Antibiotic prescribing trends among Iranian GPs during COVID-19: a longitudinal analysis of antimicrobial resistance risks
- Open Access
- 01.12.2025
- Research
Erschienen in:
Abstract
Introduction
Antimicrobial resistance (AMR) is a mounting global health crisis, driven largely by the inappropriate and excessive use of antibiotics across healthcare settings. The World Health Organization (WHO) has repeatedly warned that unchecked AMR could undermine decades of progress in infectious disease control, leading to increased morbidity, mortality, and healthcare costs worldwide. Iran, like many countries, entered the COVID-19 pandemic already facing high rates of antibiotic consumption and rising resistance among key bacterial pathogens, including Streptococcus pneumoniae, Escherichia coli, and Klebsiella pneumoniae [1, 2].
The emergence of COVID-19 profoundly disrupted clinical practice and antibiotic stewardship globally. In Iran, both hospital-based and community-based studies have documented a sharp increase in antibiotic use during the pandemic, often in the absence of confirmed bacterial co-infection. Salehi et al. (2022) found that, among 43,791 hospitalized COVID-19 patients across 12 major centers, 121.6 defined daily doses (DDD) of antibiotics were used per 100 hospital bed-days, despite bacterial co-infection being detected in only 14.4% of cases. Alarmingly, higher antibiotic use correlated with increased mortality. Similarly, Raoofi et al. reported a significant rise in resistance rates among Gram-negative bacteria, especially Pseudomonas aeruginosa and Klebsiella pneumoniae, during the pandemic, further narrowing treatment options and endangering public health [3].
Anzeige
Beyond the hospital setting, the pandemic also fueled self-medication and unsupervised antibiotic use. Faraji et al. (2022) showed that nearly 60% of COVID-19 outpatients in western Iran practiced self-medication, with easy access to antibiotics and fear of COVID-19 as key drivers. Economic barriers to accessing medical care further exacerbated this trend, particularly among women and those with lower socioeconomic status [4]. Khoshbakht et al. (2023) similarly highlighted widespread self-medication and empiric antibiotic use before hospitalization among COVID-19 patients, raising concerns about the unchecked spread of AMR in the community [5].
The overuse of antibiotics during the pandemic has had measurable consequences for resistance patterns. Khodashahi et al. (2022) documented increased minimum inhibitory concentrations (MICs) for several antibiotics among bacterial and fungal isolates from COVID-19 patients, with high rates of methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant Enterococcus (VRE), and extensively drug-resistant Acinetobacter baumannii [6]. These findings are corroborated by environmental surveillance data, which show increased antibiotic residues in wastewater during COVID-19 peaks, indirectly reflecting clinical overprescription.
Primary care settings, where general practitioners (GPs) and family physicians provide the majority of outpatient care, are particularly important in shaping antibiotic use patterns. Karimi et al. (2023) found that 57% of prescriptions by family physicians in Alborz province contained at least one antibiotic, with amoxicillin being the most commonly prescribed. Notably, nearly 60% of these prescriptions did not meet scientific criteria for rational use, and factors such as physician experience, patient demographics, and seasonality influenced prescribing behavior [7]. Sami et al. (2022) further revealed that Iranian physicians’ knowledge of AMR and stewardship guidelines remains limited, with less than half regularly consulting microbiology laboratories or established protocols before prescribing antibiotics [8].
Despite these alarming trends, no study has systematically examined longitudinal antibiotic prescribing patterns among Iranian GPs across the pre-pandemic, pandemic, and post-pandemic periods. Given the central role of GPs in outpatient antibiotic use, understanding their prescribing behaviors is critical for designing effective stewardship interventions and policy responses tailored to Iran’s unique healthcare landscape [9].
Anzeige
The present study addresses this crucial knowledge gap by analyzing over 1.4 million prescriptions from 73 GPs in a major Iranian urban clinic, spanning three distinct phases: pre-COVID, COVID, and post-COVID. By characterizing temporal shifts in antibiotic classes, prescribing intensity, and the impact of the pandemic on GP practice, our findings aim to provide actionable evidence to guide stewardship programs and inform strategies to contain AMR in Iran’s evolving healthcare system.
