Lifespan of COVID-19 living guideline recommendations: A survival analysis

Abstract

Background

NICE has maintained a portfolio of coronavirus disease 2019 (COVID-19) living guidelines since March 2020. Recommendations within these living guidelines are subject to continuous surveillance and updates in response to triggers. However, the lifespan of individual living guideline recommendations and features that may impact whether a recommendation becomes out of date sooner is unknown.

Objectives

This study describes the length of time NICE COVID-19 living guideline recommendations have remained valid.

Methods

All guidelines within NICE's COVID-19 portfolio were included to determine the lifespan of living guideline recommendations. Data were collected on all recommendations that had been developed, undergone surveillance or updated between 1 March 2020 and 31 August 2022. The initial publication date, decision to update and updated publication date were extracted. Updates were labelled as major changes in evidence synthesis or minor changes without a substantial change in the evidence base. Any recommendation that had not been updated or withdrawn was censored. Survival analysis (Kaplan–Meier curve) was carried out to determine the lifespan of recommendations.

Results

Overall, 26 COVID-19 living guidelines and 1182 recommendations were included in the analysis. Living recommendations had a median survival time of 739 days (interquartile range [IQR]: 332, 781). Based on recommendation type, intervention recommendations had a shorter survival time (354 days, IQR: 312, 775) compared to diagnosis (368 days, IQR: 328, 795), patient experience (733 days, IQR: 345, 795) and service delivery (739 days, IQR: 643, 781). Within intervention type, pharmacological recommendations had the shortest survival time versus nonpharmacological recommendations (335 days, IQR: 161, 775 vs. 775 days, IQR: 354, 775). Updates were published an average of 29.12 days following a surveillance decision.

Conclusion

Within living guidelines, some recommendations need to be updated sooner than others. This study outlines the value of a flexible responsive approach to surveillance according to the pace of change and expectation of update triggers.