Evaluation of the human activity profile for use with persons with arthritis

Authors


Abstract

Objective

To investigate the reliability, validity, and responsiveness to change of the Human Activity Profile (HAP), a questionnaire measuring physical activity, in persons with arthritis.

Methods

Twenty-eight subjects completed the following self-report questionnaires: HAP, Modified Health Assessment Questionnaire, Medical Outcomes Study 36-Item Short Form Health Survey, and Arthritis Impact Measurement Scale 2. Subjects also completed a submaximal treadmill test, Timed-Stands Test, and 50-Foot Walk Test. Responses on the HAP resulted in 2 scores: the maximum activity score (MAS) and the adjusted activity score (AAS). These scores were correlated with the other tests and examined for test-retest reliability. A subset of subjects participated in a 12-week exercise program, repeating the same tests when finished.

Results

For all subjects, the intraclass correlation coefficient (ICC) was 0.76 for the MAS and 0.87 for the AAS. Significant correlations were found between the HAP scores and the questionnaires, submaximal treadmill test, Timed-Stands Test, and 50-Foot Walk Test. In response to 12 weeks of exercise, both HAP scores had an effect size of 0.5, similar to that of the other questionnaires.

Conclusion

The ICCs demonstrate that the HAP is reliable, and the correlations between the HAP and other questionnaires, Timed-Stands Test, and 50-Foot Walk Test demonstrate that the HAP is a valid measure of physical function in persons with arthritis. The HAP's correlation with maximum oxygen consumption estimated by the treadmill test validates it as a measure of physical activity. The medium effect size (0.5) demonstrates that the HAP is moderately responsive to change. Its ease of use and broad scope make it a valuable assessment tool for persons with arthritis.

INTRODUCTION

In 2001, approximately 33% of adults in the United States reported that they were affected by arthritis or chronic joint symptoms (1). This prevalence is expected to increase as the population ages (2). Although arthritis is not a fatal disease, it impacts quality of life. Arthritis and other rheumatic diseases are the leading cause of disability for adults in the United States (3). During the 1996–1998 Behavioral Risk Factor Surveillance System conducted by the Centers for Disease Control and Prevention, persons with arthritis reported an average of 2.3 more days of activity limitation during the previous 30 days than persons without arthritis (4).

According to a 1990–1991 National Health Interview Survey, the prevalence of leisure-time physical activity among persons with arthritis and other rheumatic conditions was less than that of the general population (5), and may be due in part to the belief that individuals with arthritis will further damage their joints and exacerbate their symptoms by exercising. However, research has shown this is not the case. Recent reviews have found that Tai Chi (6) and moderate to high-intensity strength training (7) are tolerable and not detrimental for adults with rheumatoid arthritis (RA), and aerobic and strengthening exercise programs are safe and improve pain, function, and disability for adults with osteoarthritis (OA) (8, 9). Minor et al found that exercise can improve aerobic capacity, strength, and self-reported physical activity in persons with OA and RA without exacerbating their arthritic signs and symptoms (10). De Jong et al found that long-term, high-intensity exercise for persons with RA resulted in improvements in functional ability and did not increase radiographic damage in large joints (11), and that exercise may have a protective effect on the joints of the feet and hands (12).

A simple way to measure the impact of arthritis on a person's health is through self-report questionnaires. Self-report questionnaires provide a way to measure baseline function and detect changes over time or in response to intervention. Three questionnaires that have been validated for arthritis populations are the Modified Health Assessment Questionnaire (M-HAQ) (13), Arthritis Impact Measurement Scale 2 (AIMS2) (14), and Medical Outcomes Study 36-Item Short Form Health Survey (SF-36) (15). However, these questionnaires may have weaknesses in their ability to detect changes in physical function. The M-HAQ has been found to be less sensitive to change than other measures (16–18) and may have a ceiling effect (16). The SF-36 is a global measure of health with scales that evaluate physical, emotional, and social functioning; bodily pain; mental health; vitality; and general health perception (19). Likewise, the AIMS2 evaluates many areas in addition to physical function (14). Therefore, the SF-36 and AIMS2 collect data that are not needed for assessing physical function, and interpretations of a person's physical function based on the entirety of the SF-36 or AIMS2 must be guarded. Furthermore, researchers have noted the length of the SF-36 and AIMS2 to be burdensome to patients and clinicians (15, 18).

