Personalized Activity Intelligence (PAI) for Prevention of Cardiovascular Disease and Promotion of Physical Activity

Abstract

Purpose: To derive and validate a single metric of activity tracking that associates with lower risk of cardiovascular disease mortality.

Methods: We derived an algorithm, Personalized Activity Intelligence (PAI), using the HUNT Fitness Study (n = 4631), and validated it in the general HUNT population (n = 39,298) aged 20-74 years. The PAI was divided into three sex-specific groups (≤50, 51-99, and ≥100), and the inactive group (0 PAI) was used as the referent. Hazard ratios for all-cause and cardiovascular disease mortality were estimated using Cox proportional hazard regressions.

Results: After >1 million person-years of observations during a mean follow-up time of 26.2 (SD 5.9) years, there were 10,062 deaths, including 3867 deaths (2207 men and 1660 women) from cardiovascular disease. Men and women with a PAI level ≥100 had 17% (95% confidence interval [CI], 7%-27%) and 23% (95% CI, 4%-38%) reduced risk of cardiovascular disease mortality, respectively, compared with the inactive groups. Obtaining ≥100 PAI was associated with significantly lower risk for cardiovascular disease mortality in all prespecified age groups, and in participants with known cardiovascular disease risk factors (all P-trends <.01). Participants who did not obtain ≥100 PAI had increased risk of dying regardless of meeting the physical activity recommendations.

Conclusion: PAI may have a huge potential to motivate people to become and stay physically active, as it is an easily understandable and scientifically proven metric that could inform potential users of how much physical activity is needed to reduce the risk of premature cardiovascular disease death.

Keywords: Activity tracking; Cardiovascular disease mortality; Physical activity; Prevention.