Identifying unwarranted variation in clinical practice between healthcare providers in England: Analysis of administrative data over time for the Getting It Right First Time programme

Rationale, aims, and objectives The Getting It Right First Time programme aims to reduce variation in clinical practice that unduly impacts on outcomes for patients in the National Health Service (NHS) in England; often termed "unwarranted variation." However, there is no "gold standard" method for detecting unwarranted variation. The aim of this study was to describe a method to allow such variation in recorded practice or patient outcomes between NHS trusts to be detected using data over multiple time periods. By looking at variation over time, it was hoped that patterns that could be missed by looking at data at a single time point, or averaged over a longer time period, could be identified. Methods This was a retrospective time-series analysis of observational administrative data. Data were extracted from the Hospital Episodes Statistics database for two exemplar aspects of clinical practice within the field of urology: (a) use of ureteric stents on first emergency admission to treat urinary tract stones and (b) waiting times for definitive surgery for urinary retention. Data were categorized into 3-month time periods and three rules were used to detect unwarranted variation in the outcome metric relative to the national average: (a) two of any three consecutive values greater than two standard deviations above the mean, (b) four of any five consecutive values greater than one standard deviation above the mean, and (c) eight consecutive values above the mean. Results For the urinary tract stones dataset, 24 trusts were identified as having unwarranted variation in the outcomes using funnel plots and 23 trusts using the time-series method. For the urinary retention data, 18 trusted were identified as having unwarranted variation in the outcomes using funnel plots and 22 trusts using the time-series method. Conclusions The time-series method may complement other methods to help identify unwarranted variation.