Writing therapy: a new tool for general practice?

EXISTING RESEARCH

Almost all studies into this area have been either in an experimental or secondary care setting, with primary care almost neglected. Writing has been effective in a number of conditions, many for which stress can be a causative or exacerbating factor (Box 1).

Such results are promising, especially since many of these conditions are predominantly managed in primary care. However, writing is unlikely to be a universal panacea (Box 2), and it is possible that some participants may consider it unhelpful, unenjoyable, or even disagreeable.

PRIMARY CARE STUDIES

Only two studies exploring the use of writing in primary care have been conducted. The first was a randomised single-blinded feasibility study of 45 patients aged >65 years without a psychiatric diagnosis in a university-based geriatric/internal medicine primary care clinic in the US.17 It explored whether writing could reduce somatic and distress symptoms in older patients and found that three 20-minute writing sessions reduced the use of outpatient services and associated costs to half that of the control group, but with minimal reduction in symptoms. Another study of 41 frequent attendees in Israel found that writing led to lower symptom levels and fewer clinic visits among the writing group.18

A BROADER LOOK

The question of ‘does it work?’ and if it works ‘how well does it work?’ may be best answered by looking at the meta-analyses currently available. The latest and most comprehensive meta-analysis used 146 randomised control studies and reported encouraging results.19 All outcome types (psychological health, physiological functioning, reported health, subjective impact of the intervention, and general functioning) had positive and significant effect sizes except health behaviours, which showed a positive but non-significant benefit. The overall effect size (Cohen’s d) was 0.15, which is almost identical to that of statins in secondary prevention of cardiovascular events.20 (Cohen’s d 0.20 = small effect; 0.80 = large effect.)

Earlier smaller studies supported these findings with even more positive results. The initial meta-analysis by Smyth looked at 13 studies upon writing’s effect on healthy participants and found overall significant benefits for writing (d = 0.47).21 This effect was greater for psychological outcomes (d = 0.66) than for health or general functioning (d = 0.42) on sub-analysis. A second meta-analysis of nine randomised controlled studies (only one of which was also in Smyth’s initial analysis) explored the effect of writing on ‘clinical populations’, that is, ill people. The average effect size (d = 0.19) was markedly smaller than that found by Smyth,21 and different insofar that writing was significantly more effective for physical health than psychological health (d = 0.21 versus d = 0.07).22 These results need to be considered cautiously because of the relatively small number of studies used, with both meta-analyses using a fixed effects approach therefore potentially limiting their generalisability. Furthermore, the variable nature and methods of the studies (such as, number/duration of sessions, writing instructions, and time of follow-up assessment) included in the meta-analyses may hinder the accuracy of the conclusions.

Indeed, not all studies produced positive results. Another meta-analysis of 61 studies found no clear improvement in objectively measured physical health and most other outcomes after writing therapy.23 This was subsequently supported in another study of 30 randomised control trials which examined the long-term (4 weeks after writing) efficacy of writing with respect to somatic and psychological health.24 Possible explanations for the differences between meta-analyses and showing positive and negative results may also be due to different selection criteria between studies.

WRITING: THERAPY FOR WHICH PATIENTS?

Outcomes may vary systematically as a result of specific writing instructions, parameters of the experimental design, or the type of trauma or illness and whether it is still present. Intriguingly, there is evidence to suggest that writing may work best when you do not want it to.25,26 The content of the writing may account for some of the variation in the results of studies. For instance, men who wrote about more stressful topics obtained the most benefit from writing.27

For practical purposes, it is important to consider the optimal conditions for administering writing. It would seem that disclosure at home, in a private setting with minimal distraction, and writing about more recent events and previously undisclosed topics helps. This could potentially be combined with higher ‘doses’ of writing; at least three sessions with sessions lasting at least 15 minutes long.19

Writing appears to be a relatively safe intervention since hardly any studies have shown writing to be detrimental to health or exacerbate symptoms, but this still requires confirmation from much larger studies. Writing is generally associated with an immediate increase in short-term distress and negative affect, although this is not detrimental to participants in the longer term.28 It may be unsuitable for patients who are chronically depressed, highly disturbed, or psychotic, or those with post-traumatic stress disorder.29 Similarly, the initial psychological angst resulting from writing may be too much for some people, especially those who are unsupported. However, patients can be advised to stop writing at any point in case of severe distress and contact a healthcare professional if necessary. Clearly, the inherent nature of writing requires participants to be able to both read and write, thus excluding patients who are illiterate or children not yet able to express themselves through writing. Could a potential way of circumventing this be through expression of emotions through art?

CONCLUSION

Over 30% of consultations in general practice are wholly or partially for psychological problems and potentially amenable to writing therapy. After disclosing a problem to their GP patients may be reluctant to repeat the process to a stranger, and often want immediate help. Writing should be cheap, simple, and accessible. It may also be an effective means of reaching patients unwilling or unable to engage in counselling or conventional psychotherapy.

Because writing may be thought of as a psychotherapeutic activity, it is reasonable to compare it with other therapies aimed at the same end. Cognitive behavioural therapy (CBT) has a reported effect size of d = 0.42 in depression,30 which is larger than most of the effect sizes of writing quoted above in meta-analyses. Since CBT has a considerably higher cost in patient and professional time, writing therapy may be a rather more cost-effective and accessible alternative to CBT.

Evidence regarding the efficacy of writing therapy is varied but encouraging enough to suggest a use for it in general practice. Indeed we have found that most of a small sample of GPs in St Helens were already suggesting informal writing therapy to some patients. Some of the reasons cited included ‘enabling them to take control of their own destiny’ and ‘to allow the patient to clarify and see their thoughts’; especially if the patient volunteers that they find it difficult to express their feelings.

Much remains unknown about how durable the effects are of writing therapy and whether findings in healthy and clinical populations in secondary care can be extrapolated to primary care. However, there is enough evidence to warrant full trials of writing therapy in primary care, both for psychological problems and conditions where there may be a substantial psychological element. The next decade could see a useful new intervention for a variety of conditions that can be difficult to treat in general practice. Since writing therapy may help in a variety of problems, we may come to use it very widely indeed.31