Long-term exercise adherence after public health training in at-risk adults.

OBJECTIVES: Sustainment of healthy exercise behavior is essential in preventing cardiovascular disease and diabetes. Few studies have explored long-term exercise adherence after an exercise referral scheme. The objective of this study was to examine 12-month exercise adherence after an exercise intervention program. METHODS: This was a pragmatic follow-up study in at-risk people performed between June 2012 and January 2014. The main outcome measure was self-reported single-item exercise adherence. Secondary outcomes were change in exercise level, quality of life rated on a visual analog scale and self-rated health. Predictors of long-term exercise adherence were assessed by logistic regression, estimating crude odds-ratios (OR) and 95% confidence intervals (95% CIs) and adjusting for age, gender, education, smoking, moderate and vigorous exercise. RESULTS: In total, 214 adults (mean age 58.8±11.97 years, 71% women) participated in the study and received a 12-week training intervention: 62% had hypertension, 64% dyslipidemia and 15% impaired glucose tolerance. Attrition rate was 84% (n=179). During the 12-month follow-up, 48% (n=85) reported long-term exercise adherence. The main predictors of long-term exercise adherence were participation in sport activities at baseline (adjusted odds-ratio [aOR] 4.22, 95% CI 1.72-10.40), self-rated health (aOR 2.60, 1.00-6.75) and quality of life (aOR 2.39, 1.03-5.54). Long-term non-adherence was associated with low education (<10 years; aOR 3.27, 1.14-9.43) and age<50 years (aOR 3.53, 1.32-9.43). CONCLUSIONS: In this pragmatic study, long-term exercise adherence was associated with participation in sport activities and self-rated health at baseline.

Comparative study of T-amplitude features for fitness monitoring using the ePatch® ECG recorder.

This study investigates ECG features, focusing on T-wave amplitude, from a wearable ECG device as a potential method for fitness monitoring in exercise rehabilitation. An automatic T-peak detection algorithm is presented that uses local baseline detection to overcome baseline drift without the need for preprocessing, and offers adequate performance on data recorded in noisy environments. The algorithm is applied to 24 hour data recordings from two subject groups with different physical activity histories. Results indicate that, while mean heart rate (HR) differs most significantly between the groups, T-amplitude features could be useful depending on the disparities in fitness level, and require further investigation on an individual basis.