Table 1

Framework for developing a complex intervention8 applied to EoLC interventions and MORECare guidelines11

PlanningIdentify the problem, and why a change is required, its importance and what would be the benefits of a new intervention. Identify current interventions—what does and does not work. Identify resources required both in terms of time (be realistic) and funding
Understand the context (population and individuals; location or geographical setting; social, economic, cultural and political influences) where the intervention will be implemented. Understand the facilitators and barriers to reaching the relevant population
Involve stakeholdersIdentify and involve all stakeholders (commissioners of services, health professionals who deliver interventions, and patients and carers who will benefit from the intervention) throughout the development process
  • It may be appropriate to adjust eligibility criteria for participants in EoLC in order to recruit patients who may benefit the most from the intervention

The teamIt is essential the development team includes individuals with relevant expertise in relation to the problem being addressed, including stakeholders, and the methodological experts to support the collection of evidence
  • Researchers need to be familiar with the challenges in EoLC studies. Ideally involvement with stakeholders would be established prior to the study and continue beyond individual studies. It is also important to minimise the burden for clinical staff who participate

Establish the evidence-baseReview and synthesis published research evidence before starting to develop the intervention to understand the evidence base for each component of the intervention
Draw on theoryIdentify existing theory/ theories or frameworks that will inform the development of the intervention, eg, psychological and organisational theories
Develop and design the interventionCollect qualitative data to generate ideas about the content, format and delivery of the intervention. Identify outcomes and what measures exist to evaluate outcomes using quantitative methods
EoLC measures should be chosen that meet the following criteria:
  • Established validity and reliability in a relevant population

  • Responsive to change over time

  • Capture clinically important data

  • Easy to administer and interpret (concise and not complex)

  • Applicable across care settings to capture a change in outcomes by location (eg, home, hospice and hospital)

  • Able to be integrated into clinical care

  • Minimise problems of response shift (a change in conceptualisation or calibration of what is being measured)

  • In addition, and central to complex interventions is the experience of ‘care’- allowing enough time for the intervention to have had an effect is important. The plan of investigation and methods should not cause unnecessary burden on participants. The therapeutic effect of interviewing, if a component of the intervention should also be considered during evaluation

Testing the intervention Write up the intervention development process so that judgments can be made about the quality of the process. Develop a protocol for the feasibility/pilot or evaluation of the intervention
  •  Consider issues of attrition when developing protocols in EoLC, because attrition rates can influence decisions during the implementation of interventions. Estimate in advance possible reasons for attrition and missing data when calculating sample sizes and allow for data collection phases from proxies. Use the MORECare classification to describe the causes of attrition:

  •  Attrition due to death

  •  Attrition due to illness

  •  Attrition at random

  •  Consider implementation implications, eg, workforce and training needs at all stages of intervention development, testing and integration into practice.

  •  Data on the use of services needs to be collected to explore impacts on care costs

  •  Cost-effectiveness outcome measures need to be justified

  • EoLC, End of Life Care; MORECare, Methods Of Researching End of life Care.