Physical therapy treatment choices for musculoskeletal impairments.

[Jette AM, Delitto A. Physical therapy treatment choices for musculoskeletal impairments. Phys Ther. 1997;77: 145-154.]

Key Words: Health services research, Low back pain, Musculoskeletal system, Orthopedics, Outcomes.

In the United States, 10.9% of all medical office visits are primarily for musculoskeletal pathology or impairments, with 2% of all visits for back symptoms.[1] Physical therapy is frequently part of the conservative (nonsurgical) care for a myriad of musculoskeletal disorders. For example, over 80% of physicians from a variety of specialties who were surveyed nationally identified physical therapy as a preferred management strategy for low back pain.[2] Although few data exist on the actual use of outpatient physical therapy, in the second National Health and Nutrition Examination Survey (NHANES II), over 16% of the sample of adults who sought care from a health care professional for the management of low back pain received care from a physical therapist. Frequency of use of physical therapy did not vary by race, region of the country, or level of education.[3]

In spite of frequent utilization, however, little is known about the physical therapy treatments provided to patients with musculoskeletal disorders. Some members of the physical therapy profession have viewed physical therapy intervention as synonymous with procedures, including the administration of physical agents (eg, heat, cold, electrical currents), massage, and manual or manipulative therapy.[4] This view has been shared by some persons outside the field. Many state physical therapy practice acts, for instance, actually define the role of the physical therapist as predominantly an administrator of physical agents. In contrast to these traditional perceptions, contemporary scholars within the profession have reasoned that physical therapy treatment decisions should (do) follow a decision-making process that flows from a careful evaluation by the therapist (eg, pain, limitations of strength and range of motion) that leads to an impairment diagnosis, with subsequent treatment choices based on that diagnosis.[5-7] Following this approach, a physical therapy episode of care for a patient with an acute back injury may first entail predominantly passive modes (eg, modalities or manipulation) directed toward alleviating pain during the acute period, progressing to more active therapy (eg, aerobic exercise and work hardening) as the severity subsides. Such an approach suggests that the dynamics of a physical therapy episode of care may be complex and involve different treatment choices that derive from the clinical examination findings observed across the phases of an episode of care or that are based on the chronicity of the injury. Consistent with this view of physical therapy, as the course of a physical therapy episode of care progresses, the physical therapy treatment program should shift from more passive treatments to more active forms of intervention.

Little is known about the degree to which physical therapy practice patterns match theory or more traditional perceptions within and outside of the profession. In the United States, for example, Battie et al[8] sampled 293 physical therapists working in both managed care and fee-for-service environments to evaluate their treatment preferences for patients with low back pain. They found that, in general, patient education (eg, body mechanics), strengthening exercises, flexibility exercises, and aerobic exercises were the treatments most often preferred. They also reported that treatment preferences varied depending on practice setting (eg, managed care versus fee-for-service) and chronicity of the patient's condition. These data, however, only reflect therapists' stated preferences and not their actual behavior.

The only data on actual treatment use that we could locate came from the NHANES II study, where subjects with low back pain reported on specific treatment choices. In this sample, almost three quarters of the subjects reported using heat, 40% reported using exercises or other types of physical therapy, 21% reported using traction, and 7% reported using ice as part of their treatment of low back pain.[3]

The current analyses were undertaken to describe out-patient physical therapy treatments provided for patients with musculoskeletal impairments using a database containing physical therapy practice information on patients with spinal (cervical and lumbar) and knee impairments. Patients with these impairments represent a majority of outpatients seen by physical therapists in the United States.[9] The chief aims of these analyses were

1. To describe the treatments administered during the first, middle, and final thirds of an outpatient physical therapy episode of care for each impairment group.

2. To examine the mix of active and passive treatments provided to these outpatients and test the hypothesis that use of passive treatments would decrease during the course of an episode of care. Treatments with patient involvement were coded as active. Treatments done to the patient were coded as passive.

3. To determine whether treatment choices varied as a function of chronicity of the condition or by type of reimbursement.

Based on Battie and colleagues' findings on physical therapists' treatment preferences,[8] we hypothesized that treatment choices would vary under managed care versus fee-for-service reimbursement and as a function of time since onset of the injury.

