Validation of the PASE in older adults with knee pain and physical disability

Abstract

Validation of the PASE in older adults with knee pain and physical disability. Med. Sci. Sports Exerc., Vol. 31, No. 5, pp. 627-633, 1999.

Purpose: 

To examine the validity of the Physical Activity Scale for the Elderly (PASE) among individuals with disability.

Methods: 

A sample of 471 participants (mean age = 71.36) in an epidemiological study of chronic knee pain completed the PASE and self-report measures of knee pain, perceived physical function, satisfaction with physical function, and importance of physical function. A 6-min walk test and an isokinetic assessment of knee strength were also administered.

Results: 

PASE scores were significantly correlated in expected directions with performance on the 6-min walk, knee strength, frequency of knee pain during transfer, and perceived difficulty with physical functioning. Gender and age were identified as significant moderators of PASE scores and the scale's construct validity was supported by testing a conceptually driven hypothesis regarding patterns of physical activity.

Conclusions: 

These results support the PASE's validity for the assessment of physical activity among older adults with pain and disability.

Little is known about the patterns of physical activity among older people with chronic illness, symptoms, and disability. This is due, in part, to a lack of reliable and valid measures of physical activity among the aged. Specifically, most activity questionnaires do not assess involvement in less strenuous activities such as light housework, low-intensity recreational sports, gardening, and casual walking; yet, these activities may best represent energy expenditure and physical activity in older populations ^(10,11,19). An exception, however, is the Physical Activity Scale for the Elderly (PASE; ²¹).

The PASE is a self-report/interview-based measure designed to capture and assess occupational, household, and leisure activities typically performed by older adults; that is, those who are 65 yr of age and older. For example, under the category of leisure activities, separate items query the amount of time individuals spend each week (a) walking, (b) in light sport and recreational activities, (c) in moderate sport and recreational activities, (d) in strenuous sport and recreational activities, and (e) exercising to increase muscular strength and endurance. Time spent participating in each activity area is multiplied by a weighted value that reflects the amount of energy expended by an older person engaged in that activity. These weighted values are then summed to yield a composite PASE score ⁽²¹⁾.

Preliminary data indicate that the PASE is a valid and reliable measure of physical activity among adults who have no serious physical limitations ⁽²⁰⁾. In a sample of 222 individuals between the ages of 65 and 100 yr, PASE scores were significantly correlated with balance, grip strength, leg strength, self-assessed health status, and the Sickness Impact Profile. Over repeated administrations 3-7 wk apart, the test-retest reliability coefficient of the PASE was 0.75 for self-administration and 0.68 when administered during a telephone interview. Although these preliminary results are encouraging, we are unaware of any other published work on the validity of the PASE. Moreover, it is not known whether the PASE is applicable to older adults whose physical function is compromised by chronic disease. This is significant considering the prevalence of chronic diseases in older adults and the increased attention being paid to physical activity interventions in the treatment of chronic disease and physical disability ⁽¹⁷⁾.

Given the need for establishing a valid and reliable measure of physical activity among older individuals with physical limitations ⁽¹⁷⁾, four objectives were established for the present study. The first was to provide normative data on the PASE for an older group of patients who are participating in an observational arthritis study (OASIS). OASIS is a 30-month prospective epidemiological trial of older adults who have activity restrictions due to knee pain. Given their age and physical limitations, OASIS patients were considered an ideal population in which to examine the psychometric properties of the PASE.

A second objective was to identify moderators of PASE scores as there has been a call for research to identify moderators of physical activity among individuals with disability ⁽¹⁷⁾. In this study the moderating effects of demographic variables, measures of disease, and comorbid conditions were assessed.

The third objective was to demonstrate convergent validity for the measure by correlating PASE scores with objective and subjective indices of patients' symptoms and disability. These measures included performance on a 6-min walking task, leg strength, perceived difficulty performing activities of daily living, and knee pain.

