{"title":"Patient prioritization for pharmaceutical intervention in the hospital setting: a retrospective cross-sectional study.","authors":"Maude Plourde, Chantal Gilbert, Mélanie Noël, Sophie Ruelland, Pierre-Hugues Carmichael, Danielle Laurin","doi":"10.1093/ijpp/riaf033","DOIUrl":null,"url":null,"abstract":"<p><strong>Objectives: </strong>Prioritization of patients requiring pharmaceutical intervention is critical given limited resources. Data from pharmacy software could be used to target patients. This retrospective cross-sectional study aimed to describe the method implemented in a hospital care unit to prioritize hospitalized patients and compare the characteristics of those receiving a pharmaceutical intervention and those not. This study also explored the possibility of predicting an intervention using pharmacy software data.</p><p><strong>Methods: </strong>All patients admitted to a hospital care unit between November 2019 and April 2020 were included. Prioritization with the pharmacy software was based on preselected admission diagnoses and by operating an antimicrobial stewardship programme. Medications and patients' characteristics were extracted from the pharmacy software. Pharmaceutical interventions and drug-related problems were collected from medical records. Two machine learning algorithms were used to produce rule-based models for pharmaceutical intervention prediction.</p><p><strong>Key findings: </strong>A total of 850 admissions were included. A medication review following prioritization with the pharmacy software or due to external requests was carried out by clinical pharmacists in 45% of admissions, followed by an intervention in 81% of them. Patients who received an intervention had lower creatinine clearance levels and more regular medications including antibacterials for systemic use, diuretics, and psychoanaleptics. The two resulting interpretable models comprised either 6 or 17 predictors of a pharmaceutical intervention.</p><p><strong>Conclusions: </strong>Pharmacy software data may be used for more efficient prioritization of patients using specific criteria. Rule-based models are promising avenues to help clinical pharmacists systematically identify patients requiring pharmaceutical intervention, but further work is warranted.</p>","PeriodicalId":14284,"journal":{"name":"International Journal of Pharmacy Practice","volume":" ","pages":""},"PeriodicalIF":1.5000,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Pharmacy Practice","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1093/ijpp/riaf033","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"PHARMACOLOGY & PHARMACY","Score":null,"Total":0}
引用次数: 0
Abstract
Objectives: Prioritization of patients requiring pharmaceutical intervention is critical given limited resources. Data from pharmacy software could be used to target patients. This retrospective cross-sectional study aimed to describe the method implemented in a hospital care unit to prioritize hospitalized patients and compare the characteristics of those receiving a pharmaceutical intervention and those not. This study also explored the possibility of predicting an intervention using pharmacy software data.
Methods: All patients admitted to a hospital care unit between November 2019 and April 2020 were included. Prioritization with the pharmacy software was based on preselected admission diagnoses and by operating an antimicrobial stewardship programme. Medications and patients' characteristics were extracted from the pharmacy software. Pharmaceutical interventions and drug-related problems were collected from medical records. Two machine learning algorithms were used to produce rule-based models for pharmaceutical intervention prediction.
Key findings: A total of 850 admissions were included. A medication review following prioritization with the pharmacy software or due to external requests was carried out by clinical pharmacists in 45% of admissions, followed by an intervention in 81% of them. Patients who received an intervention had lower creatinine clearance levels and more regular medications including antibacterials for systemic use, diuretics, and psychoanaleptics. The two resulting interpretable models comprised either 6 or 17 predictors of a pharmaceutical intervention.
Conclusions: Pharmacy software data may be used for more efficient prioritization of patients using specific criteria. Rule-based models are promising avenues to help clinical pharmacists systematically identify patients requiring pharmaceutical intervention, but further work is warranted.
期刊介绍：
The International Journal of Pharmacy Practice (IJPP) is a Medline-indexed, peer reviewed, international journal. It is one of the leading journals publishing health services research in the context of pharmacy, pharmaceutical care, medicines and medicines management. Regular sections in the journal include, editorials, literature reviews, original research, personal opinion and short communications. Topics covered include: medicines utilisation, medicine management, medicines distribution, supply and administration, pharmaceutical services, professional and patient/lay perspectives, public health (including, e.g. health promotion, needs assessment, health protection) evidence based practice, pharmacy education. Methods include both evaluative and exploratory work including, randomised controlled trials, surveys, epidemiological approaches, case studies, observational studies, and qualitative methods such as interviews and focus groups. Application of methods drawn from other disciplines e.g. psychology, health economics, morbidity are especially welcome as are developments of new methodologies.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
