Avoiding patient harm through the application of prescribing safety indicators in English general practices (acronym: PRoTeCT)

Background and rationale There are over 1.8 million serious prescribing errors in English general practices each year and hazardous prescribing is a contributory cause of around one in 25 hospital admissions, costing around £650 million per year. Our team has developed a set of prescribing safety indicators that can detect hazardous prescribing. These can be used in two complementary ways to: Prevent hazardous prescribing using computerised decision support when a prescribing decision is being made; Identify on-going hazardous prescriptions by searching GP computer systems to identify patients at on-going risk, so that corrective action can be taken These initiatives offer substantial scope for improving patient safety, preventing medication-related harm, and saving money for the NHS. Using a series of linked work packages (WPs) the overall aim of the PRoTeCT Programme is to evaluate two interventions aimed at improving the safety of prescribing in English general practices, and generate NHS policy recommendations. Objectives and methods WP1 will provide new evidence about the serious harm, and economic impact, of hazardous prescribing in general practice by examining existing large population-based observational studies, and undertaking one new GP-database cohort study. From this, the absolute attributable risk of serious harm associated with 11 specific types of hazardous prescribing will be determined, and Markov models used to estimate the effect of hazardous prescribing on Quality-Adjusted-Life-Years (QALYs). WP2 will evaluate the effectiveness and cost-effectiveness of a new computerised decision support (CDS) system providing point-of-care alerts to clinicians (a relatively simple low intensity intervention) in reducing hazardous prescribing. Interrupted time series analysis will be used, from 24 months before the implementation of CDS to 12 months after in at least 100 general practices English general practices that have adopted the CDS. Monthly data relating to hazardous prescribing will be obtained from ResearchOne, linked to the date each practice implemented CDS. The primary outcome will be number and proportion of patients in each general practice exposed to at least one type of hazardous prescribing; secondary outcomes will examine rates of hospitalisation and mortality. Segmented regression will estimate the changes in the level and trend over time of the outcomes, adjusting for baseline trend and known confounders. The main economic evaluation will generate incremental cost-effectiveness ratios (ICERs). WP3 will evaluate the effectiveness and cost effectiveness of the complex and intensive PINCER intervention in reducing hazardous prescribing, and patient harm when rolled out at scale (there is already excellent trial evidence that PINCER is efficacious in volunteer practices). The overall design is a multicentre stepped-wedge study in up to 530 general practices clustered in 17 participating CCGs in the East Midlands of England. The PINCER intervention comprises: Automated interrogation of GP computer systems to identify patients with hazardous prescriptions; A pharmacist-led educational outreach intervention in each practice, including agreeing on a plan for action; The pharmacist working with the practice to correct hazardous prescriptions and improve safety systems. The primary outcome is the number and proportion of patients in each general practice exposed to at least one type of hazardous prescribing. Secondary outcomes include serious harm associated with hazardous prescribing, hospital admission, and mortality. Outcomes will be measured using pseudonymised data obtained from GP computer systems suppliers. Analysis will used mixed effects logistic regression to account for clustering of patients within practices, repeated measures over time, and the implementation of the intervention at different times. A full economic evaluation will generate ICERs, a