2026-10-07T12:40:35Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33980896 2026-09-24T09:25:22Z category_37103 category_33935 category_37043 category_33752 category_31226 portal_2 item_type_3 month_year_09_2026

Supplementary information files for "Heterogeneity in type 2 diabetes onset across combinations of lifestyle behaviours in the UK Biobank" Ansuman Swain (17084815) Natalie Pearson (1255488) Scott A Willis (25103196) George Leckie (742648) Axel Finke (19442869) Joseph Hutchinson (19839522) Will Johnson (2797078) Clinical and health psychology Public health Psychology Health sciences Clinical sciences Endocrinology & Metabolism Incident diabetes MAIHDA Lifestyle behaviours Bayesian survival analysis UK Biobank <p dir="ltr">Supplementary information files for "Heterogeneity in type 2 diabetes onset across combinations of lifestyle behaviours in the UK Biobank"<br><br><b>Background: </b>To examine how combinations of key lifestyle behaviors influence type 2 diabetes (T2D) onset using multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA).<br><br><b>Methods:</b> We analysed 259,448 UK Biobank adults with up to 16 years of follow-up. Fruit and vegetable intake, physical activity, sleep duration, and alcohol consumption were dichotomised by meeting public health recommendations, while smoking was categorised as never, previous, or current. Cross-classification generated 48 lifestyle strata. Sex-stratified MAIHDA models were fitted with individuals nested within strata: Model 1 (no predictors), Model 2 (lifestyle behaviors), Model 3 (lifestyle plus confounders). Posterior summaries estimated variance partitioning coefficients (VPC), proportional change in variance (PCV), and strata-specific median T2D-free survival times.</p><p><br></p><p dir="ltr"><b>Results</b>: In Model over 7% (VPCs) of the variation in T2D onset on the log-hazard scale was attributable to strata differences. In both sexes, additive effects of individual behaviors explained around 93% (PCVs) of between-strata variance. Strata meeting most recommendations had the longest predicted survival, whereas those meeting few had the shortest. Never smoking and meeting physical activity guidelines consistently characterised longer survival.</p><p dir="ltr"><b>Conclusions:</b> Variation in T2D onset is driven primarily by additive lifestyle effects rather than interactions. Each additional recommendation met progressively extends T2D-free survival.<br><br>© The Author(s), CC BY 4.0</p> 2026-07-13T00:00:00Z Dataset Dataset 10.17028/rd.lboro.33980896.v1 https://figshare.com/articles/dataset/Supplementary_information_files_for_Heterogeneity_in_type_2_diabetes_onset_across_combinations_of_lifestyle_behaviours_in_the_UK_Biobank_/33980896 CC BY 4.0