The Human Activity Profile (HAP), originally named the Additive Daily Activities Profile Test (ADAPT) Quality of Life Scale, is a self-report questionnaire that was designed to assess activity level in patients with chronic obstructive pulmonary disease (20). The HAP consists of 94 self-report items to which the subject responds in 1 of 3 ways: Still doing this activity,” “Have stopped doing this activity,” or “Never did this activity.” The HAP assesses a variety of activities ranging from activities of daily living and instrumental activities of daily living to recreational and sport activities. Each activity is based on estimated metabolic equivalents (METs), with each successive question representing a slightly higher MET level. The HAP has been found to correlate with the physical functioning scale of the SF-36 and physical performance measures in a population of adults undergoing hemodialysis (21), as well as measures of physical function in a population of adults with knee OA (22). These results suggest that the HAP can be used to measure physical function as well as activity, which was the original intent of its authors (20).

Several characteristics of the HAP make it well-suited for assessment of persons with arthritis in the clinic and in research trials. Researchers have found the HAP quick to complete (5–10 minutes) (22–24) and simple to score (22). Scores on the HAP correlate with maximum oxygen consumption (VO2max), making it useful to estimate a person's fitness level (20, 25), and normative data are available for comparison with individual scores (25). Furthermore, the HAP assesses a wide range of activities from low-level tasks, such as working at a desk, to high-level activities, such as running or jogging, which should minimize floor and ceiling effects (23).

We hypothesized that the HAP would be an effective method for measuring function and activity in all individuals with arthritis, including those with few impairments. If so, the HAP could be used to detect early declines in function, leading to earlier intervention and reduction of subsequent consequences of arthritis. It could also be used to detect subtle improvements in function to determine efficacy of treatment interventions.

The purposes of this study were to establish the test-retest reliability and validity of the HAP as a measure of physical function and activity for persons with RA and OA, and to determine its responsiveness to changes in physical function after participation in a structured exercise program.

SUBJECTS AND METHODS

Subjects.

Reliability/validity phase.

Twenty-eight adults (16 with OA and 12 with RA) participated in the reliability/validity phase of this study. Subjects were selected independent of race or sex. Subjects with OA were diagnosed via radiograph or arthroscopy. The age range for inclusion of subjects with OA was 40–69 years, and was chosen because many people in this age range have OA, yet are free of other comorbidities that could limit their activity and function. Subjects with RA were diagnosed according to the 1987 American College of Rheumatology (formerly the American Rheumatism Association) revised criteria (26). Subjects with RA were included if their age was within the range of 20 to 69 years (to distinguish our subjects from those with juvenile RA).

Subjects were recruited through physician contacts, flyers, and published advertisements. Interested subjects were screened via telephone to detect any major health problems other than arthritis that would impact their daily activity level or preclude them from performing the tests required. Based on this screening, patients with significant comorbidities and pregnant women were excluded from this study. To participate, subjects diagnosed with RA were to be stable in their disease process. Subjects who met the inclusion criteria obtained physician consent to confirm safety for participation in the submaximal graded exercise test.

Responsiveness-to-change phase.

Thirteen of the original 28 subjects (7 with OA and 6 with RA) volunteered to participate in the responsiveness-to-change phase of the study. This portion of the study required subjects to participate in 12 weeks of exercise training. People who were already participating in a regular exercise program were not included. Subjects in this phase obtained physician consent to participate.

This study was approved by the Institutional Review Board at the University of Nebraska Medical Center. Informed consent was obtained from all participants.

Procedure: reliability/validity phase.

Participation in the reliability/validity phase of the study involved one visit to the Clinical Movement Science Laboratory of the Division of Physical Therapy Education of the University of Nebraska Medical Center. The subjects were screened with the Physical Activity Readiness Questionnaire (27) and a health history questionnaire to identify any factors that would make participation in the study unsafe or confound the test results. To evaluate test-retest reliability of the HAP, subjects completed the questionnaire 2 times 5–14 days apart. To evaluate the validity of the HAP as a measure of physical function in persons with arthritis, the initial HAP scores were compared with other self-report questionnaires that measure physical function as well as with 2 tests of lower extremity function. To evaluate the validity of the HAP as a measure of physical activity in persons with arthritis, the HAP scores were compared with estimated VO2max. All self-report questionnaires and physical performance tests are detailed below and were completed once in a single visit, with the exception of the second copy of the HAP.

