shorter than either of these figures. This short length of stay partially explains the finding that, although 88.5% of the patients in our study were admitted from their home, only 37.0% of the patients were discharged to their home. The patients who were not discharged to their home had much lower initial functional scores than did patients who were discharged to their home. Although their scores improved, they did not reach a level of function adequate to return to their home.

Because patients are staying in hospitals for such short periods of time, the amount of physical therapy that they can receive is limited. On average, the patients in our study received only 4 hours of physical therapy over a 4-day period. The patients' functional status, however, improved, as demonstrated by an average increase of 15.4 points in ACIF score from evaluation to discharge. Initially, over 50% of the patients were unable to ambulate with an assistive device. By discharge, however, only 19.3% of the patients remained unable to ambulate. Thus, nearly 80% of the patients were ambulating to some degree at discharge.

The fact that patients improved in physical function during hospitalization does not establish the effectiveness of physical therapy. This study demonstrated, however, that the number of minutes of physical therapy that a patient received could be used to predict functional status at discharge. The longer the duration of physical therapy, the more a patient's functional status increased, as measured by the ACIF. This relationship remained after taking into account the possible effect of length of stay in the hospital, age, number of diagnoses, and initial functional status. Although this type of evidence is not as strong as evidence that would be provided by a clinical trial, the evidence supports the assumption that physical therapy plays an important role in improving physical function of patients in acute care settings.

This study has several limitations. Data were gathered from only one hospital. The patients seen at this hospital may differ in some ways from those seen in hospitals in other regions of the country. The type of physical therapy services provided to these patients also may differ from the physical therapy services provided in other acute care settings. It would be very useful to replicate this study in a variety of acute care settings.

This study is also limited because we only explored how the duration of physical therapy related to changes in functional status for patients in a broad diagnostic category. We need to examine this relationship in much greater detail. Future studies should use specific physical therapy diagnostic categories rather than the broad medical diagnostic categories used in this study. Within these physical therapy diagnostic categories, we need to identify those treatment modalities that produce the greatest improvement in functional status in the time available. In addition, we need to examine the relationship between physical therapy and changes in functional status in other settings such as inpatient rehabilitation and home care.

Conclusion

This study provides evidence that the amount of physical therapy that a patient with orthopedic problems receives in an acute care setting is directly related to the degree of improvement in function that occurs during the hospitalization. Although this study could not answer all of the questions relevant to the delivery of physical therapy services in the acute care setting, it provides a starting point for further, much-needed work.

(*) SAS Institute Inc, PO Box 8000, Cary, NC 27511.

References

[1] Fitzgerald JF, Moore PS, Dittus RS. The care of elderly patients with hip fracture: changes since implementation of the prospective payment system. N Engl J Med. 1988;319:1392-1397.

[2] Jalovaara P, Berglund-Roden M, Wingstrand H, Thorngren KG. Treatment of hip fracture in Finland and Sweden. Acta Orthop Scand. 1992;63:531-535.

[3] Steffen TM, Meyer AD. Physical therapists' notes and outcomes of physical therapy: a case of insufficient evidence. Phys Ther. 1985;65: 213-217.

[4] Van Dillen LR, Roach KE. Reliability and validity of the Acute Care Index of Function for patients with neurologic impairment. Phys Ther. 1988;66:1098-1101.

[5] Roach KE, Van Dillen LR. Development of an acute care index of function for patient with neurologic impairment. Phys Ther. 1988;66: 1102-1108.

[6] SAS/STAT User's Guide, Version 6, Volumes 1 and 2. 4th ed. Cary, NC: SAS Institute Inc; 1989.

[7] Verbrugge LM. From sneezes to adieux: stages of health for American men and women. In: Ward RA, Tobin SS, eds. Health in Aging. Sociological Issues and Policy Directions. New York, NY. Springer Publishing Co Inc; 1987:17-57.

[8] Wolinsky FD, Fitzgerald JF. The risk of hip fracture among noninstitutionalized older adults. J Gerontol. 1994;49:5165-5175.

[9] Kahn KL, Keeler EB, Sherwood MJ, et al. Comparing outcomes of care before and after implementation of the DRG-based prospective payment system. JAMA. 1990;264:1984-1988.

KE Roach, PhD, PT, is Assistant Professor, Division of Physical Therapy, Department of Orthopaedics and Rehabilitation, University of Miami School of Medicine, 5915 Ponce de Leon Blvd, Coral Gables, FL 33146 (USA) ([e-mail]). Address all correspondence to Dr Roach.

D Ally, PT, is Staff Therapist, Parkway Regional Medical Center, North Miami, Fla. She was a graduate student in the Division of Physical Therapy, Department of Orthopaedics and Rehabilitation, University of Miami School of Medicine, at the time this study was conducted.

B Finnerty, PT, was a graduate student in the Division of Physical Therapy, Department of Orthopaedics and Rehabilitation, University of Miami School of Medicine, at the time this study was conducted.

D Watkins, PT, is Staff Therapist, Outpatient Service, Miami Veterans Affairs Hospital, Miami, Fla. He was a graduate student in the Division of Physical Therapy, Department of Orthopaedics and Rehabilitation, University of Miami School of Medicine, at the time this study was conducted.

BA Litwin, PT, is Health Care Consultant, Polaris Group, Plantation, Fla. She was Director, Rehabilitation Services, Parkway Regional Medical Center, at the time this study was conducted.

Bjanz-Hoover, PT, was Senior Therapist and Center Coordinator of Clinical Education, Parkway Regional Medical Center, at the time this study was conducted.

T Watson, PT, is Senior Therapist, Acute Care Service, Parkway Regional Medical Center.

KA Curtis, PhD, PT, is Associate Professor, Physical Therapy Department, California State University, Fresno, Calif.

This project was supported by Foundation for Physical Therapy Research Grant 94R-05-ROA-01.

This study was approved by the University of Miami Medical Sciences Subcommittee for the Protection of Human Subjects in Research.

This article was submitted November 4, 1996, and was accepted June 27, 1997.