Development and pilot testing of PHARAO—a decision support system for pharmacological risk assessment in the elderly

European Journal of Clinical Pharmacology https://doi.org/10.1007/s00228-017-2391-3 PHARMACOEPIDEMIOLOGY AND PRESCRIPTION Development and pilot testing of PHARAO—a decision support system for pharmacological risk assessment in the elderly Ylva Böttiger 1 & Kari Laine 2,3 & Tuomas Korhonen 2,3 & Janne Lähdesmäki 2,4 & Tero Shemeikka 5 & Margaretha Julander 5 & Maria Edlert 5 & Marine L. Andersson 6 Received: 29 August 2017 / Accepted: 27 November 2017 # The Author(s) 2017. This article is an open access publication Abstract Purpose The aims of this study are to describe the development of PHARAO (Pharmacological Risk Assessment Online), a decision support system providing a risk profile for adverse events, associated with combined effects of multiple medicines, and to present data from a pilot study, testing the use, functionality, and acceptance of the PHARAO system in a clinical setting. Methods About 1400 substances were scored in relation to their risk to cause any of nine common and/or serious adverse effects. Algorithms for each adverse effect score were developed to create individual risk profiles from the patient’s list of medication. The system was tested and integrated to the electronic medical record, during a 4-month period in two geriatric wards and three primary healthcare centers, and a questionnaire was answered by the users before and after the test period. Results A total of 732 substances were tagged with one or more of the nine risks, most commonly with the risk of sedation or seizures. During the pilot, the system was used 933 times in 871 patients. The most common signals generated by PHARAO in these patients were related to the risks of constipation, sedation, and bleeding. A majority of responders considered PHARAO easy to use and that it gives useful support in performing medication reviews. Conclusions The PHARAO decision support system, designed as a complement to a database on drug-drug interactions used nationally, worked as intended and was appreciated by the users during a 4-month test period. Integration aspects need to be improved to minimize unnecessary signaling. Keywords Adverse drug reaction . Clinical decision support system . Drug interactions . Polypharmacy . Elderly * Ylva Böttiger 1 Division of Drug Research, Department of Medical and Health Sciences, Linköping University, 581 83 Linköping, Sweden 2 Medbase Ltd., Turku, Finland 3 Department of Pharmacology, Drug Development and Therapeutics, University of Turku, Turku, Finland 4 Division of Clinical Neurosciences, Turku University Hospital and University of Turku, Turku, Finland 5 Department of E-health and Strategic IT, Stockholm County Council, Stockholm, Sweden 6 Division of Clinical Pharmacology, Department of Laboratory Medicine, Karolinska Institutet, Karolinska University Hospital, 141 86 Stockholm, Sweden Introduction Background The number of drugs per patient has increased steadily over many years, especially in the elderly, as new treatment options have become available and patients live longer with chronic disease [1–4]. Even though the oldest part of the population is underrepresented in clinical trials, it can in many cases be assumed that the benefit of a given treatment, as measured by numbers needed to treat, increases with increasing age, in parallel to the increased risk of illness, as has been shown for, e.g., stroke prevention in atrial fibrillation [5]. However, with the increasing number of drugs follows an increased risk for unwanted adverse effects due to drug-drug interactions and additive pharmacological effects, leading to both patient suffering and healthcare costs [6–11]. Together, Eur J Clin Pharmacol the need for a careful, patient-oriented risk-to-benefit analysis for each treatment increases with increasing age. Decision support systems are important tools to assist healthcare personnel, as well as patients themselves, in this work [12, 13]. Previous experience of decision support systems We have previously developed a drug-drug interaction database, Sfinx, to be integrated in computerized clinical decision support systems [14]. This system is now in use in most electronic health record systems in both Sweden and Finland, as well as in several other countries. To decrease the risk of alert fatigue, due to the system signaling too often, it has been an ambition from the start of the Sfinx project to have a restrictive approach to what drug pairs should be included. Still, the database contains more than 20,000 possibly interacting drug pairs (based on 1427 active substances). Rationale for the new decision support system It has often been difficult to clearly define and act upon the limits on what interaction pairs to include into Sfinx. This is probably one of the main reasons for the great variation of content when comparing different drug-drug interaction databases [15–18]. We have also experienced a pedagogical challenge in explaining to prescribers the abilities, as well as the limits, of the database. Furthermore, for the prescriber, there is the problem of more than two drugs interacting, which will not be directly evident from any drug interaction database. To meet some of these challenges and as a complement to Sfinx, we have developed the PHARAO system for Pharmacological Risk Assessment Online. The PHARAO database contains a scoring of the risk of any of nine common and/or serious adverse effects for all substances also included in Sfinx, allowing for an evaluation of the overall risk from the patients’ complete medication list. The aim of the present work We describe the development of the PHARAO system and present data from a pilot study, testing the use, functionality, and acceptance of PHARAO in a clinical setting. We, thus, hope to support not only the rational use of drugs, but also the rational use of decision support systems. orthostatism, constipation, anticholinergic side effects, serotonergic side effects, nephrotoxicity, QT prolongation/arrhythmia, and seizures. These properties were chosen with focus on adverse events relevant in the elderly patient. The choice was based on the clinical experiences in the Sfinx working group and the challenge of addressing pharmacodynamic effects in the format of pairwise drug interactions, as well as comments and questions received from end users of the Sfinx database. Thus, the aim was for the two systems to complement each other. A standard set of literature sources for the evaluation and scoring of substances was determined, including pharmacological handbooks, summary of product characteristics, scientific evaluations available through EMA (European Medicines Agency), and relevant review articles retrieved through PubMed. All 1427 substances included in the Sfinx database at the time were scored from 0 (for no pharmacological effect) to 3 (for a strong pharmacological effect) for eight of the nine properties, according to a standard operating procedure. The scorin (...truncated)