Method

Subjects

Patients in this study received outpatient physical therapy for a primary orthopedic impairment between July 1993 and June 1994 from one of 68 physical therapy practices participating in a large database generated by the Focus on Therapeutic Outcomes (FOTO) network. The FOTO network, at the time of this study, was a privately funded consortium of six outpatient rehabilitation companies that was developed for the purpose of generating standardized outcome-oriented information for use in outpatient physical therapy settings. Companies were required to maintain an agreed-on data quality standard and record completion rate to remain in the FOTO network. The database contained information from 5,061 patients admitted for physical therapy for knee, lumbar, and cervical impairments, out of which discharge information was available for 2,598 completed episodes of care. Of these episodes of care, 1,279 episodes were for a lumbar impairment, 613 episodes were for a cervical spine impairment, and 706 episodes were for a knee impairment. The project was reviewed and approved by the Human Subjects Review Board of New England Research Institutes, which was responsible for the design and administration of the database during its first year of operation.

Data Collection

Data profiling each practice were provided by the practice managers on entry into the FOTO project and included practice location in terms of state and region of the United States, number of full-time-equivalent physical therapists, and number of new patients seen in an average day. Data profiling each physical therapist were provided by the individual practitioner on entering the FOTO network. Data on patient demographic background and duration since the onset of symptoms were obtained by patient self-report at the initial physical therapy visit.

The primary physical therapist obtained information on each patient's surgical history and referral source when the patient was first seen. For our analysis, impairments were defined as acute, subacute, or chronic. For spinal injuries, we used the following guidelines from the Quebec Task Force on Spinal Disorders[10]: 0-14 days since onset of injury=acute, 14-180 days since onset of injury=subacute, and [is greater than] 180 days since onset of injury=chronic. For knee disorders, we established the following criteria: 0-21 days since onset of injury=acute, 21-90 days since onset of injury=subacute, and [is greater than] 90 days since onset of injury=chronic. Method of payment was classified as either managed care (including health maintenance organizations and preferred provider organizations) versus fee for service (including private insurers, Medicare, Medicaid, Workers' Compensation, self-pay).

During the discharge visit, the primary physical therapist provided information on primary payment source, duration of the episode of care, number of visits, and treatments provided during the initial, middle, and final thirds of the episode of care, defined by the duration of the episode of therapy. An episode of care consisted of all therapy visits provided between the initial and discharge visits for the primary impairment, as defined by the physical therapist.

Data Analysis

As a first step, frequency distributions were calculated for all physical therapy treatments provided to each patient during each third of the episode of physical therapy. For each of the three types of impairment, a repeated-measures logistic analysis model was used to test for changes in each of the treatments across the episodes of care. The statistical significance of changes in treatments across each third of the episode was determined using a repeated-measures logistic analysis with a compound symmetric covariance matrix. Changes in treatments were assessed with the Wald chi-square test statistic.[11]

Therapists' treatment choices during the total episode of care was reported by type of reimbursement and by duration since the onset of the primary impairment or surgery for the primary impairment and compared using the chi-square statistic for equal proportions. These comparisons were done separately for patients with knee impairments and those with spinal impairments. Lumbar and cervical impairment episodes were combined into one spinal impairment group for these comparisons because the findings were very similar for both.

The final phase of the analysis consisted of examining contingency tables for each of the three types of impairments to investigate shifting patterns in choice of treatment. For each impairment, a 3 x 3 table of passive treatments, active treatments, or combinations of active and passive treatments for each of the initial and final phases of the episode of care was constructed. The statistical significance of shifts around the diagonal was determined using Bowker's test for symmetry, an extension of the McNemar statistic.[12]

An alpha level of .01 was used for all pair-wise comparisons to account for multiple testing. All analyses were conducted with the Statistical Analysis System.*

The physical therapy interventions were combined into passive and active categories as follows. Passive treatments consisted of cryotherapy, diathermy, electrical stimulation for pain, ice, iontophoresis, phonophoresis, massage, moist heat, myofascial techniques, muscle energy techniques, ultrasound, and mobilization and manipulation. Active treatments consisted of electromyography/biofeedback, electrical stimulation for strength, orthotic training, craniocervical techniques, joint mobility exercises, exercises to modulate pain, flexibility exercises, aquatic exercises, body mechanics training, gait training, strengthening exercises, home exercises, postural exercises, closed-chain exercises, endurance exercises, functional training, proprioceptive exercises, plyometrics, balance training, and running/ agility drills. A "user's manual" provided to all practices contained brief definitions of the treatments.

