Improving Assessment of Lifetime Solar Ultraviolet Radiation Exposure in Epidemiologic Studies: Comparison of Ultraviolet Exposure Assessment Methods in a Nationwide U.S. Occupational Cohort

Solar ultraviolet radiation is the primary risk factor for skin cancers and sun‐related eye disorders. Estimates of individual ambient ultraviolet irradiance derived from ground‐based solar measurements and from satellite measurements have rarely been compared. Using self‐reported residential history from 67 189 persons in a nationwide occupational US radiologic technologists’ cohort, we estimated ambient solar irradiance using data from ground‐based meters and noontime satellite measurements. The mean distance moved from city of longest residence in childhood increased from 137.6 km at ages 13–19 to 870.3 km at ages ≥65, with corresponding increases in absolute latitude difference moved. At ages 20/40/60/80, the Pearson/Spearman correlation coefficients of ground‐based and satellite‐derived potential solar ultraviolet exposure, using irradiance and cumulative radiant exposure metrics, were high (=0.87–0.92). There was also moderate correlation (Pearson/Spearman correlation coefficients = 0.51–0.60) between irradiance at birth and at last‐known address, for ground‐based and satellite data. Satellite‐based lifetime estimates of ultraviolet radiation were generally 14–15% lower than ground‐based estimates, albeit with substantial uncertainties, possibly because ground‐based estimates incorporate fluctuations in cloud and ozone, which are incompletely incorporated in the single noontime satellite‐overpass ultraviolet value. If confirmed elsewhere, the findings suggest that ground‐based estimates may improve exposure assessment accuracy and potentially provide new insights into ultraviolet radiation–disease relationships in epidemiologic studies