NHANES Multi-Cycle Diabetes and Glycemic Status Dataset, 2011-2023

A harmonized participant-level dataset derived from five NHANES survey cycles: 2011-2012, 2013-2014, 2015-2016, 2017-2020, and 2021-2023. The published file, df_with_weights.csv, contains 55,405 participants and 307 columns covering demographics, diabetes questionnaire variables, laboratory biomarkers, dietary intake, lifestyle, physical activity, comorbidities, medication-derived indicators, and retained NHANES survey design/weight variables.

The target variable is DIQ010, recoded as a three-class glycemic-status label: 0 = no diabetes, 1 = prediabetes, and 2 = diabetes. Pediatric participants are included, and users should account for this when studying adult type 2 diabetes. Survey weights are retained but not applied automatically; users are responsible for survey-aware analyses and population inference. No global imputation was applied.