Development and Evaluation of a Web Based Integrated Clinical Evidence Search System for Drug Information Services

Background : This study developed and evaluated a web-based clinical evidence search system designed to improve the efficiency of drug information services.By integrating multiple medical databases with generative Artificial Intelligence (AI), the system enables pharmacists to more efficiently evaluate clinical evidence and make informed decisions. Methods :We developed a search system that combines APIs from Europe PMC, PubMed, and ClinicalTrials.gov.This system utilizes MeSH-based term expansion for searches and assesses journal impact using the OpenAlex API.Clinical recommendations were generated through a two-stage process.First, evidence strength was measured using an evidence profile scoring system based on four factors: volume, quality, effect, and time.Second, AI reviewed these scores to assign traffic light grades.Additionally, an equivalent dose search module was developed, employing a database matching method combined with AI analysis for multi-drug comparisons.Quantitative data are presented as mean ± SD and were analyzed using the Mann-Whitney U test (IBM SPSS Statistics 28). Results :The system incorporates filters for specific populations, such as pregnancy trimesters and pediatric stages.Safety summaries were generated by extracting data from multiple sources.In performance testing with 20 clinical questions, search time was reduced by 68.5% (p < 0.001) compared to traditional methods.The equivalent dose search achieved a 78.3% (p < 0.001) reduction.AI filtered unrelated articles and displayed