De-implementation of healthcare interventions post-COVID-19: evidence from the Evidence-Based Interventions (EBI) programme in England

Objectives: We explored the effectiveness of the Evidence-Based Interventions (EBI) programme which aimed to reduce the number of inappropriate procedures or diagnostic tests (‘interventions’) in the English National Health Service (NHS). For 12 interventions, we evaluated whether expected intervention counts fell or geographical variation was reduced across England following the publication of EBI Wave 2 guidance in November 2020. Design: We conducted a retrospective observational study utilising routine hospital data. We applied a controlled interrupted time series analysis to examine both step and trend changes in the use of EBI interventions following the publication of the EBI Wave 2. We compared geographical variation in intervention rates pre-EBI and post-EBI.

Setting: English hospitals where NHS funded elective healthcare interventions are performed.

Participants: Patients who had NHS-funded elective interventions in England from March 2018 to February 2022 inclusive. Results: Half of the 12 interventions evaluated had significantly lower counts compared with their controls by February 2022. For five of these interventions, counts were between 20% and 57% lower than expected. This was the result of either (1) a larger fall in procedure counts by the end of the pandemic period (n=4 interventions); (2) declining counts post-pandemic (n=1); or (3) a combination of both (n=1). The remaining six interventions had similar or greater than expected counts by February 2022. There was limited evidence of reductions in geographical variation in procedure rates after publication of EBI guidance.

Conclusions: Wave 2 of the EBI programme had a mixed impact on the use of the interventions targeted. There were some successes, however, following significant reduction in the use of elective care during the pandemic, half of the interventions evaluated returned to utilisation at or above predicted levels.