Time-series analyses using Hospital Episode Statistics (HES) data have been used in NIHR ARC West Midlands to track uptake of knowledge derived from surgical trials.[1] However, this methodology was used to monitor uptake across the health system as a whole, rather than to examine heterogeneity across individual providers. Walker et al, recently writing in the BMJ,[2] report on the use of a method to compare longitudinal data across general practices. They used the monthly prescribing data sets from the NHS Business Services Authority. The authors selected two types of prescribing behaviour where new guidelines prescribed a change in practice. They measured the timing of break points in prescribing behaviour and the steepness of any change in slope. They examined for heterogeneity across general practices in the English health service.

While significant change of practice in line with the guidelines was demonstrated across the service, there was considerable heterogeneity between practices and also different patterns of change. Some practices changed abruptly, while in others change was more incremental.

We propose to use these methods in our future work. The CRASH-3 trial has shown that tranexamic acid improves recovery after severe head injury.[3] We will evaluate uptake of this knowledge across the health service. We will also try to use this method in tracking uptake of knowledge from surgical trials.

Richard Lilford, ARC WM Director

References:

1. Reeves K, Chan S, Marsh A, et al. Implementation of research evidence in orthopaedics: a tale of three trials. BMJ Qual Saf. 2019.
2. Walker AJ, Pretis F, PowellSmith A, Goldacre B. Variation in responsiveness to warranted behaviour change among NHS clinicians: novel implementation of change detection methods in longitudinal prescribing data. BMJ. 2019; 367: I5205.