What is the best evidence for community-centred care?
Power to Change is keen to understand how community business can contribute usefully to the delivery of high quality, community-centred care.
Investigating the effectiveness of community care models can be difficult. Not only is research on the topic scarce, it can also prove difficult to identify and assess relevant studies. Power to Change’s ongoing review into cooperative, community-based nursing interventions has exposed some methodological difficulties that have prompted us to reconsider how literature reviews are designed in the future.
One problem involves the use of search terms. Searching for phrases like “community nursing interventions” or “community nursing models” may lead to misleading analysis. Many of the studies included do not define what is meant by ‘community nursing’. They presumably follow NHS England’s vague definition of a “diverse range of nurses and support workers who work in the community, including district nurses, intermediate care nurses, community matrons and hospital at home nurses”. Only a handful of studies provide evidence of interventions that fulfil our criteria of shared ownership, accountability to the local community, and local assets aiming to return profit. It is necessary to include alternative search terms, such as “social enterprise care”, “asset transfer care”, and “cooperative ownership care” to capture interventions that demonstrate the community business model, even if they do not use this label.
Additionally, the quality assessment system used for this study has proved problematic. The review replicated the approach used for Power to Change’s previous research into mental health provision.
For primary studies, the system gives equal weighting to four principles of primary research: conceptual framing, methodological transparency, internal and external validity, and relevance to the research topic. For secondary studies, equal weighting is given to topic relevance, transparency, internal validity, use of the Cochrane protocol (an internationally-recognised standard for evidence-based healthcare research), and reliability of conclusions drawn. Only studies that scored below 50% were excluded.
This system is loosely based on the Maryland Scientific Method Scale. Introduced by Farrington et al. in their 2002 book on criminology research entitled “Evidence-Based Crime Prevention”, this scale rates studies from Level 1 (basic cross-sectional comparative studies with no control variables) to Level 5 (involving Randomised Control Trials and extensive evidence on the suitable comparability of treatment and control groups).
For our purposes, this system places too much importance on methodological rigour and insufficient importance on relevance. For the review on community nursing, studies that are only partially relevant or irrelevant to the hypothesis – for instance, interventions that do not have any of the characteristics of a community business model – can still score high enough for inclusion if their methodologies are sufficiently thorough. Similarly, whether secondary studies observe field-specific standards like the Cochrane protocol is deemed as important as whether they directly address the research question. This prioritising of the Cochrane protocol seems especially inappropriate in this case, given that only one study has observed the principle so far.
While the Maryland scale may be effective for topics that are widely and deeply researched, it is perhaps ill-suited for research into the community business model. While Farrington et al. proposed the system be used for assessing large-scale, generalisable interventions, community-business interventions are too rare and too context-specific to warrant a research design involving randomized control trials and extensive evidence on the suitability of control groups.
Furthermore, given the narrow and specific conditions that make interventions ‘community businesses’, it may be necessary for future research to embed an assessment of how closely studies observe the definition of community businesses. When evidence is thin, judging field-specific methodological standards like the Cochrane protocol is perhaps less important than interrogating the relevance of the study to the research question.
Finally, it may be worth considering how much a room an inclusion assessment system should allow for the use of researcher’s discretion. It may be that the best way to identify appropriate research for community business is similar to identifying pornography: researchers know it when they see it.
Alex is a freelance researcher based in London. He graduated from Cambridge University with a BA in History in 2014 and an MA from Yale in 2017. Follow: @alexdefroand, https://www.linkedin.com/in/alexdefroand/