Human Activity Profile Test (25).

Each subject was given the standard instructions for the HAP, as written by its authors. A copy of the HAP was given to the subjects along with a stamped envelope addressed to the investigators. The subjects were asked to complete this questionnaire again in 5–14 days and return it by mail.

Two scores were obtained from the HAP: the maximum activity score (MAS) and the adjusted activity score (AAS). The MAS is the activity with the highest MET level the individual still performs. This is reflective of the subject's current maximum activity level. The AAS represents the respondent's average activity level on a daily basis. It is calculated by subtracting from the MAS the total number of activities below the MAS that the individual has stopped doing. Potential scores for the HAP range from 0–94 for both the MAS and the AAS, with higher scores indicating higher levels of function.

SF-36 (19, 28).

Each subject was given instructions to answer the questions on the SF-36 regarding their overall general health. They were cautioned not to focus solely on the effects of arthritis on their wellbeing. The SF-36 contains 8 multi-item scales. Scores were calculated for the Physical Functioning and Role Functioning-Physical scales, the 2 scales that relate to a person's function. Possible scores for each of the scales range from 0–100, with higher scores reflecting higher functional levels.

M-HAQ (13).

The M-HAQ evaluates an individual's level of difficulty in completing 8 daily activities. The subjects chose 1 of 4 responses to rate their level of difficulty for each activity: 0 (without difficulty), 1 (with some difficulty), 2 (with much difficulty), or 3 (unable to do). The overall score for an individual is the average of the patient's score for all items in the instrument. Potential scores range from 0 to 3, with lower scores indicating higher levels of function.

AIMS2 (14, 29).

The AIMS2 contains multiple questions that make up 12 scales to describe several aspects of health such as mobility, mood, pain, and social support. There are also items to assess satisfaction with health, overall impact of arthritis, and health perception. Various combinations of the 12 scales are averaged to form 5 components of health status, including physical, affect, symptom, social interaction, and role components. Scale and component scores range from 0 to 10, with lower scores representing good health status.

Single-Stage Submaximal Treadmill Test (30).

Each subject was allowed to become familiar with walking on a motor-driven treadmill, the Quinton Medtrack CR60 (Quinton Instrument Company, Bothell, WA), on both a level surface and a 5% grade. After this practice period, each subject rested until his or her heart rate approximated his baseline level. The test began with a 4-minute warmup at 0% grade at a self-selected walking speed. The treadmill grade was then increased to 5% and the subjects continued walking at the same speed for 4 additional minutes. Heart rate was assessed with the Polar monitor (Polar Electro, Port Washington, NY). Final heart rate and selected walking speed were recorded to estimate aerobic capacity by calculating VO2max with the following equation:

equation image

Timed-Stands Test (31).

This test of lower extremity function measured the time it took to stand up 10 times from a chair (43 cm deep and 44 cm high) without using the upper extremities for assistance. Each subject was instructed to perform this test as quickly as possible. Time to complete the task was recorded to an accuracy of one-hundredth of a second.

50-Foot Walk Test (10, 31).

A length of 50 feet was measured and marked off by tape. Subjects stood at one end of the 50-foot length and were instructed to walk this distance as fast as they comfortably could without an assistive device. Further instruction was given to the subjects to continue walking beyond the end of the 50 feet. Time from marker to marker was recorded to an accuracy of one-hundredth of a second.

Procedure: responsiveness-to-change phase.

To evaluate the HAP's responsiveness to changes in physical function, a subset of 13 subjects completed the questionnaire before and after 12 weeks of exercise training upon completion of the reliability/validity phase of the study. They also completed all of the other self-report questionnaires used in the reliability/validity phase of the study, so that the responsiveness to change of the HAP could be compared with these other questionnaires. To determine if the exercise intervention was of sufficient intensity to cause a change in physical status, subjects completed the same physical performance tests used in the reliability/validity phase of the study before and after 12 weeks of exercise training.

Exercise intervention.

