Open Access Highly Accessed Research article

Can we use the pharmacy data to estimate the prevalence of chronic conditions? a comparison of multiple data sources

Francesco Chini*, Patrizio Pezzotti, Letizia Orzella, Piero Borgia and Gabriella Guasticchi

Author Affiliations

Agency of Public Health, Lazio Region; via di S. Costanza 53, 00198 Rome, Italy

For all author emails, please log on.

BMC Public Health 2011, 11:688  doi:10.1186/1471-2458-11-688

Published: 5 September 2011



The estimate of the prevalence of the most common chronic conditions (CCs) is calculated using direct methods such as prevalence surveys but also indirect methods using health administrative databases.

The aim of this study is to provide estimates prevalence of CCs in Lazio region of Italy (including Rome), using the drug prescription's database and to compare these estimates with those obtained using other health administrative databases.


Prevalence of CCs was estimated using pharmacy data (PD) using the Anathomical Therapeutic Chemical Classification System (ATC).

Prevalences estimate were compared with those estimated by hospital information system (HIS) using list of ICD9-CM diagnosis coding, registry of exempt patients from health care cost for pathology (REP) and national health survey performed by the Italian bureau of census (ISTAT).


From the PD we identified 20 CCs. About one fourth of the population received a drug for treating a cardiovascular disease, 9% for treating a rheumatologic conditions.

The estimated prevalences using the PD were usually higher that those obtained with one of the other sources. Regarding the comparison with the ISTAT survey there was a good agreement for cardiovascular disease, diabetes and thyroid disorder whereas for rheumatologic conditions, chronic respiratory illnesses, migraine and Alzheimer's disease, the prevalence estimates were lower than those estimated by ISTAT survey. Estimates of prevalences derived by the HIS and by the REP were usually lower than those of the PD (but malignancies, chronic renal diseases).


Our study showed that PD can be used to provide reliable prevalence estimates of several CCs in the general population.