Finally, consistent with recommendations by Nunnally ⁽¹³⁾, a test of construct validity was performed by testing a conceptually driven hypothesis regarding patterns of physical activity. This hypothesis was based on research examining physical disability. Specifically, we recently completed a study of the determinants of satisfaction with physical function among patients with knee pain ⁽¹⁶⁾. In this investigation, we found that the value that patients placed on their physical function moderated satisfaction ratings. The greatest dissatisfaction was expressed among those who placed high value on their function despite considerable disability. We surmised hat disabled individuals who placed less value on their physical function would also be less active, a status that is a risk factor for the progression of disability ⁽¹⁸⁾. Thus, in the present study, we predicted that value judgments would have a moderating effect on PASE scores. That is, patients with relatively severe limitations who valued their physical function would have higher PASE scores than those who placed a low value on their physical function.

METHODS

Participants. This investigation involved a series of analyses performed on baseline data that were collected on 471 patients (240 females and 231 males) who are part of a prospective study of chronic knee pain-the Observational Arthritis study in Seniors (OASIS). Table 1 presents demographic information for the sample.

TABLE 1: Demographic characteristics of the OASIS sample.

The eligibility criteria for participation in the study were (a) age 65 yr or older, (b) knee pain on most days of the week, and (c) difficulty with at least one of the following due to knee pain: walking a quarter mile, climbing stairs, getting in and out of a car, rising from a chair, lifting and carrying groceries, getting out of bed, getting out of the bathtub, or performing shopping, cleaning, or self-care activities. Potential participants were excluded if they (a) were moving from the area within 3 yr, (b) were under hospice care, (c) were receiving active treatment for cancer (other than skin cancer), (d) had shortness of breath or chest pain at rest, (e) had a mini-mental score < 24 as measured by the Mini-Mental State Examination ⁽¹⁹⁾, (f) had rheumatic or psoriatic arthritis, or (f) were currently participating in another study.

Participants were recruited via a commercial list on which names were stratified into blocks of equal size, and blocks were contacted randomly. All potential participants within a randomly selected block received a brochure in the mail before being contacted by telephone. During the telephone calls, a recruiter administered a brief (less than 5 min) screening questionnaire that assessed basic demographic information and presented the major inclusion/exclusion criteria described above. Responses to these latter questions were essential to determine whether a person qualified for a screening visit. All participants provided informed written consent for their participation with the experimental procedures conducted in conformance with the human subjects policy statement published by Medicine and Science in Sports and Exercise. The study was approved by the Institutional Review Board at Wake Forest Medical School.

Measures. The measures in this investigation included questionnaires, a 6-min walk test, an isokinetic assessment of knee strength, and anthropometric measures used to determine BMI. The self-report, performance, and anthropometric measures were taken under controlled laboratory conditions on a schedule that precluded interference between measures.

All questionnaire items and response scales were presented in their original published and validated formats. Consequently, there was variability in the questionnaire response scales used in this study (e.g., the PASE has a 4-point response scale and the Knee Pain Scale has a 5-point response scale). Detailed explanations of these measures and their response scales are provided below.

PASE. The PASE is a 12-item scale that measures the average number of hours per day spent participating in leisure, household, and occupational physical activities over the previous 7-d period. Under the category of leisure time activity, frequency and duration of participation in the following five activities were assessed: walking outside the home, light sport/recreation, moderate sport/recreation, strenuous sport/recreation, and muscle strength/endurance exercises. Frequency of participation was assessed by asking respondents how often they engaged in the activity over the past week. They responded using the following 4-point scale: 0 = never; 1 = seldom/1-2 d; 2 = sometimes/3-4 d; and 3 = often/5-7 d. Duration was assessed by asking participants on average, how may hours per day they spent on that activity. Responses were indicated using a 4-point scale: 1 = less than 1 h; 2 = 1 but less than 2 h; 3 = 2-4 h; and 4 = more than 4 h. By using a conversion table ⁽²¹⁾, these categorical indices of frequency and duration were used to calculate the average daily frequency of participation for each activity. To ensure that participants correctly classified their sport and recreational activities as light, moderate, or strenuous, participants recorded the specific activities (e.g., walking the dog, golfing, or dancing) in which they participated.

Under the category of household physical activities, participation in six activities was assessed: light housework, heavy housework or chores, home repairs, lawn work or yard care, outdoor gardening, and caring for another person. Using a binary scale (1 = no, 2 = yes) respondents indicated whether they engaged in these behaviors over the past 7 d. These scores were then converted to a frequency. If a household activity was performed in the past 7 d, it was assigned an average daily frequency of 1. If it was not performed, then its average daily frequency was 0.