Results

Practice, Provider, and Patient Characteristics

The 68 outpatient practices in the FOTO network were located across the United States: 2% from the Middle Atlantic States, 20% from the South Atlantic States, 9% from the South Central States, 62% from the North Central States, 7% from the Mountain States, and less than 1% from the Pacific States. The practices ranged in size, with 24% having one full-time physical therapist, 38% having two full-time physical therapists, 15% having three full-time physical therapists, and 23% having more than three full-time physical therapists.

The average age of the 141 physical therapists in the FOTO network was 32.6 years (SD=7.8, range=22-60); 70% of the therapists were women. The highest earned degree of 84% of the physical therapists was a bachelor's degree. Nine percent of the therapists had an entry-level master's degree, and 7% had an advanced master's degree. The average years of practice was 8.4 (SD=7.4, range= 1-33). Eighty-three percent of the therapists worked full-time, treating an average of 50 patients per week.

Fifty-one percent of the patients admitted in the database had complete data on their episode of outpatient physical therapy available in the database. Patients with complete data from the episode did not differ from patients entered in the database but did not have discharge data with respect to age, gender, level of education, practice setting, or health status on admission, as measured by the SF-36.[13] Patients with a knee injury who had complete data were more likely to be [is less than] 22 days since onset of injury compared with those patients with knee impairments who did not have complete data (83% versus 32%; P=.0001). Patients with a spinal impairment who had complete data did not differ from those with incomplete data with respect to acuity of symptoms. Table 1 shows the characteristics of the patients by primary impairment.

Table 1.
Background Characteristics of Patients with Spinal and Knee
Impairments(a)

                                                 Lumbar
                                               Impairment
                                               (n= 1,279)
Age (y)
X                                                  40.8
SD                                                 13.2

Gender
Female                                             48
Male                                               52

Ethnicity
White                                              87
Black                                              10
American Indian                                    [is less than] 1
Asian                                              [is less than] 1
Hispanic                                            2

Employment status (at admission)
Full-time                                          38
Light duty                                         13
Off due to health                                  31
Retired                                             7
Unemployed                                          9
On disability                                       2

Household income
[is less than] $15,000                              9
$ 15,000-$25,000                                   18
$26,000-$35,000                                    17
$36,000-$45,000                                    15
[is greater than] $45,000                          24
Refused to answer                                  17

Payment source
Private pay                                         6
Medicare/Medicaid                                   5
Preferred provider organization                    11
Health maintenance organization                    25
Workers' Compensation                              35
Other insurance                                    18
Duration since onset of symptoms
0-21 d                                             ...
[is greater than] 21 - [is less than] 3 mo         ...
[is greater than] 3 mo                             ...
0-14 d                                              22
[is greater than] 14 - [is less than] 6 mo          48
[is greater than or equal to] 6 mo                  30

                                                         Cervical
                                                        Impairment
                                                         (n=613)
Age (y)
X                                                          40.9
SD                                                         12.6

Gender
Female                                                       64
Male                                                         36

Ethnicity
White                                                        85
Black                                                        10
American Indian
Asian
Hispanic                                                      3

Employment status (at admission)
Full-time                                                    52
Light duty                                                   11
Off due to health                                            15
Retired                                                       8
Unemployed                                                   12
On disability                                                 2

Household income
[is less than] $15,000                                        7
$ 15,000-$25,000                                             17
$26,000-$35,000                                              20
$36,000-$45,000                                              16
[is greater than] $45,000                                    25
Refused to answer                                            15

Payment source
Private pay                                                   9
Medicare/Medicaid                                             5
Preferred provider organization                              12
Health maintenance organization                              33
Workers' Compensation                                        14
Other insurance                                              27
Duration since onset of symptoms
0-21 d                                                       ...
[is greater than] 21 - [is less than] 3 mo                   ...
[is greater than] 3 mo                                       ...
0-14 d                                                        21
[is greater than] 14 - [is less than] 6 mo                    51
[is greater than or equal to] 6 mo                            28

                                                          Knee
                                                       Impairment
                                                        (n=706)
Age (y)
X                                                         41.2
SD                                                        14.1

Gender
Female                                                      52
Male                                                        48

Ethnicity
White                                                       89
Black                                                        7
American Indian                                              1
Asian                                                        1
Hispanic                                                     2

