Inappropriate antibiotic prescribing in primary care remains an important contributor to antimicrobial resistance. Despite the availability of evidence-based clinical guidelines, antibiotics are still frequently prescribed for self-limiting infections. Digital clinical decision support tools may help general practitioners (GPs) align prescribing decisions with guideline recommendations during patient consultations.

This study evaluates the impact of a digital Prescription Support System (PSS) designed to support antimicrobial prescribing in Belgian primary care. The PSS provides guideline-based recommendations derived from the Belgian BAPCOC guidelines for common infections in ambulatory care. Recommendations are presented through a user-friendly decision tree that is integrated into existing electronic health record systems and can be consulted during routine care.

The study is embedded within the national implementation strategy of the PSS coordinated by the National Institute for Health and Disability Insurance (RIZIV-INAMI). A stepped-wedge cluster randomized design is used, in which participating general practices transition sequentially from usual care to access to the PSS over four predefined implementation steps. This approach ensures that all participating practices eventually receive access to the system while allowing comparisons over time.

The primary objective is to assess whether implementation of the PSS is associated with changes in antibiotic prescribing in Belgian general practice. Prescribing outcomes are measured using routinely collected indicators from the Belgian Antibiotic Barometer, including overall antibiotic prescribing rates and the use of broad- versus narrow-spectrum antibiotics. Secondary objectives include assessing the usability and acceptability of the PSS among clinicians and identifying factors that influence its adoption in daily practice. The study will also monitor potential unintended consequences, such as changes in workflow or concerns about underprescribing.

Findings from this study will inform future decisions regarding further optimization and wider implementation of the PSS in Belgian primary care.

Inappropriate antibiotic prescribing in primary care contributes to the development of antimicrobial resistance. Despite guidelines and public health efforts, overprescribing remains prevalent, especially for self-limiting infections which often resolve without antibiotic treatment. Digital tools such as a Prescription Search Support (PSS) may aid general practitioners (GPs) in making more informed prescribing decisions.

This implementation project builds upon a previously conducted mixed-methods implementation study evaluating the usability of a PSS, which was approved by the Ethical Review Board UZ KU Leuven (Onderwijs-Begeleidingscommissie) on 27 September 2024 (MP031770).

The PSS under study contains the current recommendations for the proper use of antimicrobial agents for common infections in ambulatory care in Belgium. In other words, it represents the BAPCOC guidelines in a digital and user-friendly environment, by using a decision tree. Advice for the following diagnoses (ICPC-2 codes) is included in the PSS: sore throat (R21, R74), pertussis (R71), common cold (R74), acute rhinosinusitis (R75), acute bronchitis / bronchiolitis (R78), influenza (R80), pneumonia (R81), ear pain (H01), otorrhea (H04), acute otitis media (H71), acute epiglottitis (R77), acute exacerbations of chronic obstructive pulmonary disease, aspiration pneumonia (R81), laryngitis stridulosa (R77), acute cystitis (U71), acute prostatitis (Y73), acute pyelonephritis (U70), acute vulvovaginitis (X84), asymptomatic bacteriuria, epididymo-orchitis (Y74), genital herpes (Y72, X90), partner treatment and treatment of asymptomatic sexually transmitted infections, pelvic inflammatory disease (X74), syphilis (X70, Y70), and urethritis (U72).

This PSS was developed as part of the Belgian National Action Plan on AMR by the Federal Public Service Health, Food Chain Safety and Environment and the National Institute for Health and Disability Insurance (NIHDI; RIZIV-INAMI), in collaboration with various experts and stakeholders, including physicians, infectiologists, hygienists, policymakers, guideline developers, and software developers. The company Smals, under instruction of RIZIV-INAMI, developed the software for this application in close collaboration with the RIZIV-INAMI working group. Guideline recommendations regarding appropriate prescribing of antimicrobials in ambulatory care were converted into decision rules, which serve as the knowledge source for the PSS.

Digital education offers several advantages, such as easy access without time or location constraints, varying possible forms and levels, and the ability to disseminate evidence on a large scale. Previous research has shown that standalone web- or smartphone-based applications in this context have the potential to increase physicians' knowledge of antimicrobials and guideline recommendations, potentially improving guideline adherence.

In Belgium, antibiotic use in ambulatory care is closely monitored through the Antibiotic Barometer, a national surveillance system coordinated by the Academic Centre of General Practice of KU Leuven and financed by the National Institute for Health and Disability Insurance (RIZIV-INAMI). The Antibiotic Barometer was approved by the Sociaal-maatschappelijke Ethische Commissie (SMEC; G-2023-6352) and provides quarterly feedback to general practices based on routine health insurance claims data.

This study evaluates the implementation and impact of the PSS in routine Belgian primary care using a stepped-wedge cluster randomized trial design. Participating general practices are recruited to ensure broad representation across Belgian primary care. General practices constitute the unit of randomization and are randomly allocated to one of four predefined implementation steps. All practices start in a control condition without access to the PSS and transition once to the intervention condition according to their assigned step. Each implementation step lasts three months, after which practices retain access to the PSS for the remainder of the study period. By the end of the study, all participating practices will have received access to the intervention.

Data for this study are derived from routine sources and handled within the secure data infrastructure of KU Leuven. Prescribing data obtained via the Antibiotic Barometer are aggregated and pseudonymized at practice level prior to analysis, ensuring that individual patients cannot be identified. Data linkage, storage, and analysis are conducted in accordance with applicable data protection regulations. Quantitative and qualitative analyses are used to evaluate both the effectiveness and implementation of the PSS in real-world primary care. Findings from this study will inform future policy decisions regarding national rollout, optimization, and long-term integration of the PSS within Belgian antimicrobial stewardship strategies.

Inclusion Criteria:

• General practices in Belgium using compatible EHR systems.
• Willingness to use and activate the PSS by giving consent.
• Agreement to participate in the study, provided signed informed consent form is in place.

Exclusion Criteria:

• Practices using an EHR system that solely allows a single sign-on to the standalone PSS.
• Practices already using a similar PSS or participating in overlapping projects.

Individual patient data from this trial will not be published in the public domain. Deidentified participant data is available for further analyses. Requests for data, with justification, should be sent to Jan Verbakel (jan.verbakel@kuleuven.be). The trial protocol is available online for an indefinite period.

Study protocol is published and accessible. SAP, CSR, ICF and analytical code will be available after publishing the results in a medical journal. Individual patient data from this trial will not be published in the public domain.

Individual patient data from this trial will not be published in the public domain. Deidentified participant data is available for further analyses. Requests for data, with justification, should be sent to Jan Verbakel (jan.verbakel@kuleuven.be). The trial protocol is available online for an indefinite period.