Data transparency matters in healthcare research, and a new UK Biobank analysis provides a sobering reminder of how methodology shapes our understanding of disease burden. Among 474,397 UK adults aged 40-69, multimorbidity prevalence ranged from 1.0% using clustering approaches to 35.3% using count-based methods with extended condition lists—a staggering 35-fold difference in the same cohort.
The research team examined count-based definitions alongside clustering analysis to understand why estimates diverged so dramatically. Count-based approaches using comprehensive condition lists identified the highest prevalence at 35.3%, while the most prevalent condition approach yielded 23.1%, body systems clustering generated 12.4%, and pure clustering analysis produced just 1.0%. These disparities have profound implications for epidemiological estimates, treatment guidelines, and healthcare resource planning across institutions and countries that may unknowingly adopt incompatible definitions.
Read the full article on GMJ Newsroom.
Was this article helpful?
