Randomised controlled trials have demonstrated positive physical and/or psychological benefits of expressive writing [1]. In the expressive writing paradigm, participants write about traumatic, stressful or emotional events in their lives for 3–5 sessions of 15–20 minutes each, over consecutive days or few weeks. Control groups write for the same time about emotionally neutral topics. Early studies with nonclinical groups (e.g., college students) found benefits in objective health outcomes (e.g., doctor visits and grade point average) and some benefits in self-reported physical and psychological health outcomes (e.g., physical symptoms and mood). Lately, researchers have investigated expressive writing as an intervention for clinical groups with physical and/or psychological health problems [2, 3]. Meta-analyses have shown significant improvements in health outcomes for both non-clinical [4] and clinical populations [5]. Despite numerous attempts to elucidate an underlying mechanism, to date no single theory can account for the benefits of expressive writing [2]. Indeed, it seems that the benefits probably reflect a host of cognitive, emotional, social and biological processes [6]. However, there is evidence that the development of a coherent narrative, which may help to reorganise and structure traumatic memories, is an important contributor [2].

Because the expressive writing paradigm is a brief, easily administered intervention with potential health benefits, this pilot study examined its acceptability and effectiveness in a population of opioid dependent patients, as they constitute a stigmatised group in whom treatment noncompliance and poor retention rates hinder the delivery of proven psychosocial treatments [7]. Since it is commonly thought that patients with substance use disorders tend to use substances to avoid thoughts and feelings about emotional issues, writing about such experiences may force them to directly confront their emotions and organise their thoughts in ways they may never have done [6]. To provide some context, dependence on drugs such as opioids is a complex, chronic disorder with high comorbidity that requires ongoing multidisciplinary management and the need for effective psychological interventions [8]. In Australia it is estimated that 3.2 percent of males and 1.3 percent of females, aged 18 years and over, meet criteria for a drug use disorder [9]. Treatment is complicated by high rates of co-existing disorders such as depressive disorders, generalised anxiety, phobias, and personality disorders requiring simultaneous management [10], by complex medical needs [11], and by issues such as homelessness, disrupted living conditions, and social isolation [12]. Methadone maintenance programs are efficacious [8], but psychosocial interventions aiming to improve individuals' medical and social functioning are particularly important for those responding less well to methadone maintenance programs. Interventions that are brief, effective, and acceptable to individuals are important both for harm reduction and to enhance the effectiveness of their stabilisation treatment [13].

Expressive writing studies have also explored the relationship between how participants write and observed health benefits using computerised text analysis software [14], to determine how and for whom expressive writing is beneficial. Pennebaker [1, 15] suggested that improved health is associated with a linguistic "fingerprint" involving higher use of positive emotion words (e.g., happy, laugh), moderate use of negative emotion words (e.g., sad, angry), and an increasing number of cognitive process words (insight words (e.g., understand, realise) and causal words (e.g., because, reason)) over writing sessions. Several subsequent studies support this relationship for positive emotion words [15, 16], although others do not [17–20]. Health benefits were associated with moderate use of negative emotion words in one study [15], higher use in another [17], but unrelated in others [16, 18–21]. Increasing use of cognitive words is associated with improved physical health in many [15–17, 22, 23] but not all studies [18, 20, 21]. Overall, the findings for positive and negative emotion words are somewhat inconsistent, and increasing cognitive word use seems more consistently related to physical health benefits in healthy samples compared to clinical samples. More recent findings suggest that variation in pronoun use (e.g., I, my) over the course of writing is strongly linked to health improvements [24], perhaps reflecting flexibility in the way people think about interrelationships among themselves, others and events.

This study explored the effectiveness and acceptability of expressive writing in high-risk drug dependent patients in a primary care clinic, in terms of psychological health (depression, anxiety and stress) and physical health (perceived physical and mental disability). The study also investigated the relationship between linguistic markers and health outcomes.

This preliminary investigation of expressive writing in a drug dependent population found no statistically significant benefits in self-reported physical and psychological health at two-week follow-up, although all outcome measures changed in the direction of improvement. Given the high-risk population, it is noteworthy that participants' health did not get worse following the intervention. There are several possible methodological explanations for the non-significant effects.

First, and most obviously, the power of the study to detect significance was limited by the small final sample size, due to high attrition (74%). Expressive writing studies usually assess health outcomes after a month or longer, however the transient nature of this population necessitated a shorter follow-up period because of the likelihood of even greater attrition. Significant health improvements may have been detectable over a longer timeframe or with a larger sample. Also, availability of resources prevented the inclusion of a control group, which would have enabled a comparison between outcomes for expressive writing and neutral writing.

The intervention's impact may have been limited by the instruction to write about a "recent stressful event", in contrast to "the most traumatic and upsetting experiences of their entire lives" generally used in expressive writing studies [1]. The high-risk nature of this population prompted a more cautious approach to minimise the possibility of adverse consequences, however, participants may have written about less emotionally salient events than in other studies, limiting the potential for health benefits. Although participants were instructed to "really let go and explore their very deepest emotions and thoughts", examination of writing tasks showed that some participants wrote about stressful events in a more descriptive manner rather than writing about their feelings or thoughts. Some participants may have avoided the more in-depth emotional and cognitive processing regarded as a critical element underlying the benefits of expressive writing [6].

Participants' preexisting health status was significantly poorer than both the general population [28] and participants in non-clinical expressive writing studies [4]. The improvements in immune functioning thought to relate to improved health after writing [34] may have had limited short-term impact in a population with poorer health and chronic conditions. In addition, it is possible that participants who simultaneously use a number of substances may have poorer outcomes, but detailed assessment of poly-substance use was beyond the scope of this study.

Finally, the availability of resources prevented access to objective measures of health and limited the number and complexity of self-report measures that could be administered. Significant benefits may have been observable on objective or other self-report measures more specific to the health problems common to this population.

We found that using more positive emotion words was significantly associated with improvements in depression and stress. This finding extends previous research linking increased use of positive emotions with improvements in physical health [15, 16]. Thinking positively is generally beneficial in improving psychological functioning [35] and optimistic individuals show better psychological adjustment [36]. Flexibility in use of first person pronouns was significantly associated with improvements in anxiety, which complements previous research showing an association between variation in pronoun use and physical health benefits [24]. Unexpectedly, increasing number of cognitive process words from first to last writing was associated with worsening depression scores. Whilst increasing cognitive word use generally predicts improved physical health, other studies have also found it to be unrelated [21] or negatively related to psychological health [18]. In this sample, increasing use of cognitive words may reflect the start of a process of dealing with emotional material that had previously been 'kept at bay' by drug use. No linguistic markers were related to SF-12 physical or mental health outcomes. The linguistic predictors of physical and psychological health are likely to be different and further exploration of these relationships in different populations is warranted.

It is nevertheless noteworthy that in a population that generally has difficulty accessing psychological interventions, expressive writing was well-received and regarded as useful by the participants themselves. Researchers and counsellors were surprised by the degree of enthusiasm with which diary writing was received, with many participants continuing to write in their diaries after the study.

In this preliminary study of high-risk drug dependent patients in a primary care clinic, a brief expressive writing intervention was acceptable and well-received by participants. Although no statistically significant improvements in self-reported physical or psychological health outcomes were found, participants' health did not worsen and linguistic analysis demonstrated some significant relationships. Given the difficulty implementing psychosocial interventions in this population and previous findings of health benefits after expressive writing, further research with a larger sample is warranted, as expressive writing may prove to be a useful intervention for harm reduction in people with substance abuse problems.