Validity and reproducibility of the Physical Activity Scale for the Elderly (PASE) questionnaire for the measurement of the physical activity level in patients after total knee arthroplasty

Because a sample size ≥ 50 should be used to evaluate the validity and reproducibility of physical activity questionnaires[12], a total of 50 consecutive TKA patients (25 men and 25 women) were included in the study (Table 1). The main inclusion criterion was primary unilateral TKA implanted at the Schulthess Clinic (Zurich, Switzerland) from April 2011 to March 2012. Specifically, patients were evaluated between 3 and 12 months postoperatively, because physical function and pain have recently been shown to attain preoperative values 3 months after TKA surgery, but no further significant PAL improvements have been reported 12 months postoperatively[2, 7]. Exclusion criteria were other artificial joints in the lower extremities, and symptoms or signs referable to overt cardiorespiratory, orthopedic, neurological or general diseases that could have negatively influenced the physical activity evaluation with questionnaires or accelerometers. The study was conducted according to the Helsinki declaration, and the protocol was approved by the Ethics Committee of the Canton of Zurich. All the subjects signed a written informed consent form before participating in the study.

Patients were invited to attend the laboratory on two occasions, separated by 7 days. They were explicitly asked to maintain their usual physical activity habits between the two test sessions. Therefore, we assumed that all patients maintained their usual PAL during this time period. During the first test session (hereafter referred to as test session 1), patients were initially asked to fill in a questionnaire enquiring about their sociodemographic characteristics, and the Western Ontario and McMaster Universities Arthritis Index (WOMAC) questionnaire. The WOMAC is a valid and reliable questionnaire widely used for evaluating knee pain, stiffness, and function in patients with osteoarthritis of the knee[22]. It has 24 items, and the scores vary between 0 (worst score) and 100 (best score). Patients were then asked to complete the PASE questionnaire. Subsequently, they were instructed to wear a portable accelerometer as much as possible until test session 2. The accelerometer was attached on an elastic belt and firmly fixed on the right hip. Patients were advised to remove the accelerometer for sleeping and during water-related activities (e.g., showering, swimming, water-gymnastic), and to note the wearing and non-wearing time periods in a daily log. Patients brought the accelerometer back at test session 2, and filled out the PASE questionnaire again. All patients received the same instructions by the same interviewer, and filled in the questionnaire together with the interviewer. Test-retest reproducibility was investigated by comparing the PASE scores obtained at test session 1 and 2. The construct validity of the PASE questionnaire was investigated by comparing the PASE scores at test session 2 to the physical activity outcomes measured by accelerometry.

The PASE is a 12-item scale evaluating the PAL of the past 7 days in three life domains: recreational, household and work-related activities. In the present study, the PASE was administered through an interviewer, because this modality has proven to be more reliable than the self-administered one[14]. Patients were asked to rate the weekly frequency and daily duration for each of the following recreational activities: walking outside the home, light, moderate and strenuous activities and muscle strengthening. Whether household activities (light and heavy housework, home repairs, lawn work/yard care, outdoor gardening and caring for others) were performed was captured by answering yes or no. Finally, working for pay or as a volunteer was assessed by recording the amount of hours per week and the type of work performed. For each activity a score was obtained by multiplying an activity frequency value by a task-specific weight provided by the scoring manual. The main outcomes are PASE total score and activity and intensity PASE sub-scores. The PASE total score, which represents the overall PAL, is the sum of all activities together, and ranges between 0 and 400 or more. The activity sub-score for recreational, household and work-related activities are computed by adding the activities corresponding to each life domain together[14]. Additionally, with an explorative intention, we calculated two PASE intensity sub-scores: one for low intensity activities and one for moderate-to-high intensity activities. We first determined, according to the physical activity compendium[23], the average metabolic equivalents of task (MET) for each item of the PASE. The classification of each single PASE questionnaire item into the different PASE intensity sub-scores was performed by consensus of two experienced investigators. Subsequently, according to the age-related classification of physical activity intensity provided by the U.S. Department of Health and Human Services[24], we defined the threshold between low and moderate-to-high intensity activities to be 3.5 MET and 4.5 MET for patients aged over and under 65, respectively. For patients older than 65 years (n = 40), the sum of the following items represented the low intensity activity PASE sub-score: walking outside the home, light recreational activity, light and heavy housework and caring for others. If the type of work was mainly sitting, or sitting and standing with some walking, the work-related activity score was added to the low intensity activities. Otherwise, it was added to the moderate-to-high intensity PASE sub-score. The sum of the remaining items (moderate and strenuous recreational activities, muscle strengthening, home repairs, lawn work/yard care, outdoor gardening) represented the moderate-to-high PASE sub-score. For patients younger than 65 years (n = 10), the score of the items home repairs and outdoor gardening were also included as low intensity activities.

