Back performance scale for the assessment of mobility-related activities in people with back pain. (Research Report).

Measures in rehabilitation often are disease-specific and address health aspects that tend to be affected in a particular group of patients. (1) Such measures may be useful in clinical decision making and in identifying change that is closely related to the condition of interest. The consequences of back pain are described as a melting pot of dysfunction affecting psychological, social, occupational, and family life. (2) Bodily pain, limitation of physical activities, and role limitation due to physical health are frequently occurring problems in this group of patients. (3) We believe such areas of health should be assessed by the use of measurement tools. Each of the 5 physical performance tests included in the Back Performance Scale (BPS) and described in this article have been found to reflect important aspects of physical functioning. (4)

A multitude of self-report questionnaires have been developed to assess pain and daily functioning in patients with back problems. (5-8) Pain is a symptom that depends on the response of the person experiencing it (9) and can be assessed by self-report measures. Such measures are also important because they indicate how patients perceive their daily functioning, and the measures can be used to examine perceived change over time. Self-report measurements also are simple to obtain.

Observation, however, is often considered the most replicable method of assessing functional performance, (10,11) even though data for this assertion often are lacking. Low to moderate agreement often has been demonstrated between self-reported and observed disability. (7,12-15) Salen et al (7) demonstrated a moderate correlation (r=.48) between patients' self-reported difficulty in performing various daily tasks, as assessed by the Disability Rating Index (DRI), and observers' assessments of actual performance. After the patients had traversed an obstacle course, however, the correlation increased substantially to r=.78. Disability ratings derived from a questionnaire have been found to be higher (worse) than those derived from observation. (15) Among 24% of all subjects in a health survey who reported at least one disability, no disability was registered by an external tester who observed the tasks being performed. The correspondence among questionnaire, interview, and clinical examination was addressed in another study. (14) Among those respondents who reported no previous low back trouble on a Nordic questionnaire for the low back, only 63% gave the same report in a personal interview, and they were found not to have low back trouble in a blinded clinical examination. As there may be a mismatch between how patients believe they function and how they function as observed by others, we contend that self-report measures should be supplemented with observational methods to guide treatment and to register change in physical performance over time.

The impact of back pain on physical performance may be classified according to the World Health Organization's International Classification of Functioning and Disability (ICF) (16) into dimensions of impairment, activity (limitation), and participation (restriction). Traditional physical tests tend to address impairments. (17-19) Impairment measures such as those of postural aberrations, decreased muscle force, and range of motion may not be good indicators of musculoskeletal dysfunction (20) and disability. (21,22)

The need for developing appropriate tools for measuring mobility and activities of daily living was recently characterized as a priority for research by an international task force on back pain. (23) Timed tests of activities such as walking, the sit-to-stand task, and repeated trunk flexion have been examined in patients with back pain and shown to have what we consider acceptable reliability, to be able to discriminate between people with and without back problems, and to be sensitive to change over time. (24-26) The ability to discriminate between people with and without back pain also has been demonstrated in tests of lifting capacity. (27,28) Decreased mobility of the trunk often is considered a manifestation of back pain, (2,29,30) and activities such as bending, twisting, stooping, crouching, lifting, dressing, and picking up objects often are limited. (2,31,32) We believe that performance tests used in patients with low back pain should be useful to elucidate such key aspects of functioning. To relate to daily tasks, we believe we probably need a variety of tests.

Five physical performance tests of compound activities were part of a test battery in a randomized controlled study of patients with long-lasting musculoskeletal pain. (33) The tests were presumed by the authors to be useful to assess activity limitations in patients with back pain because all the tests required mobility of the trunk. They were the Sock Test, (34) the Pick-up Tests, (35) the Roll-up Test, (36) the Fingertip-to-Floor Test, (37) and the Lift Test. (38) Discriminate validity and responsiveness to important change are known for each of the tests. (4)

A simple physical performance test, in our view, may not be sufficient to characterize physical functioning for all individuals with back problems. Although some activities cause pain in some people, they relieve pain in others. (39) People can be considered, we believe, more disabled if performance is limited in several activities rather than in one activity. This viewpoint is consistent with the findings of Waddell et al (22) and Thomas et al, (40) who showed that the sum of the scores of several physical tests could be better to discriminate between people with and without back pain than the use of separate tests. Provided that the tests measure a common construct of physical performance, the sum of the scores of several tests might offer a more comprehensive measure of trunk mobility than a score of a single test. The purpose of our study was to develop a condition-specific, simple, feasible, and clinically useful outcome measure of physical performance in back pain.

Our hypotheses were:

1. The Sock Test, the Pick-up Test, the Roll-up Test, the Fingertip-to-Floor Test, and the Lift Test all contribute to the assessment of a common dimension of physical performance in patients with back pain.

2. The sum of test scores is higher (worse) in patients with back pain than in patients with other musculoskeletal pain.

3. The sum of test scores discriminates between patients with different levels of functioning (working versus not working).

4. The sum of test scores is responsive to change over time and is more responsive than scores of the single tests.

Method

Subjects

The patients participated in a randomized controlled study conducted on patients on long-term sick leave (>2 months, [less than or equal to] 1 year) because of musculoskeletal conditions ([bar]X=3.3 months, SD=2.0). (33) Exclusion criteria were pregnancy, substance abuse, and illnesses such as progressive nervous system disease, serious cardiac disease, and acute infection. The patients must also have sufficient knowledge of the Norwegian language to understand instructions and questionnaires. The patients' conditions were diagnosed by their general practitioner according to the Norwegian translation of the International Classification of Primary Care (ICPC). (41) The study was accepted by the Regional Ethics Committee, was approved by the National Data Inspectorate, and was performed according to the Helsinki Declaration. All patients were to be followed for several years using data regarding sickness benefits from the National Health Insurance Register.

