Jeroen van de Pol

102 Chapter 4 Strengths and limitations There were several strengths of this study. First, this economic analysis is based on the data from a large pragmatic RCT performed in daily clinical practice, which increases the generalisability of the results. Second, because this analysis was trial based, we could use the actual costs and did not use rates or price agreements. Third, we measured a broader range of healthcare costs compared to most other studies, which results in a complete overview of effects compared to costs. There were also several limitations of this study. First, due to the nature of the intervention, blinding was not feasible, which might have influenced the results of this trial. To minimise the risk of bias, all questionnaires were captured and recorded by independent research assistants. Control patients were offered a CMR after the end of the 6 months follow-up. Pharmacists are unlikely to have given extra attention to control patients, as they generally lacked time to perform additional reviews during the study period. However, it is possible that control patients could be triggered by participating in this study to consider obtaining advice about their medication, health problems or goals, but this would rather lead to an underestimation of the study results. Second, the healthcare consumption was measured by the medical consumption questionnaire by telephone interviews every 3 months. Although this is a validated method of collecting these data, it could have introduced recall bias as 3 months is a fairly long period of time. However, this bias is unlikely to be different between both groups. Also, drug dispensing records were obtained via the pharmacy information system of the community pharmacy. Medication dispensed outside this pharmacy, as well as over-the-counter drugs, could have been missed in the dataset. However, in the Netherlands, patients prefer to visit one pharmacy [45]. Finally, the follow-up period in this study was 6 months, so we do not know what the results are over a longer period. Conclusion A CMR focused on patient's preferences, personal goals and health related complaints slightly improved health related quality of life measured with EQ- VAS and slightly reduced the number of health related complaints with impact on patients' daily lives in older persons with polypharmacy, but had no effect on health-related quality of life measured with the EQ-5D-5L. Additionally a CMR could potentially be cost-saving from a societal perspective.