Temporal changes in a novel metric of physical activity tracking (personal activity intelligence) and mortality: the HUNT study, Norway

Background: Personal Activity Intelligence (PAI) is a novel activity metric that translates heart rate variations during exercise into a weekly score. Weekly PAI scores assessed at a single point in time were found to associate with lower risk of premature cardiovascular disease (CVD) mortality in the general healthy population. However, to date, the associations between long-term longitudinal changes in weekly PAI scores and mortality have not been explored. Purpose: The aim of the present study was to prospectively examine the association between change in weekly PAI scores estimated 10 years apart, and risk of mortality from CVD and all-causes. Methods: We performed a prospective cohort study of 11,870 men and 13,010 women without known CVD in Norway. By using data from the Nord-Trøndelag Health Study (HUNT), PAI was estimated twice, ten years apart (HUNT1 1984-86 and HUNT2 1995-97). Mortality was followed-up until December 31, 2015. Adjusted hazard ratios (AHR) and 95% confidence intervals (CI) for death from CVD and all-causes related to temporal changes in PAI were estimated using Cox regression analyses. Results: During a mean (SD) of 18 (4) years of follow-up, there were 4782 deaths, including 1560 deaths caused by CVD. Multi-adjusted analyses demonstrated that participants achieving a score of ≥100 PAI at both time points had 32% lower risk of CVD mortality (AHR 0.68; CI: 0.54–0.86) for CVD mortality and 20% lower risk of all-cause mortality (AHR 0.80; CI: 71–0.91) compared with participants obtainin