Influence of methodology upon the identification of potential core outcomes: recommendations for core outcome set developers are needed

The linked study by Herman and colleagues mapped the availability of outcomes in studies of heavy menstrual bleeding, identifying widespread variation including 12 different primary outcomes, measured by an array of different definitions and instruments (Herman et al. BJOG;123:1593–8). They found the most commonly reported primary outcome, blood loss, to be measured in five different domains and by eight different methods. This heterogeneity would doubtless be worse had around 40% of trials initially identified not have been excluded for a lack of primary outcome or power calculation. The primary outcome should be the outcome of greatest therapeutic importance (Schulz et al. BMJ 2010;340:c332). Researchers may need to make pragmatic decisions when designing trials. Funding and time limitations may mean outcomes with higher event rates that are easy to measure are more attractive, increasing the statistical power of the trial at the expense of relevance for patients. Researchers may be unable to select otherwise appropriate outcomes because of the lack of objective definitions or validated instruments. The authors advocate the development of a core outcome set for heavy menstrual bleeding trials. Core outcome sets are agreed minimum sets of outcomes that can be measured in a standardised manner and reported consistently in the final publication. A core outcome set is developed in three broad stages: (1) identifying potential core outcomes; (2) determining core outcomes using robust consensus methods engaging key stakeholders; and (3) determining how core outcomes should be measured (Williamson et al. Trials 2012;13:132). In our speciality, the Core Outcomes in Women's and Newborn Health (CROWN) initiative, a consortium of 76 women's health journals, encourages healthcare professionals, researchers, and patients to develop such outcome sets (Khan. BJOG 2014;121:1181-2). However, to our knowledge, there is limited guidance for the most appropriate methods to develop core outcome sets (Williamson et al. Trials 2012;13:132). For example, in the absence of a standardised approach, different researchers have designed their systematic review using different methods, perhaps limiting the search strategy by date or language, limiting included studies by size or methodological quality, or only extracting primary outcomes from trial reports, decisions that are rarely justified. It is interesting to consider the potential effect on a subsequent outcome set of limiting a systematic review to primary outcomes as has been done in the linked study. Including women-orientated outcomes is vital. Amenorrhoea is not necessarily a desirable outcome for women with heavy menstrual bleeding. Quality of life — potentially relevant for all — was reported by less than a tenth of included studies. Furthermore, no harms or adverse events were reported as primary outcomes, limiting a healthcare professional or patient's ability to considering the trade-offs between benefits and harms of therapeutic interventions. Given the uncertainty in core outcome set development methods, further methodological research is urgently required. A research agenda could be designed through the CROWN initiative to ensure that future core outcome sets developed across our speciality are robust. Full disclosure of interests available to view online as supporting information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.