Summary
A three-phase consensus study was conducted with academic primary care experts to identify priorities for MLTC research. Consensus was reached on 10 priority questions relating to development of new models of care for MLTC, and methods and data. These largely reflected the core theme of discussion in the workshop: the need to deal with the complexity of this patient group.
For models of care, priorities included both development and evaluation (of new models) and examination of implementation challenges. Priorities also included outcome measurement, such as a focus on experiential outcomes (patient-centred care and reducing treatment burden), and capturing upstream efforts for reducing negative outcomes (prevention of disease and functional decline or worsening quality of life). They reflected the complex challenge of re-orienting the wider health and care system. They also included, perhaps more immediately, challenges of adapting current models of care and skill mix, and identifying and addressing training needs of the primary care workforce.25
For methods and data, priorities reflected the need for these models to adequately resource (through funding formulas) the complexity of management. There was an identified need for new (or linked) data beyond what is commonly collected within primary care, to better capture determinants of outcomes and prevention opportunities. Finally, opportunities were identified to expand use of quasi-experiments and other evaluation methods to explore what is ultimately effective, which could require new measures of relevant outcomes.
Strengths and limitations
The study used a three-stage approach to identify consensus research priorities of academic primary care. Limitations of this approach included the relatively small sample size and the modest response rate to the Delphi study. Nevertheless, the sample size is generally considered adequate for Delphi studies, where above 12 participants is likely to be subject to diminishing returns in terms of reliability.26 It was expected from the outset that the potential pool of researchers with expertise in the area would in any case be small overall. The follow-up rate in round 2 of the Delphi was high, so the sample was largely consistent between rounds. Another limitation is that responders were UK-centric, because of the funding and geographical location of the face-to-face workshop. Many of the UK participants were also affiliated with the NIHR SPCR, although this influence reduced over the phases (84% of open-ended survey responses, compared with 59% and 54% of round 1/2 Delphi responses). For all participants, researcher funding needs might have influenced responses. Nevertheless, the sample was expanded for the Delphi phase, and the final set of questions generated are likely to be relevant in many or most high-income countries. The final question set, however, does not necessarily represent all important issues in the topic area. There is an argument that creativity is what is needed, and new ideas will not necessarily reach consensus.
Comparison with existing literature
These findings complement previous MLTC priority setting work, such as the James Lind Alliance (JLA) from the perspective of older patients,4 recommendations and research priorities from the Academy of Medical Sciences (AMS),1 and the UK’s NIHR strategic framework for MLTC.27 Areas of overlap and difference have been mapped in the Supplementary Information. Five of the 10 priority questions did not clearly overlap with any of the previous suggestions. Similarities shared across the findings and all these other exercises include a focus on prevention of disease onset (or further decline, priority number 4), and the need for a coordinated response to (integrated) care across the whole health and care system (priority number 5). The AMS priorities further overlapped in recognising a need for strategies to improve MLTC management (although not specifically in primary care, priority number 2). The NIHR priorities also overlapped with the need for patient-centred care pathways (priority number 3) and to augment outcome measures to those most relevant to patients (similar to priority number 10).
In contrast with the priority findings identified here, the JLA’s priorities tended to emphasise specific outcome priorities for older patients (for example, social isolation, psychological wellbeing, independent living, and risk of falls), and specific intervention areas (for example, supporting carers, exercise therapy, and Comprehensive Geriatric Assessment). Arguably, several of these also emphasised a change to organisation of care to deliver the specific outcome or intervention.4 The AMS, on the other hand, had three priorities related specifically to clusters of conditions that did not come up in this exercise: trends and patterns (also reflected in the NIHR priorities),27 identifying clusters causing the greatest burden, and their determinants.1 Academic primary care researchers, particularly in this study’s workshop, raised potential limitations with the clinical relevance and application of cluster studies. For example, it is questionable how clusters might inform any attempt to reduce secondary care costs, where a large number of co-occurring conditions each contribute to a very small fraction for MLTC patients.28
Unlike these previous exercises, academic primary care researchers, understandably, also prioritised several methods and data questions. These addressed complexity and problems in adequately developing, evaluating, and implementing improved models of care, and being able to judge whether ‘effectiveness’ has been achieved (priorities 6, 8, and 9, which did not feature in previous exercises). Additionally, and perhaps reflecting the current pressures on healthcare systems, the specific emphasis on workforce skill mix and training (priority number 1), and resource allocation decisions (priority number 7) to support MLTC care have not featured previously.25 Ensuring these practical aspects receive attention is likely to be important for impact of research funding, policymakers, practitioners, and, ultimately, patient outcomes.
As outlined above, the priorities generated also differ quite substantially from the bulk of funded and published MLTC research to date. This perhaps also reflects the stage of the current MLTC evidence base, where the main gap is now application to build on the basic research already conducted.
Implications for research and practice
These priorities can offer funders and researchers a basis on which to build future grant calls and research plans. However, as raised in the workshop, the nature of the competitive scientific funding process might create challenges in obtaining support for novel, potentially groundbreaking, research in this area, as there might be a real or perceived higher level of risk.29 For example, there is the tension between allowing clinicians to recruit potentially eligible patients for interventions using the clinical expertise and judgement they apply in routine care, and concerns about selection bias or risks of using novel methods and data collection. Aligning research funding to the exploration of complexity might be the next frontier, and is likely to require context-specific solutions. Seeking to ignore complexity with traditional research methods that minimise variation in participants, intervention, implementation, and outcomes is no longer an option for improving systems of care or for disease prevention.30