Under the category of work-related physical activity, participants indicated whether they worked for pay or as a volunteer (1 = no, 2 = yes). Those who worked were also asked to indicate the number of hours that they worked, and to characterize their job as (1) mainly sitting with slight arm movements, (2) sitting or standing with some walking, (3) walking with some handling of materials generally weighing less than 50 pounds, or (4) walking and heavy manual work often requiring handling of materials weighing over 50 pounds. If a respondent's job was characterized as "mainly sitting with slight arm movements," then the average daily frequency for work-related physical activity was 0 h. For all other job types, average daily frequency was computed by dividing the number of hours worked in the past week by 7 (days).

Washburn et al. ⁽²⁰⁾ devised a set of activity weights for each of the 12 PASE items. These weights reflect the amount of energy expended by an older person engaged in that activity. PASE item scores were calculated by multiplying the activity weight by the activity daily frequency. The PASE items were then summed to create an overall PASE score. Washburn et al. ⁽²¹⁾ reported that PASE scores may range from 0 to 400 or higher.

Perceptions of physical function. Based on previous research ⁽¹⁴⁾ and clinical expertise provided by two rheumatologists, six activities of daily living were identified that frequently cause difficulty for patients with knee OA. These activities were stair climbing, walking, getting up and moving after having been seated, doing light work around the home, doing heavy work around the home, and lifting and carrying objects. Perceived difficulty was assessed using a measurement methodology similar to that used in self-efficacy research ⁽¹⁾. That is, for each activity, participants were presented with a hierarchy of five levels of activity arranged in ascending order of difficulty and were asked to indicate how much difficulty they would have performing at each level. For example, for stair climbing, participants indicated how much difficulty they would have climbing 3 steps, 7 steps, 14 steps, and 21 steps. Participants responded using a 5-point scale: 0 = no difficulty, 1 = a little difficulty, 2 = moderate difficulty, 3 = a lot of difficulty, and 4 = unable to do. Scores were created by reverse scoring each step, summing across all 5 steps and then multiplying this result by 5. Thus, for each activity, participants had a perceived difficulty score that could range from 0 to 100 with higher scores representing less difficulty. Previous analyses have shown that a single and valid index of perceived difficulty can be created by taking an average score across all six activities ⁽¹⁶⁾. Hence, for all subsequent analyses, an average perceived difficulty score was calculated and used as a measure of perceived difficulty with physical function.

Satisfaction and importance. After rating their perceived difficulty with the six activities, participants were then asked to rate their level of performance satisfaction with each activity on an 11-point scale ranging from −5 to +5. The following verbal anchors appeared at each odd interval: very dissatisfied (−5), moderately dissatisfied (−3), slightly dissatisfied (−1), slightly satisfied (+1), moderately satisfied (+3), and very satisfied (+5). In addition, the importance of each of the six activities to the patient was assessed using a 5-point scale: 0 = not at all important, 1 = somewhat important, 2 = moderately important, 3 = very important, and 4 = extremely important. Previous factor analytic work of ours has demonstrated that composite scores for both satisfaction and importance can be created by averaging ratings across the six activities (cf., ¹⁶).

Knee pain. Knee pain was assessed by the Knee Pain Scale (KPS; ¹⁵). The KPS consists of four three-item subscales that assess the frequency and intensity of knee pain experienced during activities of daily living that require either ambulation or transfer. For each activity, participants rated the frequency of the pain on a 5-point scale: 1 = never, 2 = almost never, 3 = sometimes, 4 = almost always, and 5 = always. In addition, the intensity of the pain was rated on a 6-point scale: 1 = no pain, 2 = mild pain, 3 = discomforting pain, 4 = distressing pain, 5 = horrible pain, and 6 = excruciating pain. Mean scores were calculated for each subscale. The factor structure, as well as reliability and validity for the KPS have been reported in published research ⁽¹⁵⁾.

