Realist synthesis

From the programme theory developed in the realist synthesis, it can be seen that further research might fruitfully be conducted in a number of areas. The primary research to date where TW (facilitated or not) is used to treat patients with LTCs appears to be highly under-theorised. Little is known about a number of aspects of these interventions. Although it is possible that a further dedicated realist synthesis might be able to make more sense of these intervention types, it is likely that greater mileage would be gained by first conducting additional primary research, particularly in patients with LTCs. This is especially the case if there is any intention to use unfacilitated EW for patients with LTCs. The existing studies demonstrate that greater attention is needed on theorising why it might be that such an intervention type would even work in the first place. In other words, in summary, better theorised TW interventions are first needed prior to undertaking any further realist synthesis. To provide guidance the following three questions are probably the most pressing areas to address with suitably designed primary research. This is because the answers to the following questions are most likely to provide data that are relevant for further programme theory refinement:

• Why do people with LTCs want, or not want, to participate in TW (of any type or form)?
• What do people with LTCs hope to get out of participating in TW (of any type or form)? How and why?
• Which has a bigger influence on outcomes in facilitated TW, the group, the facilitator or the writing technique? How and why?

Unfacilitated emotional writing

• There are large numbers of RCTs of unfacilitated TW in a variety of LTCs. The results suggest that there is no consistent and predictable benefit with this type of intervention. However, there may be some conditions where further robustly designed research with adequate sample sizes, comparing TW to usual care, might be useful, in particular, in people with substance misuse (because they may have experienced considerable trauma), and in PTSD (because there is weak evidence of possible benefit from the studies included in this review), or other areas in which people may have experienced traumatic events and have difficulty expressing themselves. If further research in unfacilitated writing is to be conducted, an explicit choice of patients and outcomes should be declared and CONSORT (Consolidated Standards of Reporting Trials) guidelines should be followed.¹⁵⁰ In view of the generally poor quality of the evidence to date, attention to statistical power and the maintenance of study quality are essential. Capturing any adverse events is important. Programme theory presented here might be a useful starting point to help researchers conceptualise unfacilitated EW. For example, if researchers are to investigate patients with COPD there needs to be a coherent and plausible reason as to why they think that FEV₁ might be affected by TW. It would be very useful if standard outcomes for the different conditions were measured [COMET (Core Outcome Measures in Effectiveness Trials)]¹⁵¹ to facilitate comparison with other interventions and length of follow-up should be considered.
• From the current research base, there are few data on patients understanding of, or expectations from, unfacilitated EW. Most of the trials appear to be explanatory and, in many cases, the researchers were at pains to withhold all information from participants about the possible nature of the intervention before they were randomised. If further research on unfacilitated EW in a health context is conducted, we recommend including qualitative studies to explore patients’ understanding of, and experience of undergoing, EW interventions.

Facilitated therapeutic writing

• An audit of the types of TW currently being used in the NHS in both primary and secondary care would be very useful.
• Further, robust research into the facilitated TW interventions that are used in clinical practice and the voluntary sector is recommended. Developmental work on the role of the facilitator, TW and group dynamics would be required. Some TW disciplines are very well developed but have not yet been formally evaluated in clinical settings.¹⁶ Programme theory presented here might be a useful starting point to help researchers conceptualise facilitated TW. Development work and feasibility or pilot studies should probably be conducted prior to full evaluation. Cluster RCTs and studies of other appropriate designs could be conducted on these interventions, evaluating patients with chronic physical or mental ill health. The comparators could be standard practice without TW or other comparable therapeutic interventions, such as relaxation CDs or reading bibliotherapy. Outcomes would be the standard clinical outcomes for the patient’s medical condition, and patient satisfaction, HRQoL and costs. The sample sizes would need to be large enough to find a potentially modest effect.