Answers are grounded in this paper and cite their sources.
Population-based health surveys and national registries offer important opportunities for public health research but may capture different information. Medication use represents a pertinent example; the Danish prescription registry captures dispensed prescriptions, whereas questionnaires capture self-reported use. This study investigated self-reported medication use in the Lolland-Falster Health Study (LOFUS) and dispensed prescriptions in the Danish National Prescription Registry (DNPR), with the aim of elucidating the similarities and discrepancies between the two data sources and potential implications for epidemiological research. This study included 14,762 adult LOFUS participants. Of 299 questions in LOFUS, 30 addressed medication use and asked whether participants used specific medications or medication categories “daily or almost daily”. Each question was mapped to relevant Anatomical Therapeutic Chemical (ATC) codes. Questions were excluded if they covered over-the-counter medications, overlapped with other ATC categories, or were insufficiently specified. For included questions, self-reported medication use was compared with dispensed prescriptions in DNPR within the preceding 180 days using 2 × 2 contingency tables. Nine of the 30 medication questions were considered sufficiently specific for comparison with DNPR through ATC codes. The proportion of participants reporting medication use in LOFUS, who also had a corresponding prescription in DNPR, varied across categories, ranging from 37.9% to 86.3%. The highest proportions were observed for medication for hypothyroidism (86.3%) and other anti-diabetic medication (84.9%), while the lowest were observed for medication for alcohol use disorder (40.6%) and anti-thyroid medication (when including symptomatic medication) (37.9%). In contrast, the proportion of participants reporting no medication use and having no corresponding prescription in DNPR was consistently high, ranging from 90.6% to 100%. LOFUS medication questions corresponded well with prescription registry data among participants reporting non-use and for some clearly defined chronic medication categories, but the degree of correspondence varied across medication types. Self-reported medication use and prescription registry data capture different aspects of medication use and should not be considered equivalent or interchangeable. Therefore, researchers and decision-makers should remain cautious when drawing conclusions about the individuals or population groups on which their studies are based.
No takes yet. Share an insight, caveat, or question.
Have et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: