Development, testing and validation of a taxonomy for outcomes in medical research to improve knowledge discovery

Article type
Authors
Dodd S1, Clarke M2, Becker L3, Mavergames C4, Fish R5, Williamson P1
1MRC North West Hub for Trials Methodology Research, Department of Biostatistics, Institute of Translational Medicine, University of Liverpool
2Northern Ireland Methodology Hub, Queen's University Belfast, Belfast
3Department of Family Medicine, SUNY Upstate Medical University, Syracuse, New York
4Informatics and Technology Services, Cochrane Central Executive, Freiburg
5Division of Molecular and Clinical Cancer Sciences, University of Manchester
Abstract
Background:
There is increasing recognition that insufficient attention has been paid to the choice of outcomes measured in clinical trials and chosen for systematic reviews and in particular a lack of patient input. The current lack of a standardised outcome classification system, fit for purpose, is holding back research as a result of inconsistency and ambiguity in how outcomes are described across different studies. There is also inefficiency in the searching of knowledge sources such as the Cochrane Database of Systematic Reviews, clinical trials registries, patient registries and the COMET database of core outcome sets (COS), which to date include outcomes as free-text entries only.

Objectives:
To develop, test and validate a workable outcome taxonomy, to provide high-level differentiation between outcome domains to facilitate uniformity of outcome classification in electronic databases.

Methods:
We carried out a literature review to determine existing outcome classification systems, none of which were sufficiently comprehensive or granular for classification of all potential outcomes from clinical trials. We developed a new outcome taxonomy and, as proof of principle, extracted outcomes from selected Cochrane Reviews and all published COS in the COMET database, and classified clinical trial registry entries using this new system. We have encouraged use and feedback of the taxonomy, particularly by those systematically reviewing outcomes in trials. The taxonomy has been used by nine COS developers to date.

Results:
Less than one-quarter (24%) of the annotated Cochrane Reviews included a measure of impact (function or quality of life, QoL) while physiological outcomes dominated, being present in 83% of reviews annotated to date. Seventy-eight percent of COS involving patients included a measure of life impact (global quality of life or functioning) compared to 54% of those not involving patients.

Conclusions:
We have monitored use of the taxonomy and collated feedback and common queries, to be presented here. Wider implementation of this standard taxonomy in trial and systematic reviews databases and registries will help to reduce waste in research by promoting efficient searching, reporting and classification of clinical outcomes for the first time.

Patient or healthcare consumer involvement:
Involvement in COS development.