A total of 1,776 patients were invited to participate, and 1,683 patients met the inclusion criteria. Only 1,071 patients responded, and 573 patients (53.5%) gave written consent. The patients were randomly assigned to 2 groups to receive different interventions. Because the comparison of treatment effectiveness was not the purpose of our study, the specifics of the interventions are not relevant in this report. Two hundred eighty-eight patients with different musculoskeletal conditions were examined using all 5 physical performance tests. They had a diagnosis of back pain with or without radiating pain (n=157), neck or shoulder pain (n=93), or unspecified musculoskeletal pain (n=38).

Of the total group of patients with musculoskeletal pain (n=573), 249 patients (47% men) had back pain. Only 157 patients were examined using all 5 physical performance tests when the study began. Thirty-six of the patients did not attend 1-year follow-up assessments and were considered dropouts, and no data on work status were available for an additional 7 patients at follow-up. Accordingly, responsiveness to change and discriminative ability of the sum of test scores examined at the 1-year follow-up were assessed in 114 patients with back pain. Baseline characteristics and comparisons of demographic and test variables of patients with back pain who were included (n=114) and were not included (n=135) in those analyses are presented in Table 1.

Assessment of Physical Performance

Physical performance tests. Performances on the Sock Test, the Pick-up Test, the Roll-up Test, the Fingertip-to-Floor Test, and the Lift Test were assessed at the beginning of the study and at the 1-year follow-up by the same physical therapist. All tests require mobility of the trunk in compound actions, primarily in the sagittal plane. The tests and scoring alternatives are described in Table 2. The Sock Test and the Pick-up Test were developed in connection with the study.

Sock Test. The patient simulates putting on a sock in a standardized way from a sitting position. Performance is assessed using a 4-point ordinal scale. Intertester reliability in 21 patients with musculoskeletal pain between 2 testers was satisfactory ([K.sub.w]=.79), and aspects of validity were indicated. (34)

Pick-up Test. The patient picks up a piece of paper from the floor in an optional way. Performance is assessed using a 4-point ordinal scale. Intertester reliability in 24 patients with musculoskeletal pain between 2 testers was satisfactory (K=.74), and aspects of validity were indicated. (35)

Roll-up Test. The test is one of a large battery of tests developed by Sundsvold et al. (36) The patient rolls up slowly, with arms relaxed, from a supine position to an extended sitting position. Performance is assessed on an 8-point ordinal scale. Intertester reliability in 21 patients with musculoskeletal pain between 2 testers was moderate (K=.59). (4)

Fingertip-to-Floor Test. The patient stands on the floor, feet 10 cm apart, and is asked to bend forward with straight knees and try to touch the floor with the fingertips. The distance between the tip of the middle finger and the floor is recorded in centimeters. Intraclass correlation coefficients (1,1) for intertester and test-retest reliability in 73 patients with low back pain of a modified fingertip-to-floor measure have been reported as .95 and .98, respectively. (42)

Lift Test. The patient is asked to repeat lifting a box containing a sandbag of 5 kg from the floor to a table and back to the floor for 1 minute. (38) The number of lifts is recorded. Reliability of counting lifts per minute has not been reported.

The Back Performance Scale (BPS). The basis for constructing the BPS sum scale was examined. The methods of score assignment for the Sock Test and the Pick-up Test are shown in Table 2. A score of 0 was considered a good performance with no signs of activity limitation, a score of 1 was considered a somewhat limited performance, a score of 2 was considered a rather distinct limitation of performance, and a score of 3 was considered a substantially limited performance, if performed at all. These criteria were used to rescale the other 3 tests (Tab. 2), and cutoff points of the separate tests were decided by (1) distribution of scores at the beginning of the study, (2) clinical experience concerning physical performance in people with back pain and those without back pain. The BPS sum score was calculated by adding the individual scores of the 5 tests. Ordinal test scores obtained at the beginning of the study of 157 patients with back pain are given in Table 3. The distribution of BPS sum scores at the beginning of the study is shown in Figure 1.

[FIGURE 1 OMITTED]

Data Analysis

Demographic and test data were examined by descriptive analysis and compared by chi-square tests or Student t tests of independent groups. Normality of BPS sum scores was examined by inspecting normality plots and by the Kolmogorov-Smirnov test, with P>.05 indicating normality.

Scale construction. We believed that the 5 performance tests share a theoretical construct of physical performance: sagittal mobility of the trunk in functional activities. Spearman correlations were calculated to examine the bivariate relationships among the tests, and between each test and the BPS sum score, assessed in patients with back pain at the beginning of the study (n=157). Internal consistency was examined by use of the Cronbach alpha, an overall correlation among the items. Streiner and Norman (43) suggested that an alpha should be above .70 to demonstrate sufficient homogeneity of the items but should not be higher than .90, as it may suggest a high level of item redundancy. A high alpha may imply that some of the items are unnecessary and that the scale as a whole is too narrow in its scope to have much validity. (43)

Discriminative ability. Because the tests of the BPS all require mobility of the trunk, we expected patients with back pain to demonstrate higher BPS scores (more limited performance) than either patients with neck or shoulder pain or patients with unspecified musculoskeletal pain. Because the distributions of BPS scores were not normal in the patients with neck or shoulder pain and in the patients with unspecified musculoskeletal pain (P<.05), we chose to analyze differences between patients with back pain and each of those 2 groups using the nonparametric Mann-Whitney test for independent groups. The level for rejecting the null hypothesis of no difference was P<.05.