Six-minute walk. For the 6-min walk, participants were asked to walk as far as they could in 6 min. The test was performed in an indoor exercise facility. Patients were asked to begin walking at the command "go" and to continue walking until they heard the command "stop." Patients were not allowed to carry a watch and were not given any feedback during the trial. Performance was measured as the total distance covered in feet. Rejeski and his colleagues ⁽¹⁵⁾ have published validity and reliability data on this test for patients with knee osteoarthritis.

Knee strength. A measure of isokinetic strength of the knee joint was obtained using an isokinetic dynamometer (Kin-Com, 125E, Chattanooga Corp., Chattanooga, TN). The measure of knee strength used in this study represents an average of standardized scores for four components: concentric flexion, concentric extension, eccentric flexion, and eccentric extension. Each component was tested through a joint range of motion from 90° to 30°. The resulting score is a maximum torque in Newton-meters for the combined components assessed at an angular velocity of 30°·s⁻¹.

Anthropometrics. Body mass index (BMI) was calculated (kg·m⁻²) from height and weight. These measures were made by a trained research assistant.

RESULTS

Due to missing data, the number of participants in each of the following analyses ranged from 453 to 471. Because no more than 4% of the total sample was missing for any given analysis, this slight fluctuation was not considered problematic.

Descriptive statistics. For the sample, PASE scores ranged from 0 to 419. The mean PASE score was 131.4 (SD = 71.1) and the median was 119.2. Table 2 presents the contribution of each scale item to the overall PASE score as determined by the product of the sample mean and the activity weight. The first six items are scored in terms of the number of hours of activity per day. The remaining six items are scored as 1 = engaged in the activity or 2 = did not engage in that activity during the previous 7 d. The activities making the largest average individual contributions to the total PASE score were walking, light housework, lawn work, and heavy housework.

TABLE 2: PASE item weights and contributions to total PASE score.

Total PASE scores for our sample were compared with preliminary scale norms that were established in a general population of older adults ⁽²¹⁾. In the PASE scoring manual, norms were presented separately for men and women across three age groups. To provide a more accurate comparison of PASE norms with scores obtained in the present investigation, study participants were grouped according to these six categories. Single sample t-test comparisons were then conducted. Table 3 presents PASE scores for the present sample and normative values as a function of age and gender. PASE scores from the present sample were significantly higher than the norms for five of the six groups.

TABLE 3: PASE scores from the scoring manual and the OASIS sample as a function of age and gender.

Moderating variables. The second objective of our investigation was to identify demographic variables and comorbid conditions that may moderate physical activity in older adults. Table 4 presents the means and standard deviations for the PASE score as a function of gender, race, emphysema, hypertension, diabetes, and history of angina and heart attack. For each of these categorical variables, a general linear model was conducted to test for between group differences. The level of significance was set at 0.05.

TABLE 4: Means and standard deviations for the PASE as a function of Moderators.

Gender was the only significant moderator of PASE scores. Consistent with previous data ⁽²¹⁾, men had higher PASE scores than women, P < 0.001. Also consistent with previous research ⁽³⁾, there was a nonsignificant trend (P = 0.06) for patients with COPD to report less physical activity than those without COPD. There were no effects for any of the other moderators.

In addition, bivariate correlations were calculated to determine the association between PASE and two continuous variables, age and BMI. PASE scores were significantly and negatively correlated with age (r = −0.21, P < 0.001). Older patients reported less activity than their younger counterparts. The relationship between PASE scores and BMI was not significant (r = −0.04, P > 0.05).

Convergent validity. As an index of convergent validity, bivariate correlations were calculated between the total PASE score and the measures of physical functioning. PASE scores were significantly correlated with perceived difficulty with physical functioning (r = 0.35, P < 0.001) and performance on the 6-min walk (r = 0.35, P < 0.001). As one would expect, individuals who reported engaging in more physical activity (i.e., higher PASE scores) reported greater ease in performing the six different activities of daily living and they were able to walk farther in 6 min.

PASE scores were also significantly correlated with knee strength (r = 0.41, P < 0.001). Individuals who were more physically active performed better on the knee strength test. Turning to the KPS, there was a small but significant correlation between PASE scores and frequency of knee pain during transfer (r = −0.09, P < 0.05). As expected, patients who experienced more frequent pain during transfer were less physically active. Contrary to expectation, the relationships between PASE scores and the other three Knee Pain Subscales (knee pain intensity during transfer, knee pain frequency and intensity during ambulation) were nonsignificant (r ranged from −0.07 to −0.08, all P > 0.05).

