Self-reported medication use in health surveys and dispensed prescriptions recorded in national registries capture different aspects of medication use and should not be considered equivalent or interchangeable, according to a Danish study published in BMC Health Services Research. Evaluating data from the Lolland-Falster Health Study, researchers found that while participants reporting no medication use matched prescription records with exceptional consistency, participants reporting active use displayed substantial variation in agreement depending on the specific drug category.
The investigation aimed to elucidate the similarities and discrepancies between self-reported survey responses and official registry filings, as well as the potential implications for epidemiological research. As the authors explained, population-based health surveys and national prescription registries offer vital opportunities for public health investigations, but they measure distinct phenomena: questionnaires record self-reported medication use, whereas the Danish National Prescription Registry documents prescriptions that have actually been dispensed.
To analyze the correspondence between these two data sources, the research team evaluated 14,762 adult participants from the Lolland-Falster Health Study, known as LOFUS. The comprehensive health survey included 299 questions in total, of which 30 specifically addressed medication use. These questions asked respondents whether they used specific medications or broader medication categories on a daily or almost daily basis.
The researchers mapped each of the 30 medication questions to corresponding Anatomical Therapeutic Chemical classification codes. However, questions were excluded from the comparison if they covered over-the-counter medications, overlapped with other Anatomical Therapeutic Chemical categories, or were insufficiently specified. After applying these exclusion criteria, nine of the 30 medication questions were determined to be sufficiently specific for comparison against the Danish National Prescription Registry.
For the nine included questions, self-reported medication use was compared with dispensed prescriptions in the national registry within the preceding 180 days using two-by-two contingency tables. The analysis revealed that the proportion of participants reporting medication use in the survey who also possessed a corresponding dispensed prescription in the registry varied widely across medication categories, ranging from 37.9 percent to 86.3 percent.
The highest levels of correspondence were documented for clearly defined chronic medication categories. Specifically, the highest agreement occurred for medications treating hypothyroidism, where 86.3 percent of participants reporting use had a corresponding prescription in the registry. Other anti-diabetic medications followed closely, with 84.9 percent of self-reported users having a matching dispensed prescription recorded within the 180-day window.
Conversely, the lowest levels of correspondence were observed for other medication classes. Only 40.6 percent of participants reporting the use of medication for alcohol use disorder had a matching prescription in the national prescription registry. The lowest correspondence overall occurred for anti-thyroid medication, which showed an agreement of only 37.9 percent when symptomatic medications were included in the comparison category.
In sharp contrast to reported use, the correspondence between self-reported non-use and registry records was consistently high across all examined categories. The proportion of participants reporting no medication use in the survey who also had no corresponding dispensed prescription in the Danish National Prescription Registry ranged from 90.6 percent to 100 percent, demonstrating strong negative agreement.
Based on these findings, the study authors concluded that survey questions align well with prescription registries for non-use and select chronic medications, but correspondence varies considerably across drug types. Because questionnaires and registries capture fundamentally different aspects of medication use, researchers and decision-makers must exercise caution when drawing conclusions about individuals or populations based on these data sources.
Subject of Research: Concordance between self-reported medication use and prescription registry records
Article Title: A comparison of self-reported and register-based data on medication use: results from the Danish Lolland-Falster Health Study
Article References: Have, S. W., Rasmussen, L. K. T., Labriola, M., St-Martin, G., & Jepsen, R. (2026). A comparison of self-reported and register-based data on medication use: results from the Danish Lolland-Falster Health Study. BMC Health Services Research. https://doi.org/10.1186/s12913-026-16181-8
Image Credits: AI Generated
DOI: 10.1186/s12913-026-16181-8
Keywords: Lolland-Falster Health Study, Danish National Prescription Registry, self-reported medication use, dispensed prescriptions, prescription registries, health surveys, Anatomical Therapeutic Chemical codes, ATC classification, epidemiological research, hypothyroidism medication, anti-diabetic medication, alcohol use disorder medication, anti-thyroid medication, contingency tables, public health research, chronic medication, health services research
News Source: Ophelia Keating. (November 17, 2026). Danish Study Finds Survey and Registry Medication Data Are Not Interchangeable. Scienmag.