Patients with a history of back pain who still received workers' compensation 1 year after rehabilitation were expected to demonstrate higher BPS scores (more limited performance) than those who had returned fully to work at that time (by chance the same number of patients [n=57] was in both groups). Because the distributions of BPS scores in both groups were normal (P>.20), an independent t test was used to determine whether there was a difference in BPS sum scores between the 2 groups. The level for rejecting the null hypothesis of no difference was P<.05.

Responsiveness. Responsiveness refers to the power of a scale to detect meaningful change when it occurs. (44) A construct for examining responsiveness was suggested by Stratford et al: "Those patients judged by an external standard as having achieved an important change will demonstrate greater improvement [by the measure] than those judged by the standard as not achieving an important change." (45(p1112)) There is no gold standard for important or meaningful change. Because the patients in our study were all on long-term sick leave at the beginning of the study, having returned fully to work at the 1-year follow-up was considered an important change. In order to indicate that the BPS is responsive to important change, those patients should have improved more on the BPS than the patients who had not returned to work, defined as those who still received workers' compensation.

Effect size and receiver operating characteristic (ROC) curve statistics can be used to indicate responsiveness of assessment tools, and both were used to examine responsiveness of the BPS. Effect size is defined as the mean change found in a variable divided by the standard deviation of that variable. (46) The mean change of BPS sum scores from the beginning of the study to the 1-year follow-up was divided by the standard deviation of the change and examined in improved and unimproved patients as defined by the dichotomous return-to-work variable. Specific benchmarks of effect size were used for interpreting the magnitude of change. Cohen (47) suggested an effect size of .20 to be small, .50 moderate, and .80 to be large. Whether such benchmarks of change are appropriate for the magnitude of change in the present study may be debated. However, a higher effect size was expected in patients who had returned to work after 1 year than in those who had not.

Responsiveness of the BPS also was examined by ROC curve statistics. Deyo and Centor (48) suggested that measures may be viewed as diagnostic tests for discriminating between patients who improved and those who did not improve and, accordingly, can be described in terms of sensitivity and specificity in detecting improvement (yes/no) as established by another criteria of important change. Sensitivity against 1--specificity was plotted for each of several possible cutoff points in change of BPS sum scores from the beginning of the study to the 1-year follow-up. Sensitivity, in our study, was defined as the number of patients correctly identified by the test as having returned to work divided by all patients who had returned to work after 1 year. We defined specificity as the number of patients correctly identified by the test as not having returned to work divided by all patients who had not returned to work. Equal importance of high sensitivity and specificity was assumed. (49) The area under the ROC curve is interpreted as the probability of correctly identifying the improved patients from randomly selected pairs of improved and unimproved patients. (48) The area ranges from 0.5 (no accuracy in discriminating improved from unimproved) to 1.0 (perfect accuracy). The areas under the ROC curves were used to compare responsiveness of the BPS with those of the separate 5 performance tests.

Results

The proportion of men and women was different (P=.047) in the group of 114 patients with back pain (40% men) who were included to assess discriminative and evaluative properties of the BPS and in the remaining group of 135 patients with back pain (53% men). No other demographic or test variable was different (Tab. 1).

Scale Construction

Test scores obtained when the study began for the 157 patients with back pain are shown in Table 3. The test scores for one patient with back problems are illustrated in Figure 2. Bivariate correlations ([r.sub.s]) ranged from .27 to .50 among the 5 tests and from .63 to .73 between the separate tests and the BPS (P<.01) (Tab. 4). Internal consistency of the BPS was .73 (coefficient alpha). The BPS sum scores at the beginning of the study were normally distributed, as illustrated in Figure 1. One patient had a score of 0 (minimum), and one patient had a score of 15 (maximum).

[FIGURE 2 OMITTED]

Discriminative Ability

Higher BPS sum scores were demonstrated in patients with back pain (median=8, interquartile range [IQR]=5) than in patients with neck or shoulder pain (median=4, IQR=4) (P<.001). Higher BPS scores also were demonstrated in patients with back pain than in patients with unspecified musculoskeletal pain (median=4, IQR=5) (P<.001). Age and body mass index (BMI) did not differ between patients with back pain and the other patients. The percentage of women was higher in the group with neck or shoulder pain (71%) and in the group with unspecified musculoskeletal pain (78%) than in the group with back pain (57%). No difference in BPS sum scores was found between men and women.

Patients with a history of back pain who still received workers' compensation 1 year after rehabilitation demonstrated higher BPS sum scores ([bar]X=6.3, SD=3.7, range=0-14) than patients who had returned fully to work at that time ([bar]X=3.7, SD=3.2, range=0-14) (P<.001).

Responsiveness

The effect size of change between pretest and posttest measurements assessed by the BPS was 1.33 in patients who had changed and 0.31 in patients who had not changed, as defined by the dichotomous return-to-work variable (Tab. 5). High sensitivity (67%) and high specificity (70%) were obtained with a cutoff point of improvement of 2.5 on the BPS. Approximately two thirds of the patients were correctly classified as having returned to work if they improved at least this much on the BPS, and approximately two thirds of the patients were correctly classified as not having returned to work if they had less improvement. The area under the ROC curve (Fig. 3) was larger (0.77) in the BPS than in each of the separate performance tests (Tab. 6). When transforming the Roll-up Test from an 8-point ordinal scale to a 4-point ordinal scale, the area under the ROC curve did not change. However, the areas under the ROC curve became smaller when the Fingertip-to-Floor Test and the Lift Test were transformed from ratio scales to 4-point ordinal scales (Tab. 6), implying less responsiveness to change.