Methods
Study design and data source
This retrospective cohort study utilized anonymized prescription data extracted from Iran’s National Hospital Information System (HIS). Access to the anonymized data was granted by the Ethics Committee of Zanjan University of Medical Sciences (approval code: ZUMS.REC.1394.322) as part of the study approval process, which waived the requirement for individual informed consent due to the retrospective and anonymized nature of the data. The dataset comprised records from a single government-run central public clinic serving a diverse urban and peri-urban patient population. Due to budgetary constraints and procurement policies, the clinic’s antibiotic formulary is predominantly limited to older, low-cost agents, resulting in underrepresentation of newer or high-cost antibiotics. Rural and private healthcare facilities were not included, a decision based on the study’s aim to focus on a representative public primary care setting and the logistical challenge of standardizing data from diverse private and rural sources with potentially different prescribing practices and formulary structures. According to national healthcare utilization data, this clinic accounts for approximately 7% of total GP prescriptions nationwide [10]. Prescriptions from specialties other than general practice (e.g., obstetrics, dentistry, pediatrics) were excluded to focus specifically on GP prescribing patterns.
Study periods
To control for seasonal variation in prescribing, three distinct study periods were defined with seasonally matched durations (12 months each, except the COVID-19 period, which lasted 16 months):
-
PreCOVID: January 2019 – December 2019 (baseline pre-pandemic period).
-
COVID: January 2020 – May 2021 (peak pandemic period, including initial confirmed cases and lockdowns).
-
PostCOVID: June 2021 – May 2022 (post-pandemic recovery phase following relaxation of restrictions).
The COVID period includes January and February 2020 to capture early prescribing anomalies such as antibiotic stockpiling. Sensitivity analyses confirmed that observed trends remained statistically significant after adjusting for differences in period length (p < 0.01) [11].
Variables and definitions
Active General Practitioners (GPs): Defined as physicians issuing at least 100 prescriptions per study period to ensure consistent clinical engagement. This threshold of 100 prescriptions was chosen to exclude physicians with minimal clinical activity (e.g., those on leave, in administrative roles, or with very low patient loads), thereby focusing the analysis on GPs with substantial and consistent patient-facing responsibilities, which provides a more accurate reflection of active prescribing behavior. The number of active GPs varied across periods due to workforce fluctuations, including redeployments during the pandemic [12].
Antibiotic items
Number of antibiotic units prescribed, categorized by antibiotic class and formulation (oral or injectable).
Total prescriptions
All medication prescriptions issued by active GPs during each period.
Antibiotic-to-Total drug ratio
The ratio of antibiotic items to total drug items prescribed, providing a normalized measure of antibiotic utilization. National data indicate that approximately 36–40% of prescriptions contain antibiotics [13].
Prescriptions per GP
Calculated by dividing total prescriptions by the number of active GPs per period, reflecting prescriber workload. This metric helps contextualize changes in prescribing intensity, such as potential reductions in in-person visits during lockdowns or increased antibiotic prescribing during telemedicine peaks [14].
Antibiotics included
The analysis focused on commonly prescribed broad-spectrum antibiotics frequently used in outpatient general practice:
Oral Antibiotics: Amoxicillin 500 mg, Ciprofloxacin 500 mg, Cephalexin 500 mg, Azithromycin 250 mg.
Anzeige
Injectable Antibiotics: Penicillin G (dosages: 800,000 IU, 1.2 million IU, 3.3 million IU). Penicillin G was the sole injectable antibiotic analyzed due to its dominance in pre-pandemic injectable prescriptions (>80%). Other injectable antibiotics (e.g., ceftriaxone) were excluded as they accounted for less than 5% of total injectable prescriptions in the dataset. This 5% threshold was applied to maintain a focused analysis on the most clinically significant and frequently prescribed agents, thereby enhancing the clarity and interpretability of the findings related to injectable antibiotic use. The selected agents reflect national prescribing patterns and clinical use for respiratory and general infections. Budget and procurement limitations at the clinic precluded inclusion of newer or more expensive antibiotics [15].