Employment status (at admission)
Full-time                                                   38
Light duty                                                  10
Off due to health                                           30
Retired                                                      8
Unemployed                                                  11
On disability                                                3

Household income
[is less than] $15,000                                       9
$ 15,000-$25,000                                            14
$26,000-$35,000                                             16
$36,000-$45,000                                             15
[is greater than] $45,000                                   27
Refused to answer                                           19

Payment source
Private pay                                                  5
Medicare/Medicaid                                            4
Preferred provider organization                             13
Health maintenance organization                             31
Workers' Compensation                                       24
Other insurance                                             23
Duration since onset of symptoms
0-21 d                                                      53
[is greater than] 21 - [is less than] 3 mo                  37
[is greater than] 3 mo                                      30
0-14 d                                                      ...
[is greater than] 14 - [is less than] 6 mo                  ...
[is greater than or equal to] 6 mo                          ...

(a) All values are percentages, unless otherwise noted.

Treatment Choices

Tables 2 through 4 display the frequency distributions of therapists' treatment choices for the management of lumbar, cervical, and knee impairments for each third of the episode. For lumbar impairments, the most commonly used treatments across all phases of the episode were multimodal exercises (exercises that combined various purposes), followed in frequency of use by heat modalities, flexibility exercises, and strengthening exercises. All other initial treatments occurred in fewer than half of the episodes. The data revealed a change in treatment pattern across the episode of care in the hypothesized direction, with a decrease in use of most modalities and manual techniques and an increase in use of strengthening and endurance exercises in the later stages of the episode. Mobilization was used in 27% of initial treatment plans and decreased in use in the final third of the episode. Manipulation techniques were used infrequently.

Table 2. Physical Therapy Treatments Used in Episodes of Care for Lumbar Impairments (N= 1,279)(a)

                                Stage of Episode of Care
                                  Initial      Middle     Final
Treatment Used                    Third        Third      Third

Modality
Heat(b)                            79.1        70.2       59.4
Cold(b)                            12.8        10.7        9.0
Electrical stimulation(b)          39.6        30.9       22.2
lontophoresis/phonophoresis         2.9         3.4        2.4
Massage(b)                         28.2        24.3       17.2
Myofascial techniques(b)            6.3         5.5        3.9
Devices                             2.8         2.5        2.6

Exercise
Strengthening(b)                   50.4        64.9       68.8
Endurance(b)                       24.7        37.7       41.5
Flexibility(b)                     75.1        72.3       67.5
Multimodal(b,c)                    89.9        82.1       81.2

Mobilization/manipulation
Mobilization(b)                    27.2        24.2       16.4
Manipulation(b)                     3.7         0.7        0.6

(a) All values are percentages.

(b) Repeated-measures logistic analysis with the Wald chi-square test statistic; P [is less than] .01.

(c) Multimodal=exercises that combine various purposes, including aquatic exercises, body mechanics, exercises to modulate pain, functional training, home exercises, postural exercises, proprioceptive exercises, closed-chain exercises, plyometrics, and running/agility skills.

Tables 3 and 4 characterize the pattern of treatment choices for episodes of care in patients with cervical and knee impairments. Treatment profiles in the cervical episodes of care included a high frequency of use of heat, mobilization, massage, multimodal exercises, and flexibility exercises. Strengthening and endurance exercises increased in the later portions of the episode of therapy.

Table 3. Physical Therapy Treatments Used in Episodes of Care for Cervical Impairments (N=613)(a)

                                Stage of Episode of Care
                                  Initial      Middle     Final
Treatment Used                    Third        Third      Third

Modality
Heat(b)                            87.6        82.4       73.7
Cold                                8.2         7.8        6.7
Electrical stimulation(b)          31.5        27.2       20.4
lontophoresis/phonophoresis         2.9         2.8        2.9
Massage(b)                         48.9        46.5       39.6
Myofascial techniques(b)           12.7        12.5        9.5
Devices                             0.6         0.5        0.3
Exercise
Strengthening(b)                   27.4        46.7       52.8
Endurance(b)                       14.0        23.5       27.9
Flexibility(b)                       76.0        72.1       67.4
Multimodal(b,c)                    87.8        81.2       81.2
Mobilization/manipulation
Mobilization(b)                    41.9        39.0       33.1
Manipulation                        1.8         0.8        1.1

(a) All values are percentages.