The 12-week exercise intervention consisted of aerobic walking, resistive exercises, and flexibility exercises. The exercise program was individualized for each subject based on his estimated VO2max from the submaximal treadmill test and strength and flexibility impairments found during a physical examination. The aerobic walking portion included walking at a velocity that resulted in 60% of the subject's age-predicted maximal heart rate (220 − age) until a duration of 30 minutes was tolerated. The intensity was increased to 70–85% of the age-predicted maximal heart rate depending on the subject's tolerance during the course of the 12 weeks. The prescribed resistive and flexibility exercises were performed after the aerobic walking portion of each exercise session. One-on-one exercise sessions with participating trained personnel were held 3 days per week. Each session lasted approximately 1 hour.

Statistical analysis.

Reliability.

MAS and AAS scores were calculated for the HAP completed at the testing site (MAS1 and AAS1), as well as for the second copy completed within 5–14 days at the subject's home (MAS2 and AAS2). These were each analyzed for test-retest reliability using an intraclass correlation coefficient (ICC).

Validity.

A Spearman's rho correlation was used to establish the correlation between the MAS1 and each of the other assessment measures (self-report questionnaires and physical performance tests), as well as the correlation between the AAS1 and each of these same measures.

Responsiveness to change.

To evaluate responsiveness to change of the HAP, the mean change between the preexercise (average of MAS1 and MAS2 and average of AAS1 and AAS2) and postexercise intervention MAS and AAS was determined, and the effect size was calculated with the following equation:

equation image

The same calculation was made for each other questionnaire for comparison. To confirm that the 12-week exercise intervention resulted in a change in physical status, Wilcoxon's signed rank test was used to compare results of the physical performance tests pre- and postexercise intervention.

SigmaStat for Windows Version 2.0 software (Jandel Corporation, San Rafael, CA) was used to calculate Spearman's rho correlations and conduct the Wilcoxon signed rank tests. SAS version 9.1 (SAS Institute, Cary, NC) was used to calculate ICCs and effect sizes. Statistical significance was defined as P ≤ 0.05 for all statistical tests.

RESULTS

Subject profile.

Twenty-eight adults participated in the initial phase of the study to establish reliability and validity of the HAP. Demographic information is listed with descriptive statistics in Table 1. Our study population was predominantly female, and the subjects with RA were slightly younger than subjects with OA.

Table 1. Demographics of subjects for reliability/validity phase
 Osteoarthritis (n = 16)Rheumatoid arthritis (n = 12)All subjects (n = 28)
Mean ± SD age, years60.06 ± 8.0553.25 ± 14.5957.14 ± 11.60
Age range, years40–6924–6924–69
Number of women/men14/210/224/4

Thirteen subjects participated in the responsiveness-to-change phase of the study (7 with OA and 6 with RA). The mean ± SD age of these subjects was 60.67 ± 7.96 years, with an age range of 43–69 years. This sample of subjects was entirely female.

Test-retest reliability.

The mean HAP scores from the first and second administration of the questionnaire were very similar (Table 2). The ICCs for both HAP scores ranged from 0.60 to 0.91, demonstrating moderate to good reliability (32). Although all ICCs shown were statistically significant, the test-retest reliability was stronger for subjects with RA compared with those with OA. The AAS was more reliable for all subjects compared with the MAS.

Table 2. Test-retest reliability of the HAP using intraclass correlation coefficients*
 Osteoarthritis (n = 16)Rheumatoid arthritis (n = 12)All subjects (n = 28)
  • *

    Unless otherwise indicated, values are the mean ± SD. HAP = Human Activity Profile; MAS1 = maximum activity score from first HAP; MAS2 = maximum activity score from second HAP; ICC = intraclass correlation coefficients; 95% CI = 95% confidence interval; AAS1 = adjusted activity score from first HAP; AAS2 = adjusted activity score from second HAP.

  • Two subjects failed to return the second copy of the HAP. Thus the n for MAS2 and AAS2 is one less in each of the subject groups.

  • P ≤ 0.05.

MAS
 MAS172.75 ± 8.2877.17 ± 10.0074.64 ± 9.15
 MAS274.87 ± 6.9076.55 ± 10.3175.58 ± 8.36
 ICC (95% CI)0.60 (0.14–0.84)0.88 (0.62–0.97)0.76 (0.53–0.88)
AAS
 AAS164.50 ± 10.6071.50 ± 13.6167.50 ± 12.26
 AAS266.33 ± 10.4069.18 ± 14.7367.54 ± 12.23
 ICC (95% CI)0.83 (0.58–0.94)0.91 (0.70–0.98)0.87 (0.74–0.94)

Validity.

