Core Outcome Sets (COS) can help us determine which outcomes to measure and how to measure them, when assessing the benefits and risks of interventions for specific conditions. COS are developed through a systematic process, identifying an inventory of potential outcomes that are refined to a minimum set of important outcomes using consensus-building methods, involving representative groups of stakeholders, including patients. Once the COS is defined, a similar process is conducted to achieve a consensus on how to measure these outcomes and which instruments are truthful, discriminatory and feasible.1 Each stage of COS development is then disseminated through journals to researchers and other consumers.
Once developed, the consistent use of COS may benefit researchers, who will be able to compare and aggregate their data with other studies, as well as ensuring their findings are relevant to clinicians and patients. The benefits of COS have been observed in rheumatology research, with 70 per cent of intervention trials reporting all COS measures, including patient reported outcomes, developed through the Outcome Measures in Rheumatology (OMERACT) initiative.2
How can COS benefit you?
The main purpose of COS is to ensure the inclusion of consistent outcomes when designing primary studies. By including the outcomes of a COS relevant to the condition studied, researchers can be confident they have included important outcomes, and depending on the COS, be guided on how to define and measure those outcomes. Similarly, when assessing studies and systematic reviews that include the COS in their outcomes, it can be assumed that the important and relevant outcomes have been reported regarding the intervention effect.
What COS are currently available and how were they developed?
There are a number of women’s health conditions for which COS3 have been developed:
- Prevention of preterm birth (13 core outcomes)
- Pregnant women with epilepsy (29 core outcomes)
- Pre-pregnancy care for women with pre-gestational diabetes (17 core outcomes)
- Reconstructive breast surgery (11 core outcomes)
- Maternity care (48 core outcomes)
In the COS, ‘prevention of preterm birth’,4 the initial inventory of outcomes was based on a systematic review of primary outcomes used in randomised control trials.5 This was expanded on by surveys of stakeholder groups including clinicians, patient representatives and patient advocacy groups, as well as semi-structured interviews with parents. Starting with this inventory, a two-round iterative Delphi survey process was used to establish a consensus, based on a process determined a priori that defined group composition, anonymity for respondents, grading importance of outcomes, participant feedback, criteria for consensus and management of bias. The large consultation group included 337 obstetricians, 152 midwives, 175 researchers, 75 neonatologists, and a large number of parents through the advocacy groups. For each round, participants ranked the potential outcomes from limited to critical importance using a nine-point Likert scale, as recommended by GRADE working group.6 During each round, other potential outcomes could also be nominated for consideration.
The following threshold was used to define consensus:
- 70 per cent of participants scoring an outcome of ‘critical importance’, or less than 15 per cent scoring an outcome of ‘limited importance’ to determine those deemed to ‘critical’ outcomes
- The contrary for those deemed to be of ‘limited’ value
- No consensus for those falling in between
With each round the potential outcomes were refined. Finally a face-to-face meeting comprised of 23 obstetricians, 10 researchers, two neonatologists, two patient representatives and one midwife reviewed the outcome scores and determined those of critical importance, to arrive at the COS. Through this process, the 72 primary and 155 secondary outcomes identified by the systematic review, and 33 outcomes added by clinicians and patients, were refined to 13 core outcomes comprised of four maternal and nine neonatal outcomes.
- Maternal mortality
- Maternal infection or inflammation
- Pre-labour rupture of membranes
- Harm to mother from intervention
- Offspring mortality
- Offspring infection
- Gestational age at birth
- Harm to offspring from intervention
- Birth weight
- Early neurodevelopmental morbidity
- Late neurodevelopmental morbidity
- Gastrointestinal morbidity
- Respiratory morbidity
While outcomes have been determined, there is ongoing work to define and determine the instruments to measure these outcomes. Until then, researchers are encouraged to explicitly state how an outcome was actually measured and provide the definition used in each trial. There is encouraging evidence that this work is already having some effect, with 15 planned pessary trials intending to include this COS.
How can I use COS?
For those planning a study, it is worth checking the Core Outcome Measures in Effectiveness Trials (COMET) register at: www.comet-initiative.org/. This provides details about which COS have been registered, their stage of development and associated resources. The Core Outcomes in Women’s and Newborn Health (CROWN) initiative website (www.crown-initiative.org/) also has a list of related COS, with resources and contact details if you would like to become involved. CROWN Australasia7 has been launched to promote COS locally and assist researchers in becoming involved, providing guidance on local resources for those wanting to develop their own COS.
While it is early days, there is a lot to be gained by achieving agreement on what outcomes are important. The inclusion of COS will make the most of the precious research dollar and provide the greatest benefit to patients and the community.