Prognostic prediction models and clinical tools based on consensus to support patient prioritization for clinical pharmacy services in hospitals: A scoping review

Identifying patients at high risk of adverse medication-related outcomes for targeted clinical pharmacy services is essential in hospital pharmacy. Models and predictive tools to prioritize patients are available to the clinical pharmacy services for hospital use. To describe and assess prognostic models and predictive tools used to identify inpatients at risk of adverse medication-related outcomes. We searched in Medline, Lilacs, Cochrane, CINAHL, Embase, Scopus and Web of Science, databases of theses and dissertations, and the references of the selected studies. The screening was carried out by two independent researchers. Cross-sectional studies, prospective or retrospective cohort studies, and case-control studies were eligible for inclusion. The studies addressed the development or validation of predictive models and clinical prioritization tools based on expert opinion to identify inpatients at risk of adverse medication-related outcomes. 25 studies were included, 13 of which were prognostic prediction models, seven were instrument development using the consensus method, and five were validation. The outcome events were drug-related problems (9), adverse drug reactions (8), adverse drug events (6), and medication errors (2). Most studies targeted adult patients (14), eight had older adult patients, one had obstetric patients, and others had pediatric patients. External validation was performed after the development study in three studies. The predictive model with a low risk of bias was the Medicines Optimisation Assessment Tool. Limited details on the method of expert involvement and the number of experts were identified in four studies. The development of patient prioritization tools to optimize pharmacotherapy by clinical pharmacy services is a complex process. The predictive models and tools analyzed are limited in their development and validation process, hindering their effective use in prioritizing patients by the clinical pharmacy services. The development of additional prognostic prediction models for drug-related problems is a priority.