Datasets for Baseline inpatient antibiotic use patterns in seven Malawian public hospitals: A 2021 point prevalence survey in antimicrobial resistance sentinel surveillance sites

The datasets shared alongside this publication were generated from a multisite point prevalence survey (PPS) on antibiotic use conducted in June 2021 across seven hospitals participating in Malawi’s national antimicrobial resistance (AMR) sentinel surveillance network: Queen Elizabeth Central Hospital (QECH), Zomba Central Hospital (ZCH), Kamuzu Central Hospital (KCH), Mzuzu Central Hospital (MCH), Mangochi District Hospital (MDH), Mzimba South District Hospital (MSDH), and Malamulo Adventist Hospital (MAH). For each hospital, five de-identified datasets are provided, corresponding to the standard World Health Organization (WHO) PPS data collection forms. All datasets are anonymised, contain no direct personal identifiers, and are suitable for secondary analysis. 1. Hospital Form Dataset Description:This dataset contains hospital-level structural and contextual information for each participating facility. It captures characteristics relevant to interpreting antibiotic use patterns at the institutional level. Key variables include: Hospital identifier Level of care (secondary or tertiary) Bed capacity Annual admission volume Unit of analysis: HospitalRecords: One record per hospital 2. Ward Form Dataset Description:The ward-level dataset describes the characteristics of inpatient wards included in the survey on the day of data collection. It provides information on patient distribution and ward types within each hospital. Key variables include: Hospital identifier Ward identifier Ward type (medical, surgical, obstetric, paediatric) Patient category (adult or paediatric) Number of eligible patients Number of included patients Unit of analysis: Ward 3. Patient Form Dataset Description:This dataset contains de-identified patient-level information for all inpatients eligible on the survey day. It captures basic demographic and clinical characteristics and whether the patient was receiving antibiotics at the time of the survey. Key variables include: Hospital and ward identifiers Patient age (in years) Sex Antibiotic use status (yes/no) Unit of analysis: Patient 4. Antibiotic Form Dataset Description:The antibiotic-level dataset documents detailed information on each systemic antibiotic prescribed to inpatients at the time of the survey. It allows analysis of prescribing patterns and antibiotic classes. Key variables include: Hospital, ward, and patient identifiers (coded) Antibiotic name (International Non-proprietary Name, INN) Dose and route of administration Type of therapy (empirical or directed) Microbiology sample collection status (yes/no/unknown) Unit of analysis: Antibiotic prescription 5. Indication Dataset Description:This dataset captures the documented clinical indication for each antibiotic prescription and assesses compliance with the Malawi Standard Treatment Guidelines (MSTG). Key variables include: Hospital and patient identifiers (coded) Clinical indication for antibiotic use Guideline compliance status (yes, no, not assessable, no information) Unit of analysis: Antibiotic indication Data Structure and Linkage All datasets use harmonised, coded identifiers (hospital, ward, patient, and prescription IDs) that allow linkage across datasets while preserving anonymity. This relational structure enables reproducible analyses of antibiotic use prevalence, prescribing patterns, guideline compliance, and microbiology utilisation within and across hospitals. Ethical and Privacy Considerations All shared datasets are fully de-identified. No names, exact dates of admission, national identifiers, or other personal identifiers are included. Ethical approval for data collection and secondary data sharing was obtained from the National Health Sciences Research Committee (NHSRC), Malawi.