Electronic prescribing data for prescribing-quality surveillance among high-prescribing general practitioners: A cross-sectional study in Iran
Nezam Armand, Hojjat Rahmani, Ali Akbar Fazaeli, Mohammad Effatpanah, Leila GhamkharObjective
To examine associations between prescriber characteristics and two World Health Organization (WHO) prescribing indicators among high-prescribing general practitioners (GPs) in Iran using routinely collected electronic prescribing data.
Methods
This retrospective cross-sectional study analyzed outpatient electronic prescribing data from the Iran Health Insurance Organization in Tehran Province, October 2024–March 2025. The analysis included 645 GPs with at least 30 prescriptions, an average number of medicines per prescription above the provincial mean, and more than 20% of prescriptions containing antibiotics. Weighted least squares regression modeled medicines per prescription, using prescription counts as weights. Antibiotic-containing prescription counts were modeled using negative binomial regression with the natural logarithm of total prescription volume as an offset. Models included sex, professional experience, and university zone; age was excluded because of collinearity.
Results
Physician-level means were 5.64 medicines per prescription and 51.06% antibiotic-containing prescriptions. Male sex was associated with more medicines per prescription (B=0.49; 95% CI: 0.245–0.735; p<0.001) and a higher antibiotic-prescribing rate (IRR=1.068; 95% CI: 1.006–1.133; p=0.030). Each additional year of experience was associated with fewer medicines per prescription (B=−0.027; p<0.001) and a higher antibiotic-prescribing rate (IRR=1.003; p=0.012). University zone showed no independent association. The medicines model explained little variance (R 2 =0.065).
Conclusion
Sex and experience showed modest associations within this selected subgroup. Without patient-level clinical information, causal or behavioral interpretations are not justified. These indicators do not establish clinical inappropriateness. Findings should not be extrapolated to the wider GP population or interpreted as annual prescribing estimates.