432 CHAPTER 6 indication (STOPP A1) was most frequently generated (28%), with over half of these signals accepted (54%). The acceptance of signals was highly variable, ranging from 2.5% to 75.8% for the ten most frequently generated STOPP and START signals. These findings emphasised the importance of an expert team’s involvement in translating population-based CDSS-generated STOPP/START signals to individual patients, as more than half the signals for potential overuse, underuse and misuse were not deemed clinically appropriate in a hospital setting. In addition, we found that country of the participating trial site was the strongest predictor of signal acceptance, while patient-related characteristics were poor predictors of acceptance. Second, the patients’ andphysicians’ agreementwithSTOPP/START-based individualised medication optimisation recommendationswere assessed in Chapter 4.2. For this study, only data from the Dutch intervention arm of the OPERAM trial were used. In total, 371 recommendations for 139patientswere discussedwith patients and attendingphysicians. The overall agreement was 62% for STOPP and 61% for START recommendations. The highest agreement was found for initiating osteoporosis agents and discontinuing proton pump inhibitors (both 74%). Factors associated with higher agreement in multivariate analysis were female gender (+17% [95% CI 3.7–30.4]), ≥1 falls in the past year (+15% [95% CI 1.5– 28.5]) and renal impairment (i.e. eGFR 30–50 ml/min/1.73 m2; +18% [95% CI 2.0–34.0]). The main disagreement (40%) was patients’ reluctance to discontinue or initiate medication. Moreover, the reasons for disagreement differed per drug class. For instance, the disagreement to discontinue benzodiazepines or z-drugs was mostly (91%) due to patient reluctance because of self-reported dependence or because patients argued that side effects (falls or sleepiness) were absent. In contrast, the most important reason for cardiovascular drugs was ‘physician does not agree or does not feel qualified to advise’ in 30% of cases of disagreement. Therefore, we concluded that better patient and physician education regarding pharmacotherapy’s benefit/risk balance, with more precise and up-to-date medical records to avoid irrelevant recommendations, would likely result in higher adherence to future pharmacotherapy optimisation recommendations. In Chapter 4.3, the detectability of medication errors with the in-hospital medication review in the year prior to a potentially preventable drug-related hospital admission was assessed. In total, 84 of 963 OPERAM intervention patients (8.7%) were readmitted within one year after the medication reviewwith a potentially preventable drug-related admission, of which 72 patients (n=77 medication errors) were eligible for analysis. We found that the prior in-hospital medication review did not address medication errors identified at readmission because either these MEs occurred after the medication review (~50%), no recommendation was given during the medication review (~25%) or the recommendation was not implemented (~25%).
RkJQdWJsaXNoZXIy MTk4NDMw