There was a wide range among the scores for the HAP (MAS 57–94; AAS 43–94). The descriptive statistics computed for the other self-report questionnaires and physical performance tests from the reliability/validity phase of the study are shown in Table 3. Like the HAP, the SF-36 Physical Functioning and Role Functioning-Physical scales and physical performance tests had a wide range of scores. However, the scores of the M-HAQ and AIMS2 were clustered near the low end of the possible range for these questionnaires.

Table 3. Results of self-report questionnaires and physical performance tests*
 Mean ± SDRange of scoresPossible range
  • *

    Unless noted, n = 28. SF-36 = Medical Outcomes Study 36-Item Short Form Health Survey; M-HAQ = Modified Health Assessment Questionnaire; AIMS2 = Arthritis Impact Measurement Scale 2; VO2max = maximum oxygen consumption.

  • The AIMS2 Physical Component consists of the average of the Mobility, Walking and Bending, Hand and Finger Function, Arm Function, Self Care, and Household Tasks scales.

  • Nine subjects either could not complete the submaximal treadmill test according to protocol, or the results were not valid due to medications.

  • §

    Four subjects were unable to complete the Timed-Stands Test according to protocol.

  • Two subjects were unable to complete the 50-Foot Walk Test according to protocol.

Self-report questionnaires
 SF-36 Physical Functioning64.76 ± 21.3115–950–100
 SF-36 Role Functioning-Physical49.11 ± 42.210–1000–100
 M-HAQ0.19 ± 0.240–0.750–3
 AIMS2 Physical Component1.27 ± 0.760.08–2.750–10
Physical performance tests
 VO2max estimate, ml/kg · min (n = 19)31.20 ± 7.5322.73–55.22
 Timed-Stands Test, seconds (n = 24)§25.03 ± 7.1112.82–41.00
 50-Foot Walk Test, seconds (n = 26)9.48 ± 2.695.31–16.89

The correlations for the MAS and AAS with each of the other self-report questionnaires are shown in Table 4. Significant correlations were found between the 2 HAP scores and the SF-36 Physical Functioning and Role Functioning–Physical scales, the M-HAQ, and the Physical Component of the AIMS2. Individual scales of the AIMS2 did not significantly correlate with the HAP scores with the exception of the Household Tasks scale and the AAS.

Table 4. Spearman's rho correlations of HAP scores with other self-report questionnaires*
 MASAAS
  • *

    Unless noted, n = 28. HAP = Human Activity Profile; MAS = maximum activity score; AAS = adjusted activity score; SF-36 = Medical Outcomes Study 36-Item Short Form Health Survey; M-HAQ = Modified Health Assessment Questionnaire; AIMS2 = Arthritis Impact Measurement Scale 2.

  • P ≤ 0.05.

  • The AIMS2 Physical Component consists of the average of the Mobility, Walking and Bending, Hand and Finger Function, Arm Function, Self Care, and Household Tasks scales.

SF-36 Physical Functioning0.780.80
SF-36 Role Functioning-Physical0.710.73
M-HAQ−0.49−0.51
AIMS2
 Mobility−0.13−0.21
 Walking and Bending−0.36−0.33
 Hand and Finger Function−0.040.05
 Arm Function−0.17−0.28
 Self Care−0.22−0.24
 Household Tasks−0.39−0.53
 Physical Component−0.49−0.54

Significant correlations were also found between the 2 HAP scores and all physical performance tests (Table 5). Persons with better scores on the HAP had higher levels of aerobic fitness, as evidenced by the positive correlation with VO2max. Persons with better scores on the HAP took less time to walk 50 feet and to complete the Timed-Stands Test, as evidenced by the negative correlations.

Table 5. Spearman's rho correlations of HAP scores with physical performance tests*
 MASAAS
  • *

    HAP = Human Activity Profile; MAS = maximum activity score; AAS = adjusted activity score; VO2max = maximum oxygen consumption.

  • Nine subjects either could not complete the submaximal treadmill test according to protocol, or the results were not valid due to medications.

  • P ≤ 0.05.

  • §

    Four subjects were unable to complete the Timed-Stands Test according to protocol.

  • Two subjects were unable to complete the 50-Foot Walk Test according to protocol.

VO2max estimate (n = 19)0.760.85
Timed-Stands Test (n = 24)§−0.77−0.72
50-Foot Walk Test (n = 26)−0.70−0.71

Responsiveness to change following a 12-week exercise program.

