Enhancing Self-Reported Physical Activity and Related Self-Efficacy Among Adolescents

Abstract

Physical activity (PA) declines significantly during adolescence. PA refers to individual’s level of bodily movement and performance of activities requiring physical actions. Self-efficacy (SE) is considered a determinant of PA behavior. PA-SE is defined as one’s belief in his/her capabilities to perform PA to achieve certain outcomes and capabilities in managing positive and negative experiences to participate in PA. Promoting PA levels in adolescents requires accurate assessment tools of PA and related SE. The existing self-regulatory PA-SE measures focus on measuring confidence against barriers to PA, and instruments have been validated for use in either younger or older adolescents. Adequate validity of instruments that demonstrates the relationship with PA has been established for older adolescents using a SE instrument, called the SE to Overcome Barriers to PA Scale. The International Classification of Functioning Disability and Health (ICF) emphasizes addressing both facilitators and barriers of contextual factors influencing health and health-related behaviors. Therefore, expanding measurement of PA-SE construct by including beliefs in doing PA with existence of perceived facilitators in addition to overcoming barriers based on perspective of both younger and older adolescents is needed to capture more possible variation in PA and changes in the relationship by age. PA should also be measured using a psychometrically robust PA instrument when examining validity of a PA-SE instrument. The objective of this dissertation work was to enhance validity of self-reported PA and PA-SE in younger and older adolescents and better understand the perceived regulatory SE related to PA in this population. The specific aims for our studies were: 1) To ensure content validity of some existing PA-SE measures’ items (items for SR to overcome barriers) in addition to a new SE-related facilitators items for younger and older adolescents; 2) To examine convergent validity of the 7-day recall Patient Reported Outcome Measurement Information System pediatric measure of PA (PROMIS-PA), and to explore whether the ecological momentary assessment (EMA)- and/or end of day (EoD) recall- versions of PROMIS-PA are more representative of daily PA activity than the 7-day recall version due to reduced recall bias expected with using these methods; and 3) To assess the construct validity, including convergent validity, of the new PA regulatory SE (PARSE) instrument (the resultant item pools from aim 1). In the first study, we conducted individual sessions of semi-structured and cognitive debriefing interviews with adolescents (n = 18, aged 11-17) to further explore the PA-SE related facilitators and barriers concepts and revise item pools. Participants completed 3 questionnaires for the cognitive interviewing: SE to Overcome Barriers to PA Scale, Perceived PA-SE Scale, and new SE-related facilitators items. Thematic analysis showed that personal and environmental facilitators and barriers all influence confidence of being physically active. The major categories of our PA-SE questionnaire were physical; psychological; interaction with surroundings; support and relationships; attitudes; and services, systems, and natural environments. The cognitive interviews resulted in 52 final items: 24 for SE with perceived facilitators and 28 for SE to overcome barriers. This study supports content validity including clarity, relevance, completeness of the resultant PA-SE instrument for younger and older adolescents. For our second study, participants were asked to complete daily PROMIS-PA-based EMA surveys (5 scheduled EMAs/day) and EoD surveys through an application for 7 consecutive days and wear accelerometry devices on these days. On the seventh day, they completed the 7-day recall PROMIS-PA and Youth Activity Profile (YAP). In our total sample (n = 84, aged 10-20), the 7-day recall PROMIS-PA showed significant positive moderate to high correlations with the YAP scores and the average of daily steps. Our further analyses in a subsample of participants (n = 25, aged 11-18) indicated that the 7-day recall version has positive high correlations with the averaged EMA or EoD ratings for the week. The associations of the average of daily steps with the averaged EMA and EoD ratings for the week were comparable to its association with the 7-day recall PROMIS-PA scores. These findings support the validity of the 7-day recall PROMIS-PA by showing that the PROMIS-PA, YAP, and daily steps measure related PA constructs, and it has a good representation of PA as EMA and EoD reporting. Our findings suggest that adolescents demonstrate higher compliance to the EoD and 7-day recall PROMIS-PA which summarize the EMA reporting or the daily steps comparably. For our third study, the total sample of adolescents (younger adolescents n = 51, aged 10-14; older adolescents n = 33, aged 15-20), who participated in the second study, completed also both PARSE sections (PA-SE to overcome barriers (PARSE-OB) and PA-SE with facilitators (PARSE-F)) before the week of activity tracking. PARSE-OB showed significant moderate and low positive linear relationships with the RPOMIS-PA and step average, respectively. PARSE-OB and PARSE-F scores were highly correlated. The PARSE-OB validity was supported also by demonstrating a significant moderation effect of gender on the relationship between the PA-SE and PA in younger adolescence and a significant positive linear relationship between these variables, with no gender moderation effect, in older adolescents. Our results did not support the PARSE-F and PA relationship. These findings indicate that both sections of the PARSE instrument measure similar constructs, and the PARSE-OB section measure a construct that is directly related to PA in younger and older adolescents. The PARSE-OB and PROMIS-PA are valid for use in younger and older adolescents. The use of PARSE-OB to assess PA-SE provides an improved understanding of the effect of SE on PA and can inform the design of more effective PA behavioral intervention strategies for younger and older adolescents. The use of the PARSE-F warrants further investigation to its potential indirect effect on PA since it is related to PARSE-OB. Using the PARSE-F items with PARSE-OB can help in planning behavioral intervention strategies since adolescents’ experiences, based on our qualitative interviews, indicate the importance of perceived SE around the facilitators for this population. The 7-day recall and EoD recall PROMIS-PA can provide sufficient representation of daily PA, and they are preferred self-reporting methods for adolescents when compared to EMA.