[FIGURE 3 OMITTED]

Discussion

Scale Construction

The 5 tests comprising the BPS demonstrated associations with each other, and each test contributed to high internal consistency, implying that the tests share a common characteristic in measuring physical performance. Because trunk mobility in forward flexion is required when performing all of the movements required by the tests, this characteristic may represent the shared physical dimension. Patrick et afro recommended that sufficient items examining the same concept should be included in a scale to obtain an acceptable high coefficient alpha. From a practical point of view, however, the scale should not include more test items than necessary to keep the testing simple and feasible. High internal consistency was demonstrated, although the BPS contains only 5 test items. Thus, the BPS may seem to represent a favorable scale construction.

"Back pain" is an ambiguous term because this condition includes different syndromes. (51) Physical performance may reflect not only the nature and severity of the underlying pathology, but also how the patient interprets and reacts to pain. (39) The tests were not more than moderately correlated, and we believe they could be used to assess somewhat different aspects of performance. For an individual patient, some activities may be very difficult to perform, whereas other activities are not difficult to perform (Fig. 2). The lowest and highest scores were frequently achieved by our subjects on the separate performance tests, indicating that ceiling and floor effects were present, which means that a test cannot register greater gains or greater decline respectively. (52) The problem of ceiling and floor effects was almost eliminated when the BPS was applied on groups of patients with long-lasting back pain.

Usually, when a new self-report instrument is developed to assess a dimension of health, the first task is to generate a pool of all potentially relevant items. Based on clinical judgment and statistical procedures, items are then selected from this pool for inclusion in the final instrument. (53) In our study, the choice of test items was post hoc, restricted to those initially included in the test armament of the Bergen Study: Back to Work (33) and further restricted to tests apparently designed to measure trunk mobility in functional tasks. It is an issue for further research to consider other tests, also involving side bending and twisting, for supplementing or replacing tests of the BPS.

Loss of physical abilities due to back pain may be attributable to physiological changes in motor patterns and reduced overall performance, rather than to isolated spinal impairments. The 5 activities tested in the BPS are performed from different starting positions, such as standing (Pick-up Test, Fingertip-to-Floor Test, Lift Test), sitting (Sock Test), and supine (Roll-up Test), causing gravity to act differently on the body, and some activities are performed with straight knees, whereas other activities are performed with flexed knees. Each test seems to us to measure the behavioural outcome that results from the interaction of the individual, the task, and the environment, which according to Gentile (54) is an analysis at the activity level. Because behavioral outcomes related to daily tasks are reflected in the BPS, we believe the BPS has the potential of being a performance measure with clinical usefulness. The BPS was intended to be a practical measure of activity limits, and we believe it is quick to perform and simple to assess.

Discriminative Ability

Discriminative ability of the BPS was demonstrated. The BPS sum scores were higher (worse) in patients with back pain than in patients with neck or shoulder pain and those with generalized pain. Activity limitation was particularly demonstrated by the high scores of the patients with back pain, indicating that the BPS is a condition-specific measure for people with back pain. All of the tests involve coordinated actions of the trunk in relation to the lower extremities. A correlation of r=.60 has been demonstrated between the fingertip-to-floor distance and flexibility of the hamstring muscles. (55) Therefore, it remains to be seen whether the BPS may also be useful to reflect activity limitation in patients with lower extremity dysfunction.

Performance scores also were higher (worse) in patients who had not returned fully to work 1 year after rehabilitation than in those who had returned fully to work. Discriminative ability is commonly examined by contrasting test scores of patients and those of subjects without the condition being studied. In our study, all participants had a history of long-lasting back pain. An even larger difference in BPS scores might have been obtained if test scores of patients with back problems had been contrasted to those of people with no recent sick leave because of back problems.

Responsiveness

Responsiveness of the BPS to change was demonstrated by 2 commonly used methods? Effect size statistics relate the magnitude of change to the variability in scores, but controversy exists as to whether variability of baseline scores or change in scores should be used in the equation. (47,56,57) Because baseline and follow-up measurements are not independent, we used standard deviation of the change, in accordance with the suggestion of Cohen. (47) Effect size was very high in patients who had improved, as defined by the external dichotomous return-to-work variable, but it was low in those who had not improved. This finding indicated to us that the BPS is responsive to change at the participation level of functioning. Test-retest reliability of BPS scores has not been examined, but the effect sizes in the 2 groups indicate to us that this aspect of reliability is sufficient.

Whether or not the return-to-work variable is an optimal external indicator to examine responsiveness of the BPS is arguable. Physical performance is only one of many factors (socioeconomic, psychological, demographic, job characteristics) that influence whether or not patients return to work after long-term sick leave. (58-60) A patient's global impression of change (49) may be shown to be an alternative external indicator of change in the future. Research, we believe, also should determine a numerical value of change by the BPS that is meaningful to monitor perceived change in an individual patient.

Responsiveness of the BPS to change also was demonstrated by ROC curve analysis. Change in BPS sum scores from the beginning of the study to the 1-year follow-up discriminated between patients who improved and those who did not improve, as defined by the dichotomous return-to-work variable. A correct classification of approximately two thirds of the patients at the particular BPS cutoff point of change (2.5) may seem satisfactory. The area under the ROC curve was greater in the BPS scores than in the scores of the separate tests, with the Sock Test demonstrating the greatest area among them (Tab. 6). Although the 4-point ordinal scale of the Sock Test is a gross measure, it apparently signifies clinically important steps in a patient's rehabilitation. Responsiveness became less when the Fingertip-to-Floor Test and the Lift Test were transformed from ratio to 4-point ordinal scales. These tests, we contend, should be used with their ordinal scales to better measure change in the separate activities, but we believe they should be transformed to 4-point ordinal scales when used as part of the BPS. The 0 to 15 ordinal sum score of the BPS provided the most responsive measure, more responsive than the single tests, irrespective of ordinal or ratio scales.