There was evidence that the intensity of the exercise program was sufficient to cause a change in physical status. Wilcoxon's signed rank tests revealed that there was a change in the estimated VO2max, the Timed-Stands Test, and the 50-Foot Walk Test (P < 0.01 for all) following the exercise intervention.

The HAP's responsiveness to change was similar to that of other questionnaires examining physical function. The baseline means and standard deviations, mean change, and effect size of scores on the self-report questionnaires in response to 12 weeks of exercise training are summarized in Table 6.

Table 6. Effect size for self-report questionnaires (n = 13)*
Self-report questionnairesBaseline mean ± SDMean change ± SE of changeEffect size
  • *

    SE = standard error. HAP = Human Activity Profile; MAS = maximum activity score; AAS = adjusted activity score; SF-36 = Medical Outcomes Study 36-Item Short Form Health Survey; M-HAQ = Modified Health Assessment Questionnaire; AIMS2 = Arthritis Impact Measurement Scale 2.

  • The AIMS2 Physical Component consists of the average of the Mobility, Walking and Bending, Hand and Finger Function, Arm Function, Self Care, and Household Tasks scales.

HAP
 MAS70.54 ± 6.322.9 ± 1.10.5
 AAS61.54 ± 9.854.5 ± 2.00.5
SF-36 Physical Functioning59.87 ± 20.559.7 ± 2.60.5
SF-36 Role Functioning-Physical38.46 ± 39.0128.8 ± 9.30.7
M-HAQ0.31 ± 0.29−0.1 ± 0.10.5
AIMS2
 Mobility0.81 ± 0.97−0.2 ± 0.30.2
 Walking and Bending4.04 ± 2.96−1.3 ± 0.60.4
 Hand and Finger Function1.92 ± 1.89−1.0 ± 0.40.5
 Arm Function0.62 ± 1.060.0 ± 0.20.0
 Self Care1.06 ± 2.21−0.8 ± 0.60.3
 Household Tasks0.48 ± 0.99−0.2 ± 0.20.2
 Physical Component1.49 ± 0.83−0.6 ± 0.20.7

DISCUSSION

The findings from this study establish the test-retest reliability and validity of the HAP as a measure of physical function and physical activity for persons with RA and OA. Our results also support its use as a measure of physical function that is responsive to change in persons with arthritis.

We found moderate to good (32) and significant reliability for both the MAS (ICCs 0.60–0.88) and AAS (ICCs 0.83–0.91) as indicated in Table 2. The HAP was more reliable for persons with RA than for those with OA, possibly due to the greater range of HAP scores for subjects with RA. The same effect holds true for the AAS versus the MAS.

Other investigators have found good reliability coefficients (r = 0.79 to 0.97) for both HAP scores (21–23, 25). Our ICCs were slightly lower than those previously reported for a sample of adults with knee OA (22). This finding may be due to the fact that we did not restrict involvement to a specific joint, and the time lapse between administrations of the HAP was slightly shorter (2–7 days) in the study of adults with knee OA (22) than in our study (5–14 days). Our slightly longer time period may have prevented subjects from answering the second questionnaire based on the memory of their previous answers, rather than truly reflecting on their activity levels.

We found moderate to excellent (33) and significant correlations between both HAP scores and each of the 3 questionnaires measuring physical function (r = −0.49 to 0.80), the Timed-Stands Test, and the 50-Foot Walk Test (r = −0.70 to −0.77). These results provide evidence of the validity of the HAP as a measure of physical function for persons with RA and OA. The lowest significant correlations found in this study were between the HAP scores and the AIMS2 Physical Component and M-HAQ (r = −0.49 to −0.54). Only 1 of the 6 subscales of the Physical Component score of the AIMS2 correlated with the HAP, the Household Tasks subscale (r = −0.53). The low correlation with the M-HAQ can be explained by acknowledging that the questions asked in the M-HAQ assess only low-level activities. Our subjects had scores clustered near the low end of the possible range, implying that the M-HAQ had a ceiling effect in our sample. The scores on the AIMS2 Physical Component were also clustered near the low end of the possible range. In contrast, the range of our subjects' HAP scores indicated a wide variety of functional levels. Therefore, the HAP appeared to be more effective than the M-HAQ and AIMS2 in detecting existing impairments in our study population.

