The study (PMID: 41057301), published in Nature Communications on October 7, 2025, analyzed accelerometer data from 73,485 UK Biobank participants (mean age ~62 years, followed for ~8 years). Researchers used machine learning to classify physical activity into light (LPA), moderate (MPA), and vigorous (VPA) intensities. They examined dose-response relationships with outcomes including all-cause mortality, cardiovascular disease (CVD) mortality and events, type 2 diabetes, and cancer incidence/mortality.
Vigorous and moderate activity showed strong, near-linear inverse associations with most outcomes, while light activity had much weaker or limited benefits.
Traditional guidelines (based on self-reported data) equate 1 minute of VPA to 2 minutes of MPA, but device-measured data reveals VPA is substantially more time-efficient.
For equivalent risk reductions (5–35%), the median minutes of MPA needed per 1 minute of VPA were:
4.1 for all-cause mortality
7.8 for CVD mortality
5.4 for major CVD events
9.4 for type 2 diabetes
~3.5 for cancer mortality (lower for cancer outcomes overall)
Across non-cancer outcomes, ~6.6 minutes of MPA per 1 minute of VPA.
LPA required far more time: 53–94 minutes per 1 minute of VPA for non-cancer outcomes (median 79 minutes), and even higher for cancer (156 minutes for mortality), reflecting its weaker protective effects.
These results highlight that accelerometer-based measurements reduce biases in self-reported data, showing VPA provides greater health benefits per minute than previously estimated. The authors suggest updating physical activity guidelines and wearable device algorithms (e.g., fitness trackers) to reflect these device-based equivalences, emphasizing VPA's efficiency for time-constrained individuals while noting LPA's limited substitutability for higher intensities.