Conclusion

The BPS, including 5 physical performance tests of daily activities, appears to be a useful instrument for reflecting key aspects of performance in patients with long-lasting back problems. Internal consistency of the BPS was high, and discriminative ability of the instrument and responsiveness to change were demonstrated. The BPS was shown to be more responsive to change than each of the 5 tests separately. As performance of the 5 tests primarily requires mobility of the trunk in the sagittal plane, we believe future research should examine whether tests using side bending and twisting also should be included or could replace other tests to have an even better measure of mobility-related activities in people with back pain. The BPS is a practical measure of performance, being easy and quick to perform, with no need for costly equipment. Reliability of the BPS sum score needs to be established.

Table 1.
Characteristics of Patients With Back Pain at Beginning of Study:
Comparisons Between Patients Included and Excluded From Main
Analysis of Measurement Properties of the Back Performance Scale

                                            Included      Excluded
Variables                                   (n=114)       (n=135)

Demographic variables
  Age (y): [bar]X (SD)                      43.9 (10.6)   43.0 (10.4)
  Body mass index (kg/[m.sup.2]):
    [bar]X (SD)                             25.5 (4.0)    24.8 (3.8)
  Sex (%): men, women                       40, 60        53, 47
  Civil status (%): married or
    cohabiting, single                      77, 23        74, 26
  Education (%): primary school,
    technical college, high school or
    university                              35, 25, 40    42, 28, 29
  Occupation (%): blue-collar, white-
    collar, unspecified                     27, 38, 35    42, 29, 29
  Smoking habits (%): smoking, nonsmoking   50, 50        58, 42
  Duration of complaints (y): [bar]X (SD)   10 (9.0)      11 (9.2)
  Previously on sick leave because of
    back pain (%): yes, no                  77, 23        79, 21

Disability
  Disability Rating Index (100 mm):
    [bar]X (SD)                             56.2 (13.3)   55.0 (14.0)

Pain
  Norwegian pain questionnaire (%):
    [bar]X (SD)                             40.4 (22.2)   42.9 (23.4)
  Visual analog scale (100 mm):
    [bar]X (SD)                             52.6 (20.5)   53.4 (19.0)

                                            Tests for
                                            Difference (a)

Variables                                   t, [chi square]   P

Demographic variables
  Age (y): [bar]X (SD)                      0.54 (1)          .591
  Body mass index (kg/[m.sup.2]):
    [bar]X (SD)                             1.46 (1)          .145
  Sex (%): men, women                       3.95 (2)          .047
  Civil status (%): married or
    cohabiting, single                      0.25 (2)          .614
  Education (%): primary school,
    technical college, high school or
    university                              3.03 (2)          .220
  Occupation (%): blue-collar, white-
    collar, unspecified                     5.60 (2)          .061
  Smoking habits (%): smoking, nonsmoking   1.47 (2)          .226
  Duration of complaints (y): [bar]X (SD)   0.33 (1)          .735
  Previously on sick leave because of
    back pain (%): yes, no                  0.24 (2)          .624

Disability
  Disability Rating Index (100 mm):
    [bar]X (SD)                             0.61 (1)          .544

Pain
  Norwegian pain questionnaire (%):
    [bar]X (SD)                             0.65 (1)          .515
  Visual analog scale (100 mm):
    [bar]X (SD)                             0.30 (1)          .765

(a) Between-group differences examined by: (1) Student t test (P<.05),
(2) chi-square ([chi square]) test (P<.05)

Table 2.
Back Performance Scale (BPS) With Operational Definitions of Ordinal
Scores

Tests           Task/Movement

Sock Test       From a sitting position: grab the toes with fingertips,
                  the leg flexed in the sagittal plane, one leg tested
                  at the time, scoring the least reach

Pick-up Test    From a standing position: pick up a piece of paper
                  from the floor, 2 or 3 times in varied ways

Roll-up Test    From a supine position: roll up slowly into a long-
                  sitting position, arms relaxed

Fingertip-to-   From a standing position: feet 10 cm apart and
  Floor Test      knees straight, reach toward the floor with
                  fingertips

Lift Test       From a standing position: repeat lifting a box (1.35
                  kg, sized 0.36x0.36x0.25 cm) containing a
                  sandbag of 5 kg, for 1 min, from the floor to the
                  table (height=76 cm), technique optional

                BPS sum scores

Tests           Scores   Categories of Performance

Sock Test       0        Can easily grab the toes with fingertips of
                           both hands
                1        Can grab the toes with fingertips, but with
                           effort
                2        Can reach beyond the malleoli, but not reach
                           the toes
                3        Can hardly, if at all, reach as far as to
                           the malleoli

Pick-up Test    0        Can do the task with ease in varied ways
                1        Can do the task with minor effort or some
                2          decreased flexibility
                3        Can do the task with marked effort or lack of
                           flexibility, may need support of hand on
                           thigh
                         Cannot perform the task at all or need
                           external support

