Usually Available Clinical and Laboratory Data Are Insufficient for a Valid Medication Review: A Crossover Study

Medication reviews are essential for minimizing errors in patient care. Prior research has shown that these reviews can reduce adverse drug events. However, the impact of available clinical data on review accuracy remains unclear. No prior work had resolved how much additional information improves review quality. This uncertainty drove the current study. Existing knowledge lacks detailed analysis of how information availability affects medication review outcomes. Researchers have not fully explored the relationship between data completeness and error detection. The gap in understanding how much data is sufficient for accurate reviews motivated this work. This study aims to clarify the role of clinical data in medication review performance.

The goal was to assess how clinical data availability affects medication review accuracy. Researchers wanted to determine if more information leads to better reviews. They focused on three professional groups: pharmacists, general practitioners, and nursing home physicians. The study used three case scenarios with three evaluation points. Each participant reviewed the same cases at different data availability stages. The main question was whether additional data improves review quality. Researchers compared participants' remarks to an expert panel's findings. The study aimed to identify which errors are detectable with more information and which remain undetected.

The study involved 15 pharmacists, 13 general practitioners, and 18 nursing home physicians. Participants reviewed three cases at three different stages. At each stage, they received more clinical information. The first stage included only the medication list. The second added laboratory results and admission reasons. The final stage included full medical history. Remarks were categorized into six types: indication without medication, medication without indication, contraindications/interactions, dosage problems, double medication, and wrong medication. Each remark was scored against an expert panel's assessment. The scoring system ranged from -1 (potentially harmful) to +3 (appropriate). Researchers calculated the percentage of participants making each type of error at each stage.

At the first evaluation, the overall mean score was -1.7% for case A, 3.9% for case B, and 8.7% for case C. After adding lab results and admission reasons, scores improved to 15.0%, 19.8%, and 22.2%. With full medical history, scores rose further to 30.0%, 36.7%, and 44%. The error of missing medication where there was an indication remained high across all stages. Even with full information, this error was frequently undetected. Other error types showed improvement with more data. Contraindications and interactions were better identified with additional information. Dosage problems and double medication also improved. However, some errors remained difficult to detect regardless of data availability. The results suggest that more information helps but does not eliminate errors.

The study found that additional clinical data improves medication review accuracy. However, even with full information, a high margin of error remains. The absence of medication where there was an indication was consistently missed. Researchers propose that current review methods may not be sufficient. The authors suggest changing how medication reviews are performed. They note that some errors are easier to detect with more information. Other errors remain undetected regardless of data availability. The authors conclude that current practices may need to be re-evaluated. They propose that medication reviews should be conducted differently to achieve higher standards.