A Smartphone App to Improve Physical Activity

Step counts (objective physical activity) were measured using the activPAL thigh-worn accelerometer during a 72-hour monitoring period.

Estimated average change is computed as the change in activPAL-tracked mean daily steps at Month 4 relative to baseline; this outcome is analyzed in a linear regression analysis model. All randomized participants who started the intervention and had valid activity monitor data were utilized in the model (ITT analysis)

Objective sedentary activity time was measured using the activPAL thigh-worn accelerometer during a 72-hour monitoring period. Expressed as average minutes/day.

Self-reported physical activity will be measured via the Physical Activity Scale for the Elderly (PASE). PASE is a ten-item instrument designed to assess engagement in physical activities commonly pursued by older adults, including those related to leisure, household, and occupational tasks. The tool is a valid and reliable measure of physical activity engagement in the older adult population. Scores range from 0 to 361. Higher scores indicate a higher level of activity.

Functional mobility will be assessed through a four-meter walk test, a commonly used, validated measure of physical and functional performance in older adults.

The potential mediators of physical activity self-efficacy, self-regulation, outcome expectation, and social support will each be measured separately using subscales of the 78-item physical activity portion of the Health Beliefs Survey. Subscales demonstrate sufficient internal consistencies (Cronbach's α=0.68-0.90) and are predictive of physical activity. Higher scores indicate greater self-efficacy. Possible scores for the self-efficacy subscale range from 0-100.

The potential mediators of physical activity self-efficacy, self-regulation, outcome expectation, and social support will each be measured separately using subscales of the 78-item physical activity portion of the Health Beliefs Survey. Subscales demonstrate sufficient internal consistencies (Cronbach's α=0.68-0.90) and are predictive of physical activity. Higher scores indicate greater self-regulation behaviors. Possible scores for the self-regulation subscale range from 1-5.

The potential mediators of physical activity self-efficacy, self-regulation, outcome expectation, and social support will each be measured separately using subscales of the 78-item physical activity portion of the Health Beliefs Survey. Subscales demonstrate sufficient internal consistencies (Cronbach's α=0.68-0.90) and are predictive of physical activity. Higher scores indicate greater family social support for physical activity. Possible scores for the family social support subscale range from 1 - 5.

The potential mediators of physical activity self-efficacy, self-regulation, outcome expectation, and social support will each be measured separately using subscales of the 78-item physical activity portion of the Health Beliefs Survey. Subscales demonstrate sufficient internal consistencies (Cronbach's α=0.68-0.90) and are predictive of physical activity. Higher scores indicate greater outcome expectations. Possible scores for the outcome expectation subscale range from 1-25

Aging self-perceptions will be assessed by the Attitudes Toward Own Aging subscale of the Philadelphia Geriatrics Center Morale Scale. This five-question tool captures the subjective aging experience, shows moderate internal consistency (Cronbach's α=0.61-0.64), and predicts mortality risk. Scores can range from 0 to 5. A higher score indicates more positive aging self-perceptions.

Views of aging will be measured using the Attitudes to Ageing Questionnaire. This 24-item assessment identifies subjective views about age-related changes in multiple domains, is cross-culturally valid, and is psychometrically sound (Cronbach's α=0.68-0.84). Its subscales include psychosocial loss, physical change, and psychological growth. A higher score for psychosocial loss indicates more negative attitude (min 8; max 40).

Views of aging will be measured using the Attitudes to Ageing Questionnaire. This 24-item assessment identifies subjective views about age-related changes in multiple domains, is cross-culturally valid, and is psychometrically sound (Cronbach's α=0.68-0.84). Its subscales include psychosocial loss, physical change, and psychological growth. A higher score on physical change indicates more positive attitude (min 8; max 40)

Views of aging will be measured using the Attitudes to Ageing Questionnaire. This 24-item assessment identifies subjective views about age-related changes in multiple domains, is cross-culturally valid, and is psychometrically sound (Cronbach's α=0.68-0.84). Its subscales include psychosocial loss, physical change, and psychological growth. A higher score on psychological growth indicates more positive attitude (min 8; max 40).

Usage behavior is defined as the proportion of days the app was opened across the 4-month trial period

The Mobile App Rating Scale User Version (uMARS) was used to measure user satisfaction with the app, particularly participant's ratings for app quality. The uMARS includes a usability feedback subindex which is comprised of the average of item responses for 4 subsections (engagement, functionality, aesthetics, information), to yield a total quality score. Ratings are on a scale of 1 (low perceived quality, min) to 5 (high perceived quality, max).

Spearman Correlation coefficients were calculated to assess correlations between change in the primary outcome (change in daily steps) and change in hypothesized mediators (i.e., Self-efficacy for Physical Activity Change, Self-regulation of Physical Activity Change, Family Social Support for Physical Activity Change, Outcome Expectation for Physical Activity Change, Aging Self-perceptions (Attitude Toward Own Aging) Change, Views of Aging--Psychosocial Loss Change, Views of Aging--Physical Change Change, Views of Aging--Psychological Growth Change).