Statistical analysis
Descriptive and Inferential statistics
Trend analysis
Percentage changes in antibiotic prescriptions between periods were estimated using Poisson regression models suitable for count data, with 95% confidence intervals (CIs) [16]. For example, the percentage difference between the PreCOVID and COVID periods was calculated as: (COVID − PreCOVID) / PreCOVID × 100.
Comparative tests
Chi-square tests with Yates’ continuity correction were applied for proportional differences such as antibiotic-to-total drug ratios. One-way ANOVA (for normally distributed data) or Kruskal-Wallis tests (for nonparametric data) compared mean prescriptions per GP across periods, reporting mean differences (MDs) with 95% CIs [17].
Multivariable regression
Negative binomial regression models adjusted for overdispersion and covariates, including GP count and total prescriptions. Results are presented as incidence rate ratios (IRRs) with 95% CIs [18].
Statistical significance
A two-sided p-value < 0.05 and non-overlapping 95% CIs were considered statistically significant [19].
Software
Analyses were conducted using SPSS version 26 and R version 4.0 (for regression model confidence intervals).
Results
Trends in antibiotic prescriptions
The analysis revealed significant increases in antibiotic prescribing during and after the COVID-19 pandemic compared to the pre-pandemic baseline [17‐19]. Total antibiotic prescriptions increased from 107,365 (PreCOVID) to 200,433 during the COVID period, representing an 87% increase (95% CI: 82%–92%; χ²=215.4, p < 0.001). PostCOVID prescribing remained elevated at 208,040 items, a 94% increase compared to PreCOVID (95% CI: 89%–99%; p < 0.001). The antibiotic-to-total drug ratio doubled from 7.6% to 14.4% during the pandemic (Δ = 6.8%, 95% CI: 6.2%–7.4%; p < 0.001) and remained high post-COVID at 13.5%. Mean prescriptions per GP increased significantly from 21,876 to 23,300 (MD = + 1,424, 95% CI: 120–2,728; F = 3.2, p = 0.03). These trends in total antibiotic items and the antibiotic-to-total drug ratio across the three periods are illustrated in Fig. 1. These prescribing metrics by study period are summarized in Table 1.
Table 1
Antibiotic prescribing metrics by study period
Period | Active GPs | Total prescriptions | Total drug items | Antibiotic items | Antibiotic/Total drug (%) |
|---|---|---|---|---|---|
PreCOVID | 21 | 459,388 | 1,417,054 | 107,365 | 7.6 |
COVID | 30 | 460,023 | 1,395,310 | 200,433 | 14.4 |
PostCOVID | 22 | 512,593 | 1,537,406 | 208,040 | 13.5 |
Fig. 1
Total antibiotic items and antibiotic/total drug ratio (%) across the three periods
Anzeige
Class-specific prescribing patterns
Oral antibiotics
Azithromycin prescriptions surged by 120% (95% CI: 110%–130%; p < 0.001), largely prescribed for respiratory tract infections (RTIs) in line with Iran’s interim COVID-19 guidelines (MoH Protocol No. 5, 2020) [20], with 72% of azithromycin prescriptions linked to RTI diagnostic codes. β-lactams such as amoxicillin showed a 40% increase in outpatient RTI. Amoxicillin and cephalexin prescribing remained relatively stable (+ 15% and + 12%, respectively), with changes not statistically significant after Bonferroni correction [21].
Injectable penicillins
Penicillin G 800,000 IU prescriptions dropped from 4,516 units preCOVID to zero postCOVID, a 100% decrease (95% CI: 99%–100%; p < 0.001), attributed to nationwide shortages and reduced injectable use during lockdowns [22]. These trends are visually summarized in Fig. 2, which shows class-specific prescribing patterns of oral antibiotics and injectable penicillins.
