Inappropriate antimicrobial use drives antimicrobial resistance. Providing clear, accessible prescribing guidance is a core component of antimicrobial stewardship (AMS) programs. After Laos released national antimicrobial prescribing guidelines in 2021, investigators tested whether delivering these guidelines via a smartphone application with AMS training would increase adherence compared with the same guidance provided in paper form with training.
The study was an open cohort stepped-wedge cluster randomized controlled trial conducted in six hospitals across Lao PDR between 2021 and 2022. The stepped-wedge design introduced the intervention in sequential steps across clusters (hospitals), allowing all sites to receive the intervention over time while enabling comparison with periods when only the paper-based guideline was available.
The intervention combined access to a smartphone application containing the national antimicrobial prescribing guidelines and AMS training on antimicrobial use principles and app use. The comparator was the same guideline content provided in a paper-based format together with training on its use. Both delivery modes used the same guideline recommendations.
The primary outcome was guideline adherence at the prescription level, defined as prescriptions that used the correct antimicrobial agent(s) and the correct dose according to the prescribing guidelines. Repeated point prevalence surveys (PPS) at approximately four-month intervals captured prescribing data in inpatient and outpatient settings. Prescribing was assessed from patients’ medical records; prescriptions could not be reliably linked to individual prescribers.
Data were analysed using mixed-effects regression models that accounted for time and clustering by hospital. Models adjusted for potential confounders including sequence of intervention introduction, patient age and gender, prescribing indication, and hospital department. Results are presented as proportions with 95% confidence intervals and adjusted odds ratios (aOR) where applicable.
Across the inpatient cohort, 3,561 antimicrobial prescriptions were assessed. Observed guideline adherence in the reference (paper) group was 17.0% (95% CI 15.0–19). In the intervention (smartphone) group inpatient adherence was 25.6% (95% CI 23.1–28.3). After adjusting for time, clustering, and confounders, there was no statistically significant difference in inpatient adherence by delivery mode (aOR = 1.26, 95% CI 0.8–1.9; p = 0.276).
For outpatient prescribing (3,905 prescriptions assessed), the adjusted analysis found no overall improvement in adherence with the smartphone app compared with paper-based guidelines (aOR = 0.91, 95% CI 0.7–1.1; p = 0.406).
Analyses revealed time-dependent variation in outpatient adherence related to the step at which hospitals received the intervention. Hospitals exposed to the intervention in the earliest group showed higher outpatient adherence (32.1% [302/940]; 95% CI 29.1–35.2) compared with hospitals in the last group (19.1% [241/1,261]; 95% CI 16.9–21.4). The adjusted odds of adherence were higher in the first group compared with the last (aOR = 1.95, 95% CI 1.4–2.7; p < 0.001).
The investigators also reported that a persistently high proportion of inappropriate antimicrobial use occurred for surgical and obstetric prophylaxis, which contributed to overall low guideline adherence.
Key limitations reported by the authors include the inability to link prescriptions to individual prescribers, preventing assessment of individual behaviour change after guideline implementation. The study assessed prescribing from medical records, which may have limitations in completeness and detail. The stepped-wedge cluster randomized design allowed within-site comparisons over time and ensured all hospitals received the intervention, which is a strength for implementation research.
The trial demonstrates that it is feasible to deliver national prescribing guidelines via a smartphone application in a low- and middle-income country setting such as Laos. However, the intervention (app plus AMS training) did not produce the pre-specified improvement in guideline adherence compared with paper-based guidelines with training. The authors conclude that availability and delivery mode of guidance alone are insufficient to secure higher adherence; adherence depends on multiple contextual, clinical, and behavioural factors that may require multifaceted stewardship strategies beyond guideline dissemination.
The trial was registered at ClinicalTrials.gov (NCT04914793). The article reports that a de-identified dataset used for analyses is publicly available on Figshare and that code for sample size calculations and statistical analyses is available as supplementary material. The full original study dataset is owned by the Ministry of Health, with access requests managed through the Mahidol-Oxford Tropical Medicine Research Unit Data Access Committee as described in the paper.
In six hospitals in Lao PDR, introducing national antimicrobial prescribing guidelines via a smartphone application with AMS training was operationally feasible but did not significantly increase prescription-level guideline adherence compared with a paper-based guideline and training after adjusting for time and clustering. The findings underscore that improving antimicrobial prescribing likely requires broader, multifaceted interventions addressing the contextual drivers of inappropriate use in addition to guideline access.