(b) Repeated-measures logistic analysis with the Wald chi-square test statistics; P [is less than] .01.

(c) Multimodal=exercises that combine various purposes, including aquatic exercises, body mechanics, exercises to modulate pain, functional training, home exercises, postural exercises, proprioceptive exercises, closed-chain exercises, plyometrics, and running/agility skills.

Table 4. Physical Therapy Treatments Used in Episodes of Care for Knee Impairments (N=706)(a)

                                Stage of Episode of Care
                                  Initial      Middle     Final
Treatment Used                    Third        Third      Third

Modality
Heat(b)                           29.5         25.5       22.7
Cold(b)                           57.5         50.8       41.4
Electrical stimulation(b)         21.1         15.9        9.1
lontophoresis/phonophoresis(b)     7.9          6.5        4.8
Massage(b)                         8.4          7.2        4.8
Myofascial techniques              2.5          2.1        1.0
Devices                            4.4          4.2        5.2

Exercise
Strengthening                      88.8        88.2       86.8
Endurance(b)                       42.5        58.5       63.9
Flexibility(b)                     77.9        69.3       62.0
Multimodal(c)                      75.2        73.8       74.9

Mobilization/manipulation
Mobilization(b)                    28.5        20.8       10.2

(a) All values are percentages.

(b) Repeated-measures logistic analysis with the Wald chi-square test statistic; P [is less than] .01.

(c) Multimodal=exercises that combine various purposes, including aquatic exercises, body mechanics, exercises to modulate pain, functional training, home exercises, postural exercises, proprioceptive exercises, closed-chain exercises, plyometrics, and running/agility skills.

Treatment choices for episodes of care in patients with knee injuries were characterized by a high frequency of use of all types of exercise and frequent use of cold modalities. Strengthening, flexibility, and multimodal exercises were used in more than three quarters of the initial stages of these episodes of therapy, with endurance exercises used in a majority of the later stages of the episodes.

Table 5 displays the frequency of use of different treatments by type of reimbursement. Overall, the frequency of treatments used for both spinal and knee impairments was quite similar in managed care versus fee-for-service environments. Differences between payment types were observed for 6 out of 25 analyses. Therapists working with patients with knee impairments covered by a fee-for-service arrangement more frequently used massage and devices than did therapists working with patients with knee impairments in a managed care environment. Therapists working with patients with spinal impairments under a fee-for-service arrangement were more likely to use devices, strengthening exercises, and endurance exercises and less likely to use heat modalities than were therapists working with patients with spinal impairments reimbursed through a managed care relationship.

Table 5 Physical Therapy Treatments Used in Episodes of Care for Knee and Spinal Impairments by Source of Payment(a)

                              Knee Impairments
                              Source of Payment
                              Managed Care        Fee-for-Service
Treatment Used                (n=308)             (n=398)

Modality
Heat                            38.0              34.1
Cold                            60.0              62.3
Electrical stimulation          24.3              23.3
lontophoresis/phonophoresis     11.3               9.3
Massage                          6.5              14.0(b)
Myofascial techniques            2.9              4.0
Devices                          4.2             10.3(b)

Exercise
Strengthening                   94.1             95.7
Endurance                       68.1             67.1
Flexibility                     80.8             79.6
Multimodal                      81.5             86.1

Mobilization/manipulation
Mobilization                    34.1             29.4
Manipulation                    ...              ...

                              Spinal Impairments
                              Source of Payment
                              Managed Care        Fee-for-Service
Treatment Used                (n=731)             (n=1,155

Modality
Heat                            88.1             83.2(b)
Cold                            11.9             14.4
Electrical stimulation          43.3             38.1
lontophoresis/phonophoresis      5.2              3.7
Massage                         40.5             39.5
Myofascial techniques           10.8             10.7
Devices                          1.1              3.3(b)

Exercise
Strengthening                   64.1             73.0(b)
Endurance                       33.9             43.4(b)
Flexibility                     80.3             82.0
Multimodal                      93.7             93.6

Mobilization/manipulation
Mobilization                    41.7             38.8
Manipulation                     3.1              4.0

(a) All values are percentages.

(b) Chi-square test of equal proportions; P [is less than] .01.