Roll-up Test    0        Can roll up with ease to long-sitting position
                1        Can roll up with marked effort or partially
                2          to long-sitting position
                3        Can roll up in supine position between the 8th
                           and 12th thoracic vertebrae
                         Can roll up in supine position above the 8th
                           thoracic vertebra

Fingertip-to-   0        Can reach to the floor, distance=0 cm
  Floor Test    1        Can reach to a distance >0 cm, [less than or
                           equal to] 20 cm
                2        Can reach to a distance >20 cm, [less than or
                           equal to] 40 cm
                3        Can reach to a distance >40 cm

Lift Test       0        Can do the lifting task >15 times
                1        Can do the lifting task >10, [less than or
                           equal to] 15 times
                2        Can do the lifting task >0, [less than or
                           equal to] 10 times
                3        Cannot/will not do the lifting task=0

                0-15

Table 3.
Ordinal Test Data at Beginning of Study in Patients With Back Pain
(n = 157)

Tests                     n (%)

Sock Test
  0                       31 (20)
  1                       57 (36)
  2                       44 (28)
  3                       25 (16)
Pick-up Test
  0                       25 (16)
  1                       51 (33)
  2                       76 (48)
  3                        5 (3)
Roll-up Test
  0                       24 (15)
  1                       38 (24)
  2                       41 (26)
  3                       54 (34)
Fingertip-to-Floor Test
  0                       28 (18)
  1                       47 (30)
  2                       51 (33)
  3                       31 (20)
Lift Test
  0                       25 (16)
  1                       49 (31)
  2                       54 (34)
  3                       29 (19)

Table 4.
Bivariate Correlation Matrix (a) of Test Scores at Beginning of Study
in Patients With Back Pain (n=157)

                          Sock     Pick-up   Roll-up
Tests                     Test     Test      Test

Sock Test                 1.00
Pick-up Test               .50 *   1.00
Roll-up Test               .37 *    .34 *    1.00
Fingertip-to-Floor Test    .41 *    .44 *     .30 *
Lift Test                  .27 *    .34 *     .33 *
BPS (b)                    .73 *    .70 *     .69 *

                          Fingertip-to-   Lift
Tests                     Floor Test      Test

Sock Test
Pick-up Test
Roll-up Test
Fingertip-to-Floor Test   1.00
Lift Test                  .27 *          1.00
BPS (b)                    .70 *           .63 *

(a) Spearman rank correlation ([r.sub.s]), * P<.01.

(b) Back Performance Scale (BPS)=sum of 5 test scores.

Table 5.
Effect Size of the Back Performance Scale Sum Scores in Improved and
Not Improved Patients

                        Baseline        Follow-up         Change

Patient
Groups         n    [bar]X     SD    [bar]X     SD    [bar]X     SD

Improved       57   7.7        3.4   3.7        3.3   4.0        3.0
Not improved   57   7.1        3.1   6.3        3.7   0.9        2.9

Patient        Effect
Groups         Size (a)

Improved       1.33
Not improved   0.31

(a) Calculated as the mean change divided by the standard deviation
of the change.

Table 6.
Areas Under the Receiver Operating Characteristic (ROC) Curves for
Each Instrument (n= 114)

                          ROI: Area
                          (95% Confidence Interval)

Instruments               Ordinal Scale       Ratio Scale

Back Performance Scale    0.77 (0.68, 0.85)
Sock Test                 0.72 (0.63, 0.81)
Pick-up Test              0.65 (0.55, 0.75)
Roll-up Test              0.69 (0.60, 0.79)
Fingertip-to-Floor Test   0.65 (0.55, 0.75)   0.71 (0.61, 0.81)
Lift Test                 0.63 (0.51, 0.76)   0.69 (0.59, 0.79)

References

(1) McColl E, Christiansen T, Konig-Zahn C. Making the right choice of outcome measure. In: Hutchinson A, Bentzen N, Konig-Zahn C, eds. Cross Cultural Health Outcome Assessment: A User's Guide. European Research Group on Health Outcomes; 1996:12-26.

(2) Harper AC, Harper DA, Lambert LJ, et al. Symptoms of impairment, disability, and handicap in low back pain: a taxonomy. Pain. 1992;50: 189-195.

(3) Jette DU, Jette AM. Physical therapy and health outcomes in patients with spinal impairments [published erratum appears in Phys Ther. 1997;77:113]. Phys Ther. 1996;76:930-941.

(4) Strand LI, Ljunggren AE, Haldorsen EM, Espehaug B. The impact of physical function and pain on work status at 1-year follow-up in patients with back pain. Spine. 2001;26:800-808.

(5) Fairbank JC, Couper J, Davies JB, O'Brien JP. The Oswestry low back pain disability questionnaire. Physiotherapy. 1980;66:271-273.

(6) Roland M, Morris R. A study of the natural history of low-back pain, part II: development of guidelines for trials of treatment in primary care. Spine. 1983;8:145-150.

(7) Salen BA, Spangfort EV, Nygren AL, Nordemar R. The Disability Rating Index: an instrument for the assessment of disability in clinical settings. J Clin Epidemiol. 1994;47:1423-1435.

(8) Stratford PW, Binkley J, Riddle DL. Development and initial validation of the back pain functional scale. Spine. 2000;25:2095-2102.

(9) Fordyce WE. On the nature of illness and disability [editorial]. Clin Orthop. March 1997(336):47-51.

(10) Jette AM. The Functional Status Index: reliability and validity of a self-report functional disability measure. J Rheumatol. 1987;14(suppl 15):15-21.

(11) Ekdahl C, Englund A, Stenstrom CH. Development and evaluation of the index of muscle function. Advances in Physiotherapy. 1999;1: 45-55.