Satisfaction ratings of physical function, values, and PASE scores. As an additional step in construct validation, we evaluated the hypothesis that patients with relatively severe limitations in physical function who valued their physical function would be more physically active than those who had similar levels of disability, yet placed a low value on their physical function. Based on median splits, participants were classified as either high or low on both the dimensions of physical function and the value placed on physical function. PASE scores were used as the dependent variable. A general linear model was employed to evaluate the 2 (high or low physical function) × 2 (high or low value) study design with an a priori contrast employed to evaluate the previously stated hypothesis ⁽⁸⁾. As predicted, patients who had low levels of physical function, yet who continued to value their physical function, had higher PASE scores than those with similar levels of physical function who placed lower value on their function, t(467) = 2.80, P < 0.01 (see Table 5 for means and standard deviations). In addition, the linear model produced significant main effects for both perceived physical function F(1,467) = 33.46, P < 0.001, and value placed on physical function, F(1,467) = 5.82, P < 0.05. As expected, patients who had greater perceived physical function were more active (mean = 149.20, SD = 71.83) than those who reported poorer physical function (mean = 113.04, SD = 65.65). Also, those who placed greater value on their physical function were more active (mean = 138.81, SD = 67.61) than those who valued it less (mean = 122.53, SD = 74.33).

TABLE 5: Means and standard deviations for PASE scores as a function of difficulty with and value placed on physical function.

DISCUSSION

In recent years, researchers have devoted increasing attention to the use of physical activity interventions to treat chronic disease and disability in the aged. Subsequently, the need for a valid and reliable measure of physical activity for older adults with physical disability has become apparent ⁽¹⁷⁾. The PASE has been identified as a suitable measure of physical activity among older adults; however, there is limited psychometric support for the measure and what little that does exist is restricted to individuals without physical limitations. To address this shortcoming, we administered the PASE to a sample of older patients who had chronic knee pain. This approach produced several noteworthy findings.

First, examination of the PASE subscale scores revealed that walking, housework, and lawn work were the most frequently performed physical activities among this cohort. This finding is consistent with other research that has examined the physical activity patterns of older people with lower extremity disabilities ⁽⁶⁾ and those without physical impairments ⁽²⁰⁾. Given that these activities represent the greatest sources of energy expenditure among the elderly (see Table 2), the utility of the PASE is clear. The PASE is one of the only self-report measures of physical activity that assesses participation in such low level, yet significant, activities.

Interestingly, despite their physical limitations, the patients in our sample had significantly higher PASE scores than the normative values reported in the PASE scoring manual. There are at least two possible explanations for this. First, climate appears to influence PASE scores. Washburn and his colleagues ⁽²⁰⁾ reported that among their sample of older Massachusetts residents, PASE scores ranged from 80 in January to 140 in August. Given that our data were collected in a southern climate over a 12-month period, and that two of the most frequently cited physical activities among older people take place outdoors (i.e., walking and yardwork), it is not surprising that our sample had higher PASE scores.

Another possibility is that our sample was more physically active because they had been prescribed exercise by a doctor. Exercise therapy is often prescribed to patients with knee pain and knee osteoarthritis ⁽²⁾. Moreover, research indicates that when older knee osteoarthritis patients are given medical advice to exercise, the majority attempt to comply ⁽⁴⁾. Thus, our sample may have been more physically active than Washburn's normative sample ⁽²¹⁾ because participants were complying with their doctors' orders. Given the potential influence of climate and motivational factors (e.g., exercise prescriptions) on PASE scores, we recommend that researchers consider the influence of these variables when comparing their scores with the norms generated in the present study.

A second relevant finding pertains to the identification of moderators of PASE scores. Men were significantly more active than women, and there was a trend for patients with emphysema to be less active than those patients who did not have emphysema. Age was also significantly correlated with PASE scores. These findings attest to the importance of identifying populations that are at high risk for sedentary behavior and developing interventions to suit their specific needs ⁽⁹⁾. For example, as the burden of comorbidities increase, it is highly likely that physical activity programs need to be taught within the broader framework of a cognitive-behavioral intervention that is designed to help patients manage the limitations and symptoms of their medical conditions.