The data in Table 6 show that therapists were more likely to use devices in their episodes of care with patients with knee injuries of long duration. No other treatment differences by duration since onset of injury were observed for knee episodes of care. Treatment choices differed in two aspects for spinal episodes of care as a function of time since onset of injury. Frequency of use of heat and electrical stimulation decreased as chronicity increased.

Table 6. Treatments Used in Episodes of Core for Spinal and Knee In juries by Duration Since Onset of Symptoms/Surgery(a)

                             Knee Impairments
                             Days Since Onset of Symptoms/Surgery(b)

                                0-21      22-90      91+
Treatment Used                (n=213)    (n=236)   (n=191)

Modality
Heat                            41        336.0     32.0
Cold                            62        466.1     58.6
Electrical stimulation          21        127.1     24.6
lontophoresis/phonophoresis     13.1        9.7      8.9
Massage                          9.8       11.0     12.0
Myofacial techniques             4.7        2.5      3.6
Devices                          6.5        4.6     12.5(d)

Exercise
Strengthening                   93.9       96.2     93.7
Endurance                       67.6       64.8     71.2
Flexibility                     82.1       81.0     78.0
Multimodal                      81.7       81.3     89.5

Mobilization/manipulation
Mobilization                    33.3       30.0     31.4
Manipulation                    ...        ...      ...

                             Spinal Impairments
                             Days Since Onset of Symptoms/Surgery(c)

                                0-14      15-18     0180+
Treatment Used                (n=412)    (n=916)   (n=548)

Modality
Heat                            89.8      86.0      80.2(d)
Cold                            16.7      13.6      11.1
Electrical stimulation          49.0      40.2      33.5(d)
lontophoresis/phonophoresis      3.1       4.9       4.0
Massage                         38.3      40.5      40.0
Myofacial techniques             8.5      10.0      13.5
Devices                          1.9       3.3       2.0

Exercise
Strengthening                   67.7      69.6      71.1
Endurance                       36.9      38.8      44.5
Flexibility                     81.5      80.6      82.4
Multimodal                      93.9      93.2      93.8

Mobilization/manipulation
Mobilization                    38.6      38.9      42.8
Manipulation                     5.8       3.0       2.9

(a) All values are percentages.

(b) Missing days since onset of symptoms/surgery for 66 patients.

(c) Missing days since onset of symptoms/surgery for 16 patients.

(d) Chi-square test of equal proportions; P [is less than] .01.

The Figure displays the shift between use of passive treatments only, active treatments only, and a combination of passive and active treatments between the initial and final thirds of the episode of therapy for patients in all three impairment groups. Use of passive treatments alone was rarely reported across all three impairment groups. Most episodes of therapy were characterized by a combination of passive and active treatments. For patients with lumbar and cervical spine impairments, most passive treatment patterns in the initial third of the episode shifted to mixed treatment or active treatment patterns by the final third of the episode. Use of active treatments alone increased somewhat during the episodes of care. Shifts in the mix of active versus passive treatment patterns from initial to final thirds of the episodes were statistically significant for all three impairment groups (P [is less than .01).

[Figure ILLUSTRATION OMITTED]

Discussion

The findings from this analysis provide new insights into contemporary outpatient physical therapy management for three prevalent musculoskeletal impairments. Contrary to some traditional perspectives of physical therapists as purveyors of physical agents (eg, heat, cold, diathermy), physical therapy during these episodes of care was characterized by diverse combinations of treatments, which included physical agents accompanied by exercise and manual therapy techniques. Only a small minority of episodes included passive treatments alone (eg, physical agents and manual therapy), ranging from 1% to 6% of episodes of therapy. This result, based on data from 2,598 completed episodes of care provided by 141 physical therapists working in 6to private outpatient practices, revealed that physical agents were used frequently in the physical therapy management of patients with lumbar, cervical, and knee disorders, in concert with active interventions.

One explanation for this finding is that various physical agents may have been used as an adjunct to active therapies, as opposed to a primary intervention. For example, active approaches such as exercise may have been preceded or followed by application of thermal agents. Such a strategy is commonly recommended with musculoskeletal disorders such as aggressive muscle strengthening exercises that are followed up by prophylactic icing of the involved structures to prevent post-exercise pain and swelling.4 To clarify such issues, future studies may need to have practitioners delineate which treatments are primarily used to address the patient's impairments versus treatments that are used as adjunctive treatments.

