How to estimate the population that is covered by community pharmacies? An evaluation of two methods using drug utilisation information.

Purpose: Pharmacy-dispensing data are valuable sources of drug information, but the population that is covered by the pharmacies is often difficult to determine. We evaluated two methods using drug utilisation information to estimate the population size: a drug-use-based extrapolation of a known part of the population and a capture-recapture estimation without any prior knowledge of the population.
Methods: Using pharmacy-dispensing data of three towns with known populations in the Netherlands, we estimated age-and-sex specific population sizes by extrapolating the proportion of drug-using inhabitants. In addition, we applied two-source and three-source capture-recapture models with all combinations of the following drug groups as different sources: anti-asthmatics, analgesics, antibiotics and anti-histamines.
Results: Drug-use-based extrapolation resulted in the best estimates with the least variability. All capture-recapture models provided underestimations of the true population. Three-source capture-recapture resulted in better average estimates than two-source capture-recapture, but also had more variability.
Conclusions: If a part of the population is known, and if there is reason to assume that drug utilisation patterns do not vary within the region, it is best to use drug-use-based extrapolation. In all other situations capture-recapture may be considered, with as main limitation that we found all models to underestimate the population considerably.