Third, the present findings attest to the validity of the PASE as a measure of physical activity among older adults with disability. Specifically, PASE scores correlated in expected directions with both objective (6-min walk and knee strength) and subjective (perceived difficulty with physical functioning, frequency of pain during transfer) measures of patients' disability. These results support the convergent validity of the scale. As an aside, we suggest that future research examine whether muscle strength is a cause or an effect of higher PASE scores. Clearly, loss of muscular strength is a significant concern among aging patients with knee osteoarthritis. Patients with knee osteoarthritis show strength deficits relative to age, mass, and gender-matched healthy populations ⁽¹²⁾. Moreover, among knee osteoarthritis patients, knee and ankle muscular strength account for approximately 25% of the variability in dynamic stability measures ⁽⁷⁾. Understanding the direction of the relationship between physical activity and knee strength has implications for developing interventions for older adults with osteoarthritis.

Fourth, the PASE's construct validity was supported by testing a conceptually driven hypothesis regarding patterns of physical activity. As hypothesized, patients who had low levels of physical function yet who continued to value their physical function were more active (i.e., had higher PASE scores) than those with similar levels of physical function who placed lower value on their function. This finding highlights the moderator role of value in determining whether people with disability will try to remain active. As reflected by the PASE scores, patients who value their physical functioning are more likely to try and maintain what functioning they do have by continuing to be physically active. This strategy can be effective for coping with the physical consequences and deteriorations associated with certain diseases and the aging process ⁽¹⁸⁾.

In contrast, some patients may cope with physical decline by decreasing the value that they place on physical functioning. This strategy may reflect a psychological coping strategy; by downplaying the value of being functional, a disabled person may not be as distressed by changes in his or her level of physical functioning (cf., ⁵). However, this tactic can be detrimental. As reflected by the PASE scores, poorly functioning individuals who place lower value on their physical function are less likely to be physically active. Hence, although devaluation of physical function probably has short-term psychological benefits to the disabled individual, these benefits are likely overshadowed by the long-term health consequences of being inactive. This possibility speaks to the need for coupling cognitive-behavioral interventions with physical activity interventions. Clearly, how people think about their physical function (i.e., the value they place on it) has implications for their motivation to enhance or maintain their level of function.

In summary, the results of this investigation support the validity of the PASE as a measure of physical activity in older individuals who have mild to moderate levels of physical disability in the lower extremities. Clearly one advantage of this measure is its content validity among older adults; PASE items reflect the types of physical activities that are most common among older people. In addition, PASE scores had reasonable levels of convergent validity and they appear sensitive to the differential values that people placed on physical activity.

Like any other self-report index of physical activity, the PASE does have limitations. Specifically, we have found through interviews that people tend to overestimate the time spent in various activities. Self-report biases can be attenuated by administering the questionnaire using an interview format ⁽²¹⁾. Trained interviewers can ask specific questions that prompt respondents to provide more accurate estimates of their time spent on physical activity. For example, a respondent may initially report 3 h of dancing at a party. However, when asked how many times she danced in 3 h, it may be revealed that she danced only seven or eight times (e.g., for a total of 45 min) in the 3-h period. The interviewer would then record 45 min of moderate sport/recreation-a value that is more accurate than the 3-h duration originally reported.

Also, the PASE does not discriminate well between levels of intensity within a given activity. For example, under the category of walking outside the home, 1 h of brisk walking for exercise and 1 h of strolling through the park receive the same PASE score. Likewise, under the category of lawn work, "lawn mowing is counted as lawn work regardless of the type of mower (riding, power, or push) used ⁽²¹⁾." Obviously, there are differences in intensity when one pushes versus rides a lawnmower. Such lack of specificity increases measurement error.

Despite these weaknesses, the PASE will enable investigators in OASIS to examine how patterns of physical activity influence the progression of osteoarthritis and chronic knee pain. In addition, the measure will permit OASIS's social scientists to explore the mediational role of social cognitive variables on physical activity participation over time. Researchers are encouraged to examine the utility of the PASE for studying the relationship between physical activity and other types of disability among the aged.