Artificial intelligence in hospital pharmacy practice: Effectiveness, internal validity, and translational readiness: A systematic review

Authors

• Kwanchai Rattanamanee, Department of Pharmacy Practice, Faculty of Pharmaceutical Sciences, Naresuan University, Phitsanulok, Thailand
• Parnupong Kiewsiri, Medical and Pharmacy Innovation Research and Development Unit, Faculty of Pharmaceutical Sciences, Naresuan University, Phitsanulok, Thailand
• Prayuth Poowaruttanawiwit, Naresuan University, Faculty of Pharmaceutical SciencesFollow

Abstract

Background: Artificial intelligence (AI) is increasingly integrated into hospital pharmacy, offering potential to enhance medication safety, therapeutic decision-making, and operational efficiency. However, the internal validity and translational readiness of current interventions remain uncertain.
Objectives: The objective of the study was to systematically review AI applications in hospital pharmacy and evaluate their effectiveness, internal validity, and readiness for clinical integration.
Methods: Following Preferred Reporting Items for Systematic reviews and Meta-Analyses 2020 guidelines and PROSPERO registration (CRD420251067936), We systematically searched PubMed, Embase, Scopus, Web of Science, IEEE Xplore, and ThaiJO (2000–June 2025). for randomized controlled trials (RCTs), cohort, quasi-experimental, and case–control studies evaluating AI interventions in inpatient or outpatient pharmacy practice. Risk of bias was assessed using RoB 2, ROBINS-I, and QUADAS-2.
Results: Thirteen studies (5 RCTs, 8 observational) met inclusion criteria. AI interventions in clinical decision support, adherence monitoring, and error prevention demonstrated significant improvements in safety and process outcomes. High internal validity was reported in multicenter, blinded, electronic health recordintegrated RCTs using objective endpoints such as glycemic control and adherence. Observational evidence supported AI’s utility in pharmacovigilance (Odds ratios was approximately 1.7). Studies with subjective measures, simulated designs, or limited confounder control showed lower validity.
Conclusion: AI shows demonstrated measurable improvements in medication safety and operational outcomes, although variability in study quality limits generalizability. Future research should focus on standardized outcomes, multicenter real-world trials, and governance frameworks to strengthen validity, scalability, and safe clinical adoption.

DOI

10.56808/3027-7922.3259

Recommended Citation

Rattanamanee, Kwanchai; Kiewsiri, Parnupong; and Poowaruttanawiwit, Prayuth (2026) "Artificial intelligence in hospital pharmacy practice: Effectiveness, internal validity, and translational readiness: A systematic review," The Thai Journal of Pharmaceutical Sciences: Vol. 50: Iss. 2, Article 6.
DOI: https://doi.org/10.56808/3027-7922.3259
Available at: https://digital.car.chula.ac.th/tjps/vol50/iss2/6