Self-efficacy and physical activity in overweight and obese adults participating in a worksite weight loss intervention: Multistate modeling of wearable device data

Michael C. Robertson, Charles E. Green, Yue Liao, Casey P. Durand, Karen M. Basen-Engquist

Research output: Contribution to journalArticlepeer-review

2 Scopus citations

Abstract

Background: Physical activity is associated with a reduced risk of numerous types of cancer and plays an important role in maintaining a healthy weight. Wearable physical activity trackers may supplement behavioral intervention and enable researchers to study how determinants like self-efficacy predict physical activity patterns over time. Methods: We used multistate models to evaluate how selfefficacy predicted physical activity states among overweight and obese individuals participating in a 26-week weight loss program (N = 96). We specified five states to capture physical activity patterns: (i) active (i.e., meeting recommendations for 2 weeks), (ii) insufficiently active, (iii) nonvalid wear, (iv) favorable transition (i.e., improvement in physical activity over 2 weeks), and (v) unfavorable transition.We calculated HRs of transition probabilities by self-efficacy, body mass index, age, and time. Results: The average prevalence of individuals in the active, insufficiently active, and nonvalid wear states was 13%, 44%, and 16%, respectively. Low self-efficacy negatively predicted entering an active state [HR, 0.51; 95% confidence interval (CI), 0.29- 0.88]. Obesity negatively predicted making a favorable transition out of an insufficiently active state (HR, 0.61; 95% CI, 0.40-0.91). Older participants were less likely to transition to the nonvalid wear state (HR, 0.53; 95% CI, 0.30-0.93). Device nonwear increased in the second half of the intervention (HR, 1.73; 95% CI, 1.07-2.81). Conclusions: Self-efficacy is an important predictor for clinically relevant physical activity change in overweight and obese individuals. Multistate modeling is useful for analyzing longitudinal physical activity data. Impact: Multistate modeling can be used for statistical inference of covariates and allow for explicit modeling of nonvalid wear.

Original languageEnglish (US)
Pages (from-to)769-776
Number of pages8
JournalCancer Epidemiology Biomarkers and Prevention
Volume29
Issue number4
DOIs
StatePublished - 2020
Externally publishedYes

ASJC Scopus subject areas

  • Epidemiology
  • Oncology

Fingerprint Dive into the research topics of 'Self-efficacy and physical activity in overweight and obese adults participating in a worksite weight loss intervention: Multistate modeling of wearable device data'. Together they form a unique fingerprint.

Cite this