Background: Corticosteroids are potent anti-inflammatory and immunosuppressive agents widely prescribed in clinical practice; however, they carry a significant risk of adverse effects if prescribed inappropriately. This study aimed to evaluate the rate of non-compliance with prescribing regulations and clinical guidelines in outpatient service prescriptions containing corticosteroids at the Outpatient Department of Military Hospital 175 in 2025, and to identify associated factors.
Methods: A descriptive cross-sectional study of 315 fee-for-service outpatient prescriptions containing corticosteroids, written by 69 physicians (July-August 2025). Prescriptions were assessed against two separate criteria sets: administrative rules (13 criteria of Circular 26/2025/TT-BYT) and clinical standards (benchmarked against the Vietnamese National Drug Formulary 2022 and summaries of product characteristics), each by two independent reviewers. Because prescriptions cluster within physicians, multivariable analysis used Firth penalised logistic regression with physician-level cluster-robust standard errors and three pre-specified covariates.
Results: Administrative non-compliance was 8.3% (26/315; 95% CI: 5.5-11.9) and clinical non-compliance 6.7% (21/315; 95% CI: 4.2-10.0); 1.9% failed both. The most frequent error was contraindication (5.1%), followed by missing dosing time, missing “0” before single-digit quantities and inappropriate indication (3.2% each); the 21 clinically non-compliant prescriptions came from only 8 physicians. For administrative criteria, ophthalmic route (aOR = 4.85; 95% CI: 1.74-13.56) and physician age ≥ 40 years (aOR = 4.76; 95% CI: 1.58-14.31) were associated with error; for clinical criteria only female physician gender was (aOR = 0.14; 95% CI: 0.03-0.73), with low stability.
Conclusion: Errors concentrated in ophthalmic prescriptions and among physicians aged ≥ 40 years for administrative rules, whereas associations with clinical criteria were unstable and need further confirmation. Audit should target error type and drug group rather than prescriber characteristics.