Decoding Traditional Health Patterns with Big Data

Scientists created a tool that turns routine health check results into a digital version of Traditional Chinese Medicine constitution. They applied it to a large group of older adults tracked over many years.

The dataset included 47,417 health exams from 11,355 participants. The study covered the years 2017 through 2026. Researchers used the first exam for each person as a baseline and paired yearly exams for follow‑up, giving 32,648 adjacent visit pairs. The system linked illnesses to constitution types, built combined risk scores from several markers, predicted next‑year outcomes, followed how constitutions shifted, and tested how well routine data could label a constitution.

To make the results robust, the team ran many sensitivity checks. They examined different time splits, used body mass index alone or with waist size, removed height and weight factors, adjusted for overall fat, and looked at both new and lasting constitution patterns. The phlegm‑dampness profile emerged as the most distinct, showing higher cardiometabolic marker levels than a balanced constitution.

The work demonstrates that everyday health records can be repurposed into clear, computable health phenotypes. This could help doctors spot subtle risk patterns that standard lab tests often miss.