Optimal prescribing of antimicrobials is becoming increasingly challenging because of the growing complexity of guidelines and constantly changing distribution of infectious pathogens. Prescribing antimicrobials appropriately according to local guidelines optimizes therapy for the individual patient and reduces the emergence of resistance. By adapting and evaluating a smartphone based app containing local guidelines we aim to study appropriate prescribing of antimicrobials by physicians in three hospitals (Netherlands, Sweden and Switzerland).

Rationale: Antimicrobials are an indispensable part of modern medicine. However, optimal prescription of these agents is becoming increasingly challenging because of the growing complexity of guidelines, and constantly changing epidemiology of infectious pathogens. Moreover, due to local variations in the prevalence of certain pathogens and antimicrobial resistance (AMR), antimicrobial choices need to be tailored to local epidemiology. Improvement of antimicrobial use, in particular prevention of overuse and suboptimal use of antimicrobials, through antimicrobial stewardship (AMS) programs is increasingly regarded as indispensable, both to optimize therapy for the individual patients as well as to reduce emergence and spread of AMR. With the widespread use of electronic health records (EHR) and handheld electronic devices in hospitals, informatics-based AMS interventions hold great promise as tools to improve antimicrobial prescribing. However, they are still underdeveloped, understudied and underutilized.

Objective: The study aims to adapt and evaluate the "AB-assistant", a smartphone based digital stewardship application that is customizable to local guidelines by local antibiotic stewards and therefore has the potential to be used worldwide, including in low- and middle-income countries.

Study design: The existing North American Spectrum app (SpectrumMD; Canada) will be adjusted and translated for the European market. During a usability study physicians will use the app for two weeks followed by individual interviews to determine facilitators and barriers of app use. Based on the results of these interviews the app will be adjusted if necessary. After adaptation and usability testing, thereafter the AB-assistant app will be evaluated in an international, multicentre, randomized clinical trial involving centres in 3 countries in different settings with appropriate antimicrobial use as a primary outcome. In a stepped wedge cluster randomized trial, wards will be randomised after stratification for specialty. At baseline a 2-week measurement period will be done, followed by the introduction of the intervention to 6 wards (in 3 hospitals) with a 4-week interval with 6 inclusion periods. This cycle will be repeated with the inclusion of all new intervention wards. We include the 36 wards in total during the 6 inclusion phases and at the end of the inclusion time we allow use of the app by everyone, also wards not included in the study.

Inclusion Criteria:

Cluster level (wards):

• Medical and surgical wards.

Physician level:

• All physicians involved in antibiotic prescribing decisions in the participating wards.

Patient level:

• All patients hospitalized in the participating wards >= 18 years of age to whom systemic antimicrobials are prescribed.

Exclusion Criteria:

Cluster level (wards):

• Outpatient clinics
• Psychiatry wards
• ICU

Physician level:

• None

Patient level:

• None

Treatment level:

• Surgical and medical prophylaxis.

De-identified raw data will be made publicly on request available from the time of publication or within a reasonable time period of the experiments being completed. General project data and metadata will be included in the supplementary data of published articles if needed and or shared through university repositories.

Masking of care providers and investigators is not feasible. Outcome assessors and data analysts will be blinded to the study arm allocation.