The unanswered research questions relevant to preterm birth identified during this process were prioritised in the United Kingdom and Republic of Ireland by people affected by preterm birth (parents, grandparents, adults who were born preterm, and others affected by preterm birth), by a range of health professionals, and by people who were both personally affected by preterm birth and a health professional. To our knowledge this is the first such process in preterm birth. People affected by preterm birth and health professionals had many shared priorities, but our process demonstrates that on some questions they have different perspectives. Priorities may also change over time and in different settings, Hence, although the top research priorities from this process should be considered by those who plan and fund research in this area, the full list of 104 unanswered questions is also relevant to decision-making about research funding. This is particularly true if we wish to make research more relevant to those whose lives have been affected by preterm birth, and the healthcare workers who care for them.
While several of the top priorities for research are broad topics already well recognised as important, such as what is the optimum milk feeding regimen for preterm infants and prevention of infection, others are indicative of areas previously underrepresented in research; for example packages of care to support families after discharge, and what is the role of stress, trauma and physical workload in the risk of preterm birth, and are there effective ways to reduce this risk and does this influence outcome. This is in keeping with findings from previous James Lind Alliance partnerships, which suggests and highlights the value of partnership and shared decision making with an inclusive stakeholder group with balanced representation of service users and clinicians [7].
In line with the literature on consensus development [19], the strengths of this Preterm Birth Priority Setting Partnership include the large numbers of participants in the process, the range of stakeholders involved, the formality of the processes, the use of facilitators for face-to-face debate to ensure that all options were discussed and all participants had a chance to voice their views, providing feedback and repeating the judgment, and ensuring that judgements were made confidentially. The first three features applied to both the consultation and the workshop; the last applied only to the consultation. The change in priorities between the survey and the workshop deserves further investigation. Although the choice of individuals within the professional groups represented is unlikely to have made a difference to the priorities, [20] difference in status across workshop participants may have [19].
Preterm birth is associated with factors such as lower socio-economic status, ethnicity (such as African origin), and maternal age (being lower than 18 years or above 35 years) [21]. Despite implementing strategies to reach a more representative population, our respondents remained primarily white and with a relatively high proportion of homeowners, hence not representative of the population affected by preterm birth. This could limit generalisablity of these priorities to other populations. A wide range of relevant health professionals participated in the public voting, including neonatologists, obstetricians, neonatal nurses, midwives, speech and language therapists, psychologists and general practitioners; strengthening generalisablity.
Maintaining balanced representation between people affected by preterm birth and the different groups of health professionals for the final prioritisation workshop was challenging. This may have had implications for the final decisions, as happens in guideline development, where consensus development research concludes that differences in how groups are constituted (but not individual members) leads to different decisions [22]. At our workshop differences in priorities between the various professional groups contributed to the difficulty in achieving consensus for a top 10 list, and to the two ‘lost priorities’ which although ranked in the top 5 at the public vote were not included in the final top 15. The difficulty in agreeing a top 10 underlines the complexity of priority setting for research, particularly for topics such as preterm birth, which involve mother and baby, as well as their wider family. This complexity, and the differing priorities of different stakeholders, make it important to publicise the top 30 list, and the full long list of 104 questions, as well as the top 15 priorities [23].
Large changes in ranking following the public vote and the final prioritisation appeared to be related to difficulty in the perspective of people affected by preterm birth being heard in the large group session, and an imbalance between the different priorities of two key types of health professional (neonatologists and obstetricians). This was further complicated by fewer obstetricians than expected attending the workshop, and by some of the healthcare professionals also being researchers. Another element of our work, reported in detail elsewhere, was a nested observational study of how service users and healthcare professionals interact when making collective decisions about research priorities [24]. This suggested that health care professionals and service users tended to use different pathways for persuasion in a group discussion, and communication patterns depended on the stage of group development. The Steering Group had worked together for some time, and when new participants joined for the workshop communication patterns returned to an earlier stage. This may have influenced quality of the consensus.
Reporting of the process for prioritisation is therefore important for transparency, and to identify ways to improve it. Future prioritisation processes, particularly those with a similar wide range of healthcare professionals, should endeavour to anticipate potential different perspectives and mitigate any imbalance where possible, and should report voting separately by ‘service users’ and healthcare professionals. Similarly, whilst it may be appropriate to include healthcare professionals who are also researchers in prioritisation, this potential conflict of interest should be declared and taken into account.
This priority setting was limited to the United Kingdom and Ireland, and is therefore most readily generalisable to settings with a similar population and health system. Previous research prioritisation processes for preterm birth [3, 25] did not include people affected by preterm birth and were for low and middle income settings. The most recent neonatal prioritisation exercise in the UK did not include people affected by preterm birth and considered only medicines for neonates [26]. Although unanswered research questions are universal, prioritisation of these questions depends on the local values, context and setting. Nevertheless, there are common priorities across these different settings and our prioritisation process in the UK, such as prevention of preterm birth, postnatal infection and lung damage.
Failure to take account of the views of users of research (i.e. clinicians and the patients who look to them for help) contributes to research waste [27]. James Lind Alliance priority setting partnerships brings together ‘patients, carers and clinicians’ to identify unanswered research questions and to agree a list of the top priorities, (http://www.jla.nihr.ac.uk/about-the-james-lind-alliance/about-psps.htm) which can then shape the health research agenda [12,13,14]. The aim is to ensure that those who fund health research, and also those who support and conduct research, are aware of what really matters to both patients and clinicians. In our priority setting partnership, people affected by preterm birth and the different groups of health care professionals had different priorities. This underlines the importance of this paper presenting the full list of 30 questions taken forward to the prioritisation workshop, and the respective priorities of people affected by preterm birth and health professionals, as well as the long list of 104 unanswered questions sent out for public voting.