Fig. 2
Class-specific prescribing patterns of oral antibiotics and injectable penicillins
Temporal patterns and policy impacts
Three distinct phases were observed: (1) PreCOVID (2019): Typical outpatient patterns with balanced oral/injectable use. (2) Pandemic peak (2020–2021): Characterized by an azithromycin surge (IRR = 2.2, 95% CI: 2.0–2.4) and an injectable collapse for Penicillin G (IRR = 0.05, 95% CI: 0.02–0.10). (3) PostCOVID (2021–2022): Marked by sustained high oral use and no recovery of injectables. Policy and supply chain factors were key drivers; the rise in azithromycin use aligns with Iran’s interim COVID-19 guidelines [20], while penicillin shortages are documented in Iran FDA reports [22]. As depicted in Fig. 3, the IRR for azithromycin prescriptions increased significantly during both the COVID (IRR: 2.2, 95%CI: 2.0–2.4) and post-COVID periods (IRR: 2.0, 95%CI: 1.8–2.2), while the IRR for penicillin G collapsed (COVID: 0.05 [0.02–0.10]; post-COVID: 0.01 [0.00–0.05]), reflecting a dramatic shift in injectable antibiotic use.
Fig. 3
Incidence rate ratios (IRRs) for key antibiotics during COVID and Post-COVID
As depicted in Fig. 3, the IRR for azithromycin prescriptions increased significantly during both the COVID (IRR: 2.2, 95%CI: 2.0–2.4) and post-COVID periods (IRR: 2.0, 95%CI: 1.8–2.2), while the IRR for penicillin G collapsed (COVID: 0.05 [0.02–0.10]; post-COVID: 0.01 [0.00–0.05]), reflecting a dramatic shift in injectable antibiotic use.
Anzeige
Discussion
Our findings demonstrate a marked 87% increase in antibiotic prescribing among Iranian general practitioners (GPs) during the COVID-19 pandemic, predominantly driven by empirical use of broad-spectrum agents such as azithromycin and β-lactams (amoxicillin, cephalexin). This surge is consistent with national and international evidence indicating widespread, often unwarranted, antibiotic use in COVID-19 patients despite low rates of confirmed bacterial coinfection (6–14% in Iran) [23, 24]. Notably, Hooshmand et al. reported that only 14.4% of hospitalized COVID-19 cases had documented bacterial coinfection, yet antibiotic consumption was extensive, reaching 121.6 DDD per 100 hospital bed-days [25], and they observed high rates of broad-spectrum antibiotic use in COVID-19 wards, with a strong association between antibiotic exposure and increased mortality. The persistence of elevated prescribing (+ 94% post-pandemic) highlights systemic challenges in Iran’s antimicrobial stewardship infrastructure. In contrast, settings such as Abu Dhabi and the SSO-insured population in Iran reported significant post-pandemic reductions in antibiotic use (30% decline), likely reflecting stronger stewardship enforcement, greater diagnostic capacity, and more robust pandemic preparedness [26, 27].
A major driver of empirical antibiotic prescribing was the clinical overlap between COVID-19 and bacterial pneumonia (e.g., fever, dyspnea), which was exacerbated by limited access to rapid diagnostic testing, especially in outpatient settings [28]. Our results align with national data showing that approximately 60% of COVID-19 patients in Iran received antibiotics, even though only about 10% had confirmed bacterial coinfections a trend mirrored in European cohorts [29]. The rapid adoption of telemedicine during lockdowns further contributed to empirical prescribing for respiratory symptoms, as documented in both the UAE and UK [30]. Additionally, increased specialist referrals in primary care (e.g., a 19.3% rise in Golestan province) suggest heightened diagnostic uncertainty at the GP level [31]. This diagnostic uncertainty and limited access to point-of-care tests directly link “diagnostic limitations” to the observed prescribing trends, justifying its mention in the conclusion as a key area for intervention.
The substantial rise in azithromycin use (65% nationally) and cephalexin prescriptions (40% increase in Tehran) raises serious concerns about accelerating resistance among WHO-priority pathogens such as Streptococcus pneumoniae and Escherichia coli [32]. This risk is compounded by Iran’s already significant AMR burden and a high baseline of irrational prescribing (59% in primary healthcare settings) [33]. Raoofi et al. documented alarming increases in resistance rates among Gram-negative bacteria, particularly Pseudomonas aeruginosa (89%) and Klebsiella pneumoniae (66.3%) during the pandemic [34]. Furthermore, Hooshmand et al. found that antibiotic use in COVID-19 patients was associated with a fourfold increased risk of death [25]. The decline in injectable penicillin use, attributed to supply-chain disruptions, was not offset by rationalization of oral antibiotic prescribing, which instead remained high and misaligned with WHO guidance [35]. This entrenched pattern of overuse threatens to further entrench AMR in Iran’s healthcare system.
