(186) Who Should Be Prioritised For Primary Care Medication Review? Consensus Validation And Point Prevalence of Explicit Prescribing Indicators

Background: While many quality indicators have been proposed to improve clinical prescribing, little work has been done to synthesize, apply, and validate case-finding indicators to prioritise patients for medication review interventions from electronic health records.

Objectives: To identify, screen and validate a set of explicit prescribing indicators that could be used to identify patients at increased risk of drug related harm within ambulatory practice and characterize the prevalence of indicators triggered.

Methods: We performed a RAND/UCLA Appropriateness Method based consensus process to identify potential case-finding indicators and operationalize those indicators in ambulatory practice within Scotland’s National Health Service. To identify indicators, we derived an initial set of candidate indicators from three validated sets: Tayside set, Nottingham set, and START/STOPP set, focusing exclusively on explicit indicators, that is, specific drug-drug, drug-disease or drug-patient interactions that increase the risk of ineffective or unsafe medication use. Then we convened an expert panel of twelve volunteer members of the Scottish government’s polypharmacy working group, including two geriatric physicians, five general practitioners, and five pharmacists and conducted a 3-round consensus rating process to generate a final list of indicators. We conducted a cross-sectional study to estimate the prevalence of the selected indicators in ambulatory practice, using electronic medical records as of April 2018 from 62 general practices within the National Health Service (NHS) region of Tayside, Scotland.

Results: A total of 207 candidate indicators were rated by the consensus panel, of which 97 indicators had a median rating of 4 or higher. For 3 of these 97 indicators, there was disagreement after three rounds of rating. Of the remaining 94 indicators, 21 were largely duplicative because they assessed a prescribing risk already reflected by another measure albeit at a lower level of risk. The final set therefore comprised 73 case-finding indicators targeting 18 types of drug-related harm. To estimate the prevalence of final indicators, 90,977 individuals who were registered in NHS Tayside before 2018 were included. The prevalence of each indicator triggered was relatively low, ranging from 0 to 273 per 10,000 patients.

Conclusions: A group of validated explicit prescribing indicators can be applied to electronic health records to assess and improve quality of pharmacologic care in ambulatory practice.