We performed a cross-cultural adaptation of the PASE questionnaire into German according to the guidelines of the American Association of Orthopedic Surgeons Outcomes Committee[25]. The questionnaire was translated from English into German by an informed and an uninformed translator, who were both native German speakers. Then the two versions were combined into one under the supervision of a research methodologist. Two native English speakers, who were fluent in German, performed each a back translation. All involved persons resolved remaining problems and approved the final version of the German PASE questionnaire. Thus, physical activity examples reported in the single PASE items were strictly culturally adapted in order to be familiar to the elderly people living in Switzerland. Finally, the German version of the PASE questionnaire was pre-tested in 15 consecutive patients after total knee or hip arthroplasty to examine comprehensibility of wording and items. The final version was then approved by the New England Research Institute (the copyright owners of the PASE questionnaire).

Physical activity was objectively monitored using the Actigraph GT3X + accelerometer (ActiGraph, LLC, Pensacola, FL, USA). It is small (4.6 × 3.3 × 1.9 cm), light-weight (19 g) and measures accelerations in magnitudes of ± 6 g. The accelerometry-derived PAL (aPAL) is characterized by the total number of counts, which is the vector resulting from accelerations in the three orthogonal axes (vertical, antero-posterior and medio-lateral). In the present study, acceleration data was sampled at 100 Hz, filtered, digitalized and added over a user-specified time interval (epoch) of 60 s. Accelerometer data were analyzed and computed with ActiLife 5 software (ActiGraph software, LLC, Pensacola, FL, USA). For each patient, a complete data set was defined to have at least 10 h∙day^-1 of monitored wear during at least 5 days[26, 27]. Nonwear periods were defined as time intervals of at least 60 consecutive minutes of zero counts, with an activity interruption allowance of 0-100 counts∙min^-1 lasting a maximum of 2 consecutive minutes[28]. The aPAL was calculated by dividing the total amount of counts by the total duration of monitored activity. The time spent per day in low intensity activities and the time spent in moderate-to-high intensity activities were calculated using the equation of Sasaki et al.[29] which allows converting Actigraph GT3X + counts into MET (MET = 0.000863∙counts + 0.668876). The same intensity thresholds as those adopted for PASE intensity sub-scores were used. For patients older than 65 years, the threshold between low and moderate-to-high intensity activity was set at 3280 counts∙min^-1. For patients younger than 65 years, the threshold was fixed at 4323 counts∙min^-1. Epochs with less than 150 counts∙min^-1 were classified as inactive and therefore excluded from the analyses[30].

Data normality was evaluated with Shapiro-Wilk tests. Descriptive data are presented as means ± standard deviations (SD), while validity and reproducibility results are presented with the corresponding 95% confidence intervals (CI). Differences between men and women were investigated with unpaired t tests (two-tailed). Reproducibility of the PASE total score was investigated using two-way mixed, single measure intraclass correlation coefficients (ICC_3,1) for reliability, percent standard error of measurements (SEM) and smallest detectable changes (SDC = 1.96 × √2 × SEM) for agreement[31]. Reliability refers to the ability of a measurement instrument to differentiate among patients (despite measurement error), while agreement refers to the precision of the measurement instrument[31]. Acceptable reliability was set at ICC ≥ 0.70 for a sample size of n ≥ 50[12]. In addition, differences between the first and second test sessions were calculated for PASE total score, activity and intensity sub-scores using paired t-tests (two tailed). The construct (criterion) validity of the PASE questionnaire was evaluated using Pearson’s correlation coefficients between PASE total score and aPAL. Low and moderate-to-high intensity activity PASE sub-scores were correlated to the min spent in low and moderate-to-high intensity activities, respectively. Adequate validity was set at r ≥ 0.50 for a sample size of n ≥ 50, according to Terwee et al.[12] for examining associations between physical activity questionnaires and accelerometry. Statistical analyses were conducted with the SPSS version 20 (SPSS, Chicago, IL, USA). The significance level was set at P < 0.05.