The results of other studies also support the use of the HAP as a measure of physical function in persons with chronic disease. In a sample of hemodialysis patients, Johansen et al (21) found significant correlations between the Physical Functioning scale of the SF-36 and the MAS (r = 0.68) and AAS (r = 0.82). The same study found significant correlations ranging from –0.48 to 0.75 between both HAP scores and physical performance tests such as the 50-Foot Walk Test, time to climb 12 steps, and time to rise from a chair 5 times (21). Bennell et al (22) also found significant correlations between the HAP scores and physical performance tests in a sample of adults with knee OA. The strength of their correlations was slightly lower, ranging from 0.34 to 0.63 (22). Furthermore, Stuifbergen (24) found significant correlations between the MAS (r = 0.67) and AAS (r = 0.91) and the Physical Functioning scale of the SF-36 in a population of adults with multiple sclerosis.

We found good to excellent (33) and significant correlations between the MAS and VO2max (r = 0.76) and the AAS and VO2max (r = 0.85), demonstrating the usefulness of the HAP as an indicator of physical activity and fitness. The developers of the HAP found a good to excellent and significant correlation (r = 0.83) between VO2max and the ADAPT Quality-of-Life Scale (the original name of the HAP) (20). Our findings were anticipated because high scores on the HAP reflect an increased activity level commonly associated with good physical fitness and a high VO2max. The usefulness of the HAP to predict VO2max and fitness level without conducting a treadmill test in this population is worth further exploration.

Despite the reliability and validity of the HAP, we recognize that this questionnaire inherently has some degree of subjectivity, as do all questionnaires. The HAP instructions ask the respondent to answer if they are Still doing this activity,” “Have stopped doing this activity,” or “Never did this activity.” We attempted to make these instructions clear, but misinterpretation on the part of the subjects may have influenced their HAP scores in ways that we could not detect. Although we were present to make clarifications at the first administration of the HAP, the subjects completed the second copy at home without any assistance. Another limitation of the HAP is that it only assesses whether a person has actually “done” an activity, not whether they “could” do an activity. In addition, several subjects stated that although they could do an activity, they had pain or difficulty in doing so.

Exercise can improve fitness and physical function in persons with arthritis (8–11). However, a weakness common in self-report questionnaires is the lack of sensitivity to change in physical function in response to an exercise program, especially in high-functioning subjects. We found that all of the questionnaires used in our study had medium to large effect sizes (34). Despite what appeared to be ceiling effects of the M-HAQ and the AIMS2 when our high-functioning subjects were initially assessed with these questionnaires, these 2 questionnaires had effect sizes similar to that of the HAP scores. One might intuitively expect little room for improvement in our subjects with the M-HAQ and AIMS2. However, because of the low variability of our sample on these questionnaires (see baseline SDs in Table 6), a small mean change in these scores still resulted in a medium to large effect size.

Despite the similar effect sizes between the questionnaires we investigated, we feel that the HAP would be more useful for assessing function in persons with arthritis than the other questionnaires investigated due to its ease of use and broad scope. Compared with the HAP, the AIMS2 has complex scoring with calculation of several scales to arrive at the Physical Component, which is the portion of the AIMS2 assessing physical function that had the strongest effect size. Although the Physical Functioning and Role Functioning-Physical scales of the SF-36 could be administered separately from the remainder of the SF-36, the scoring is also cumbersome. In contrast, to score the HAP, one simply has to note what activities a person is still performing and those they can no longer perform. In addition to being easier to complete and score than the other questionnaires, the HAP covers a wider range of activities. With one quick glance, the practitioner can identify limitations even in a high-functioning patient. With repeated administration of this questionnaire upon subsequent clinic visits, practitioners can detect subtle changes or declines in physical function, which can prompt early intervention to minimize future consequences of the disease.

In conclusion, we found the HAP to be a valid, reliable, and responsive tool to assess the physical function of individuals with arthritis. We suggest that the HAP can be used by health care professionals to identify limitations in high-functioning patients with arthritis. The HAP shows promise in evaluating the efficacy of treatment interventions and may be able to detect declines in function over time, which are 2 critical and necessary functions of useful assessment and outcome tools in health care.

Acknowledgements

The authors would like to thank Jason Moore, Jennifer Sibley, Renee Schroeder, and Amy Goldman for their assistance with data collection and analysis.

Ancillary