Another finding from this investigation was the observed change in treatments used during the course of an episode of care. As the data shown in Tables 2 through 4 revealed, the general trend across the episodes of care was a decrease in passive treatment approaches (eg, heat, cold, electrical stimulation, manual therapy), with a corresponding increase in more active forms of therapy in the latter stages of the episodes of care. Further research is needed to identify the reasons why therapists choose a specific treatment regimen and why treatment choices change or do not change during the course of an episode of care. Research is currently under way relating treatment choices and other factors to patient outcomes.[14,15]

A surprising finding was the rather low proportion of episodes of care using mobilization and manipulation, particularly for spinal impairments. Manual therapies were used in a higher proportion of episodes of care for cervical impairments than for knee and lumbar impairments, and like other passive approaches, the overall use of manual therapy appeared to decrease over the phases of the episodes of care. The positive role of manual therapy has been well-defined in past research,[16] and its use, particularly in patients with low back pain, has been advocated in clinical practice guidelines in the United States, Great Britain, and the Netherlands. In spite of this, fewer than one third of the episodes of care for lumbar impairments indicated manual therapy as a treatment. Future studies might address the possible reasons for this finding, including the therapists' confidence in their ability to administer manual therapy and state practice acts prohibiting the physical therapists' use of this treatment.

The influence of managed care on physical therapy practice is a topic of considerable debate within and outside of the profession, and these findings suggest that the method of third-party payment (health maintenance organization, preferred provider arrangements, fee for service) was not associated with most treatment choices. Although used infrequently, devices (eg, orthoses) were two to three times more likely to be used in fee-for-service environments than in managed care environments during episodes of care for spinal and knee impairments. Such a finding is not surprising given the restrictions and the control that managed care imposes, particularly on devices such as lumbar corsets. Similarly massage was almost twice as likely to be used in the treatment of knee disorders, provided that the patient was seen in a fee-for-service mode as opposed to a managed care mode. In contrast, heat was used less frequently for spinal impairments under fee-for-service arrangements. What remains unknown, however, is the degree to which differences in the treatment episodes under managed care versus fee-for-service arrangements resulted in different patient outcomes. This is a high priority for future research.

The findings from this study on actual treatment choices for episodes of care for spinal impairments were compared with the stated treatment preferences for back pain management made by a statewide sample of therapists.[8] In general, the sample in the study by Battie et al[8] reported treatment preferences for hypothetical patients with low back pain that were quite similar to the actual treatment choices made by our sample of therapists, especially with respect to frequency of use of preference for exercises, mobilization, and heat modalities. Cold modalities or ice, however, were used with less frequency than they were viewed as a treatment preference, and electrical stimulation was used with more frequency than the stated preference would indicate. To a large degree, the data suggest that there is considerable consensus in the profession as to the treatments indicated for patients with low back injury.

Contrary to Battie and colleagues' comparison of treatment preferences by practice setting,[8] we did not observe differences in the actual use of joint mobilization under managed care versus a fee-for-service reimbursement arrangement. Furthermore, unlike Battie et al, we found more frequent use of strengthening and endurance exercises and less use of heat for episodes of care for spinal impairments under fee-for-service as compared with managed care arrangements. Such discrepancies may have a variety of explanations, not the least of which is the different methodologies of the two studies. Battie et al compared preferences of therapists who were working in different practice settings, whereas we examined therapists' choices for certain treatment strategies reimbursed under different payment arrangements within the same practice setting. Additionally, Battie et al asked practitioners to list their preferences given hypothetical cases, whereas we monitored treatments administered within actual episodes of care. Thus, there could be a difference between what clinicians say they will do in a hypothetical situation and what they actually do in actual patient care. Geographical treatment preferences may have had an effect on these treatment choices, as all of the therapists in Battie and colleagues' sample were from one state.

The limitations with respect to the generalizability of these findings must be kept in mind. The data for this investigation came from a sample of practices and therapists who did not represent any defined population. Although the FOTO database during the period 1993 to 1994 contained practices that were located in different regions of the country and that varied in size, all of the practices were associated with six corporations. In addition, a sizable proportion of patients admitted to the database did not contribute discharge information and were therefore lost to analysis. Although this limitation did not appear to bias the sample with respect to baseline health status and demographic background, the degree to which the discharge sample might be biased with respect to other factors could not be determined. Because we cannot determine the extent to which the findings can be generalized to other outpatient physical therapy practices, it will be important to compare these findings with those of future studies of a similar nature done on different types of samples. An additional potential limitation of the study was that the data on treatment choices were provided by the primary therapist at the completion of each patient's episode of care, and these data were not validated by an independent source. Although this limitation is common to most clinical databases, this feature again highlights the importance of replication.

