397 General Discussion 2.1. Implementation of medication optimisation recommendations The variability of a medication review intervention started with high variability in the applicability of CDSS-generated signals after evaluating the clinical context at the individual patient level. In the OPERAM trial, this evaluation was performed by a pharmacotherapy team consisting of a physician-pharmacist pair based on additional EHR information. The acceptance of STOPP/START signals ranged from 2.5% (i.e. initiate high-potency opioids in moderate-severe pain) to 75.8% (i.e. initiate vitamin D) for the top ten most frequently generated STOPP and START signals (Chapter 4.1) [23]. Consequently, some signals were more likely to be discussed with the attending physician and the patient than others (Box 1). Strategies to improve the predictive value of CDSS-assisted screening were reflected in the previous section about the ‘applicability of screening tools for medication optimisation’. BOX 1 Two examples to illustrate the variability in frequency and acceptance of CDSS-generated signals for medication optimisation Example 1 Benzodiazepine (BZD) use and increased risk of falls Approximately 5% of STOPP signals concerned the potentially inappropriate long-term use of BZDs (STOPP D5). Pharmacotherapy teams accepted almost two-thirds of these signals, but one-third were considered not appropriate after evaluating the individual patient level and, therefore, did not result in a recommendation for benzodiazepine discontinuation [23]. Example 2 Unindicated acetylsalicylic acid (ASA) and increased risk of (gastrointestinal) bleeding A pproximately 2% of STOPP signals concerned the discontinuation of unindicated ASA combined with an oral anticoagulant for atrial fibrillation (STOPP C5). One-third of signals were considered appropriate after evaluation at the individual patient level, while in two-thirds, ASA was still indicated and was therefore not recommended to be discontinued [23]. After detecting potentially inappropriate prescribing, the next step in the medication review process was implementing proposed medication changes in agreement with patients and their attending physicians (Chapter 4.2). Overall, 62% of OPERAM intervention patients with at least one recommendation had one or more 5
RkJQdWJsaXNoZXIy MTk4NDMw