(12) Gronblad M, Hurri H, Kouri JP. Relationships between spinal mobility, physical performance tests, pain intensity, and disability assessments in chronic low back pain patients. Scand J Rehabil Med. 1997;29:17-24.

(13) Michel A, Kohlmann T, Raspe H. The association between clinical findings on physical examination and self-reported severity in back pain: results of a population-based study. Spine. 1997;22:296-303.

(14) Holmstrom E, Moritz U. Low back pain: correspondence between questionnaire, interview and clinical examination. Scand J Rehabil Med. 1991;23:119-125.

(15) Wijlhuizen GJ, Ooijendijk W. Measuring disability: the agreement between self-evaluation and observation of performance. Disabil Rehabil. 1999;21:61-67.

(16) International Classification of Functioning, Disability and Health: ICIDH-2. Geneva, Switzerland: World Health Organization; 2001.

(17) Nelson RM, Nestor DE. Atlas of standardized low-back tests and measures of the National Institute for Occupational Safety and Health. Scand J Work Environ Health. 1988;14(suppl 1):82-84.

(18) Guides to the Evaluation of Permanent Impairment. 3rd ed. Chicago, Ill: American Medical Association; 1990.

(19) Moffroid MT, Haugh LD, Henry SM, Short B. Distinguishable groups of musculoskeletal low back pain patients and asymptomatic control subjects based on physical measures of the NIOSH Low Back Atlas [published erratum appears in Spine. 1994;19:2137]. Spine. 1994; 19:1350-1358.

(20) Mooney V. Impairment, disability, and handicap. Clin Orthop. August 1987(221):14-25.

(21) Deyo RA. Measuring the functional status of patients with low back pain. Arch Phys Med Rehabil. 1988;69:1044-1053.

(22) Waddell G, Somerville D, Henderson I, Newton M. Objective clinical evaluation of physical impairment in chronic low back pain. Spine. 1992;17:617-628.

(23) Abenhaim L, Rossignol M, Valat JP et al. The role of activity in the therapeutic management of back pain: Report of the International Paris Task Force on Back Pain. Spine. 2000;25(suppl 4):1S-33S.

(24) Harding VR, Williams AC, Richardson PH, et al. The development of a battery of measures for assessing physical functioning of chronic pain patients. Pain. 1994;58:367-375.

(25) Simmonds MJ, Olson SL, Jones S, et al. Psychometric characteristics and clinical usefulness of physical performance tests in patients with low back pain. Spine. 1998;23:2412-2421.

(26) Simmonds MJ. Physical Function and Physical Performance in Patients With Pain: What Are the Measures and What Do They Mean? Seattle, Wash: IASP Scientific Program Committee, IASP Press; 1999.

(27) Ljungquist T, Harms-Ringdahl K, Nygren A, Jensen I. Intra- and inter-rater reliability of an 11-test package for assessing dysfunction due to back or neck pain. Physiother Res Int. 1999;4:214-232.

(28) Ljungquist T, Fransson B, Harms-Ringdahl K, et al. A physiotherapy test package for assessing back and neck dysfunction: discriminative ability for patients versus healthy control subjects. Physiother Res Int. 1999;4:123-140.

(29) Mayer T, Tabor J, Bovasso E, Gatchel R. Physical progress and residual impairment quantification after functional restoration, part I: lumbar mobility. Spine. 1994;19:389-394.

(30) Rainville J, Sobel JB, Hartigan C. Comparison of total lumbosacral flexion and true lumbar flexion measured by a dual inclinometer technique. Spine. 1994;19:2698-2701.

(31) Linton SJ. The relationship between activity and chronic back pain. Pain. 1985;21:289-294.

(32) Kohlmann T, Raspe H. Die patientennahe Diagnostik von Functionseinschrankungen im Alltag. Psychomed. 1994;6:21-27.

(33) Haldorsen EM, Kronholm K, Skouen JS, Ursin H. Multimodal cognitive behavioral treatment of patients sicklisted for musculoskeletal pain: a randomized controlled study. Scand J Rheumatol. 1998;27: 16-25.

(34) Strand LI, Wie SL. The Sock Test for evaluating activity limitation in patients with musculoskeletal pain. Phys Ther. 1999;79:136-145.

(35) Strand LI, Ljunggren AE. The Pick-up Test for assessing performance of a daily activity in patients with back pain. Advances in Physiotherapy. 2001;3:17-27.

(36) Sundsvold MO, Vaglum P, Denstad K, eds. Comprehensive Physiotherapy Muscle Study: For Use in Clinical Work and Research. Oslo, Norway: Ulleval Hospital; 1982.

(37) Frost M, Stuckey S, Smalley LA, Dorman G. Reliability of measuring trunk motions in centimeters. Phys Ther. 1982;62:1431-1437.

(38) Wie SL. An Evaluation of Lifting Function in Persons With Low-Back Pain: An Assessment and Evaluation of the Reliability and Validity of the Dynamic Lifting Test: "Kystatesten Lifting-Low." Bergen, Norway: University of Bergen; 1996.

(39) Waddell G. The Back Pain Revolution. Edinburgh, Scotland: Churchill Livingstone; 1998.

(40) Thomas E, Silman AJ, Papageorgiou AC, et al. Association between measures of spinal mobility and low back pain: an analysis of new attenders in primary care. Spine. 1998;23:343-347.

(41) Norwegian College of General Practitioners. ICPC, ICHPPC-2-defined, IC-Process-PC. [Klassifikasjoner og definisjoner for primaerhelsetjenesten. WONCA's internasjonale klassifikasjoner og definisjoner tilrettelagt for Norge]. Otta, Norway: TANO AS og Norsk selskap for almennmedisin; 1991.