Summary

The findings from this investigation revealed that out-patient physical therapy intervention for patients with spinal and knee impairments was characterized by a diverse array of modalities, exercises, and manual therapy treatments. Treatment choices varied by type of impairment and across thirds of the episode of care. Fee-for-service payment arrangements were associated with increased use of devices, therapeutic massage, strengthening exercises, and endurance exercises. Most treatment choices were not associated with duration since onset of symptoms. The study's findings revealed that although physical agents were frequently used in physical therapy episodes of care, they were applied along with exercise and manual therapy interventions. Future research should relate the provided treatments to variation in patient outcomes following physical therapy.

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

This study was approved by the Human Subjects Review Board of New England Research Institutes.

This article was submitted November 8, 1995, and was accepted October 7, 1996.

References

[1] DeLozier J, Gagnon R. National ambulatory care survey: 1989 summary. In: Advance Data From Vital and Health Statistics of the National Center for Health Statistics. Hyattsville, Md: National Center for Health Statistics; 1991. Publication no. 203.

[2] Cherkin DC, Deyo RA, Wheeler K, et al. Physician views about treating low back pain: the results of a national survey. Spine. 1995;20: 1-10.

[3] Deyo R, Tsui-Wu Y Descriptive epidemiology of low back pain and its related medical care in the United States. Spine. 1987;12:264-268.

[4] Pinkston D. Evolution of the practice of physical therapy in the United States. In: Scully RM, Barnes MR, eds. Physical Therapy. Philadelphia, Pa: JB Lippincott Co; 1989:2-30.

[5] Sahrmann SA. Diagnosis by the physical therapist--a prerequisite for treatment: a special communication. Phys Ther. 1988;68:1703-1706.

[6] Guccione AA. Physical therapy diagnosis and the relationship between impairment and function. Phys Ther. 1991;71:499-503.

[7] Jette AM. Diagnosis and classification by physical therapists: a special communication. Phys Ther. 1989;69:967-969.

[8] Battie MC, Cherkin DC, Dunn R, et al. Managing low back pain: attitudes and treatment preferences of physical therapists. Phys Ther. 1994;74:219-226.

[9] Jette AM, Smith K, Haley SM, Davis KD. Physical therapy episodes of care for patients with low back pain. Phys Ther. 1994;74:101-110.

[10] Spitzer WO. Diagnosis of the problem (the problem of diagnosis): scientific approach to the assessment and management of activity-related spinal disorders-a monograph for clinicians: report of the Quebec Task Force on Spinal Disorders. Spine. 1987;12(suppl): S16-S21.

[11] Murray D, Wolfinger R. Analysis issues in the evaluation of community trials: progress toward solution. Journal of Community Psychology. 1994(CSAP special issue):140-154.

[12] Bowker A. Bowker's test for symmetry. Journal of the American Statistical Association. 1948;43:572-574.

[13] Stewart Al., Hays RD, Ware JE. The MOS short-form general health survey (SF-36). Med Care. 1988;26:724-732.

[14] Jette DU, Jette AM. Physical therapy and health outcomes in patients with spinal impairments. Phys Ther. 1996;76:930-941.

[15] Jette DU, Jette AM. Physical therapy and health outcomes in patients with knee impairments. Phys Ther 1996;76: 1178-1186.

[16] Shekelle P, Adams A, Chassin M, et al. Spinal manipulation for low-back pain. Amer Intern Med. 1992;117:590-598.

AM Jette, PhD, PT, is Professor and Dean, Sargent College of Allied Health Professions, Boston University, 635 Commonwealth Ave, Boston, MA 02215 (USA) ([e-mail]). Address all correspondence to DrJette.

A Delitto, PhD, PT, is Associate Professor and Chair, Department of Physical Therapy, School of Health and Rehabilitation Sciences, University of Pittsburgh, 101 Pennsylvania Ave, Pittsburgh, PA 15261, and Director of Research, Comprehensive Spine Center, University of Pittsburgh Medical Center.