Iran’s fragmented antimicrobial stewardship infrastructure is characterized by inconsistent adherence to national and international prescribing guidelines, a lack of real-time prescription monitoring and feedback systems (especially in outpatient and GP settings), weak implementation of stewardship programs (particularly outside major hospitals), and disruptions in reverse referral systems and primary care coordination, as seen in the 36.1% decline in reverse referrals in Golestan province, which placed additional strain on GPs and likely contributed to overprescribing. In contrast, targeted interventions in the SSO-insured sector, such as direct monitoring and feedback, were associated with more substantial reductions in antibiotic use post-pandemic [36,37].
Our findings are in line with several major Iranian studies. Salehi et al. documented widespread antibiotic use in COVID-19 inpatients, with a direct correlation between antibiotic consumption and mortality [38]. Raoofi et al. reported increased resistance among Gram-negative bacteria during the pandemic, further limiting treatment options [36]. Hooshmand et al. highlighted excessive antibiotic use and its association with adverse outcomes in hospitalized COVID-19 patients [27]. Mehrizi et al. identified antibiotics, corticosteroids, and antithrombotics as the most commonly prescribed drugs for COVID-19, with notable temporal trends [39]. These studies collectively underscore the urgent need for strengthening antimicrobial stewardship programs (especially in outpatient and GP settings), expanding access to rapid diagnostic tests to reduce empirical prescribing, implementing robust, real-time prescription monitoring via national digital health platforms, enhancing prescriber education and feedback mechanisms, and coordinating intersectoral efforts among regulators, insurers, pharmacists, and laboratories to address AMR comprehensively [40].
Conclusion
Our study reveals a significant and sustained increase in antibiotic prescribing by Iranian general practitioners during and after the COVID-19 pandemic, predominantly driven by broad-spectrum agents such as azithromycin and β-lactams. This trend, coupled with Iran’s existing challenges in antimicrobial stewardship and diagnostic limitations, threatens to accelerate antimicrobial resistance nationally. The findings highlight urgent needs for strengthening stewardship programs targeted specifically at outpatient and primary care settings, expanding access to rapid diagnostic testing, and implementing national prescription monitoring systems. Collaborative efforts among policymakers, healthcare providers, insurers, pharmacists, and laboratories will be essential to curb unnecessary antibiotic use and protect public health. Future research and policy must prioritize evidence-based prescribing interventions and real-time surveillance to ensure post-pandemic recovery does not come at the cost of escalating antimicrobial resistance.
Limitations
Our analysis, like most national studies, is constrained by the lack of detailed patient-level outcomes (e.g., treatment success, AMR emergence). Data entry inconsistencies in the hospital information system (HIS) may affect the accuracy of prescription records. While the single-center design enhances internal validity by controlling for institutional confounding factors, and the clinic’s ~ 7% share of national GP prescriptions supports its representativeness for public primary care, the findings may not be fully generalizable to all Iranian GPs, particularly those in private practice or rural settings with different formularies and patient populations. The short post-pandemic observation period also limits the temporal stability of the findings.
Acknowledgements
The authors thank the staff of the Social Security Organization clinic for their support and the Ethics Committee of Zanjan University of Medical Sciences for approval.
Declarations
Ethics approval and consent to participate
This study was approved by the Ethics Committee of Zanjan University of Medical Sciences (approval code ZUMS.REC.1394.322). This study was conducted in accordance with the Declaration of Helsinki and relevant national ethical guidelines for medical research involving human data. Due to the retrospective and anonymized nature of the data extracted from Iran’s National Hospital Information System (HIS), the requirement for individual informed consent was waived.
Consent for publication
Not applicable, as no identifying images or personal details compromising anonymity are included in the manuscript.
Competing interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
