Interpretation

Our findings demonstrate the COMET register to be a comprehensive and up-to-date resource for COS developers and users to identify completed COS. Not all relevant COS were identified within our search which was limited to Pregnancy and Childbirth or Neonatal health categories despite being registered within other health areas. We recommend using a broad search strategy within COMET to avoid missing relevant COS. The CROWN register was less current. We hope our findings will be used to update this important discipline-specific resource.
While core outcome set methodology is in its infancy, exponential growth in COS development signifies an overwhelming commitment to the methodology as a strategy to address variation and research wastage. The recent publishing of COS-STAD minimum standards8 and COS-STAR reporting guidelines18 indicate a positive impact on COS development over the past three years, with COS being developed with increasing rigor. Many recent COS include published and/or registered protocols and systematic reviews improving the transparency of the project. While COMET registration is not yet mandated, prospective registration of planned COS projects outlininga priori the COS-STAD criteria could further improve transparency and is recommended.
Our review found similar limitations even in the most recent COS. To be relevant COS must include outcomes relevant to patients, clinicians and other stakeholders.7 The initial list of outcomes must then be generated taking all views into account. Patient views were not included in a significant number of outcome lists. Since most COS are funded, it is important that funding be allocated to address this neglected area to ensure COS are truly relevant.
To be effective, COS must be implemented in practice.7To optimise implementation, COS should be generally relevant. Inclusive stakeholder participation is an important consideration in COS development to facilitate wide implementation. Similar to the most recent annual review 5 patient participation was not considered in some COS (participation: 71% vs 69%, respectively). In others, patient attrition rates were high despite large recruited samples. COS developers need to identify strategies to improve retention in this stakeholder group. Similar to the most recent review,5 while most participants in the current study had international representation, low and middle-low income countries were poorly represented. This was despite median country representation being significantly higher in maternal and neonatal COS compared to general COS development (Mdn = 26 vs Mdn = 10).5Without diverse representation, COS may not be relevant in these countries. COS developers should identify strategies aimed at both improving and maintaining representation from these countries to improve global applicability of COS.14
Lastly, only two of the 26 COS specific to maternal and neonatal health included recommendations for measurement of outcomes. While core outcome set development focuses primarily on outcome generation, there is a danger that a lack of attention in determining how and when outcomes are measures will continue to contribute to ongoing variation in outcome reporting. A joint initiative by the COMET and COSMIN (Consensus-based Standards for the selection of health Measurement Instruments) initiatives provides guidance for selecting outcome measures in COS.41 COS developers are encouraged to include timely recommendations to address this gap in knowledge. Strategies to harmonise outcome measurement is an important consideration and may require concerted efforts in the future to address the issue.