(42) Gauvin MG, Riddle DL, Rothstein JM. Reliability of clinical measurements of forward bending using the modified fingertip-to-floor method. Phys Ther. 1990;70:443-447.

(43) Streiner DL, Norman GR. Health Measurement Scales: A Practical Guide to their Development and Use. 2nd ed. Oxford, England: Oxford University Press; 1995.

(44) Kirshner B, Guyatt G. A methodological framework for assessing health indices. J Chronic Dis. 1985;38:27-36.

(45) Stratford PW, Binkley FM, Riddle DL. Health status measures: strategies and analytic methods for assessing change scores. Phys Ther. 1996;76:1109-1123.

(46) Kazis LE, Anderson JJ, Meenan RF. Effect sizes for interpreting changes in health status. Med Care. 1989;27(suppl 3):S178-S189.

(47) Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates Inc; 1988.

(48) Deyo RA, Centor RM. Assessing the responsiveness of functional scales to clinical change: an analogy to diagnostic test performance. J Chronic Dis. 1986;39:897-906.

(49) Farrar JT, Young JP, LaMoreaux L, et al. Clinical importance of changes in chronic pain intensity measured on an 11-point numerical pain rating scale. Pain. 2001;94:149-158.

(50) Patrick DL, Wild DJ, Johnson ES, et al. Cross-cultural validation of quality of life measures. In: Orley J, Kuyken W, eds. Quality of Life Assessment: International Perspectives. Berlin, Germany: Springer-Verlag; 1994:19-32.

(51) Fairbank JC, Hall H. History taking and physical examination: identification of syndromes of back pain. In: Wiesel SW, Weinstein JN, Herkowitz HN, et al, eds. The Lumbar Spine. Philadelphia, Pa: WB Saunders Co; 1996:85-104.

(52) Domholdt E. Physical Therapy Research: Principles and Applications. 2nd ed. Philadelphia, Pa: WB Saunders Co; 2000.

(53) Juniper EF, Guyatt G, Jaeschke R. How to develop and validate a new health-related quality of life instrument. In: Spilker B, ed. Quality of Life and Pharmacoeconomics in Clinical Trials. Philadelphia, Pa: Lippincott-Raven Publishers; 1996:49-56.

(54) Gentile AM. Skill acquisition: action, movement, and neuromotor processes. In: Carr J, Shepherd R, eds. Movement Science: Foundations for Physical Therapy in Rehabilitation. Gaithersburg, Md: Aspen Publisher; 2000:111-187.

(55) Biering-Sorensen F. Physical measurements as risk indicators for low-back trouble over a one-year period. Spine. 1984;9:106-119.

(56) Guyatt G, Walter S, Norman G. Measuring change over time: assessing the usefulness of evaluative instruments. J Chronic Dis. 1987; 40:171-178.

(57) Deyo RA, Diehr P, Patrick DL. Reproducibility and responsiveness of health status measures: statistics and strategies for evaluation. Control Clin Trials. 1991;12(suppl 4):142S-158S.

(58) Bendix AF, Bendix T, Haestrup C. Can it be predicted which patients with chronic low back pain should be offered tertiary rehabilitation in a functional restoration program? A search for demographic, socioeconomic, and physical predictors. Spine. 1998;23:1775-1783.

(59) Brage S. Musculoskeletal Health Problems and Sickness Absence: An Epidemiological Study of Concepts, Determinants, and Consequences. Oslo, Norway: Institute of General Practice and Community Medicine, University of Oslo; 1998.

(60) Haldorsen EMH, Wormgoor ME, Bjorholt PG, Ursin H. Predictors for outcome of a functional restoration program for low-back pain patients: a 12-month follow-up study. Eur J Phys Med Rehabil. 1998;8: 103-109.

LI Strand, PT, PhD, is Associate Professor, Section of Physiotherapy Science, Department of Public Health and Primary Health Care, Faculty of Medicine, University of Bergen, Ulriksdal 8c, Bergen, Norway ([e-mail]). Address all correspondence to Dr Strand.

R Moe-Nilssen, PT, PhD, is Associate Professor, Section of Physiotherapy Science, Department of Public Health and Primary Health Care, Faculty of Medicine, University of Bergen.

AE Ljunggren, PT, PhD, is Professor, Section of Physiotherapy Science, Department of Public Health and Primary Health Care, Faculty of Medicine, University of Bergen.

Dr Strand and Dr Moe-Nilssen provided concept/research design, writing, and data analysis. Dr Strand provided data collection. Dr Ljunggren provided project management and consultation (including review of manuscript for submission). Professor Holger Ursin is acknowledged for valuable advice. The authors thank the collaborating researchers in the Bergen Study: Back to Work: Alf Erling Risa, Astrid Grasdal, Ellen MH Haldorsen, Jan Sture Skouen, Jan Utkilen, Karsten Kronholm, May Vik, Petter Bomann-Larsen, and Solveig Lill Wie. They also thank Nina Konglevoll for data entry.

Data were derived from a rehabilitation study, The Bergen Study: Back to Work, which was funded by the Department of Health and Social Welfare and administered by the Municipality of Bergen. The physical test data were collected at the College of Physiotherapy in Bergen, Norway.

The Bergen Study: Back to Work was performed according to the Helsinki Declaration and approved by the Regional Ethics Committee, Health Region III, Norway, and the Norwegian Data Inspectorate.

This article was submitted July 20, 2001, and was accepted July 6, 2002.