Culture and complex interventions: lessons for evidence, policy and practice

Summary

Evidence-based mental healthcare is evolving rapidly. There is a need for well-tested and effective interventions that are suited to culturally diverse populations. This editorial considers the findings from the SITARA study. There are a substantial number of implications for research, policy and practice.

Cultural psychiatry as a discipline can be traced back to the establishment of the ‘Transcultural Research in Mental Health’ newsletter edited by Erik Wittkower from McGill University in 1956. ¹ Since then, Transcultural Psychiatry has been established as the official international journal of the World Psychiatric Association's Transcultural Section. In 2007, the World Association of Cultural Psychiatry was established with affiliations from many national and international organisations. Academic and clinical investigations have focused on understanding the cultural influences on the manifestations and rates of mental illness, vulnerability and coping with mental illness, use of and access to effective psychiatric treatments, stigma and resilience, and the efficacy of indigenous and alternative healing methods. In more contemporary discourse, research usually focuses on public health perspectives, including comparison of prevalence, incidence and aetiological risk factors across cultural groups. ² There is insufficient attention to the cultural adaptation of clinical interventions and the way services are organised and delivered. ^3,4 The evidence base suggests a lack of randomised trials, yet a clear role for specialist services, for collaboration between sectors and for interventions that facilitate referrals between, into and out of services. ^3,4 Interventions identified in the world literature tend to accelerate transit through services, remove adverse pathways, and introduce new pathways. ^3,4 Furthermore, training the workforce to deliver effective interventions to culturally diverse populations is also lacking a generic and well-tested model. ⁵ There is also much controversy about separate or specialist provision, ⁶ and particular sensitivities in a harsh economic climate: should all provision be generic or should we invest in the cultural adaptation of interventions and services? Studies of interventions in culturally distinct groups show that these need not be expensive; for example, if developed in low-income countries, they have to be cost-effective so should be attractive and affordable in higher-income countries. ⁷

The SITARA trial

In this issue Gater et al attend to many of these matters by undertaking a cluster randomised pilot study of a complex social intervention adapted for depressed and socially isolated Pakistani women living in the city of Manchester, in the north of England. ⁸ The cultural factors of relevance were identified as linguistic isolation, social isolation and a lack of social support in marital and close relationships attributed to gender roles. These were all presumed to lead to a higher prevalence of common mental disorders, and self-harm and suicide risks. The social intervention included socialised groups for the delivery of activities and psychoeducation. The women travelled to and from the venue in a group taxi, accompanied by a female Pakistani transport facilitator, in order to tackle stigma and to provide some cultural sanction for the intervention; there were childcare facilities and food after each session and, given the care and trouble taken by the researchers, individuals felt obligated to attend. A training manual was developed and piloted and used to train staff in group facilitation that emphasised empathy, confidentiality and engagement. This social intervention was compared with a protocol-based antidepressant intervention based on National Institute for Health and Clinical Excellence guidelines to be delivered by individuals' general practitioners (GPs). The study piloted a standardised depression rating scale as the primary outcome and a social functioning scale as the secondary outcome. ⁸ An intention-to-treat analysis took account of the clustered design.

Although attention to environmental interventions is not new, this is the first pilot study of a complex intervention for Pakistani women who are depressed with this profile of problems. The importance of the findings lies in developing the intervention, testing the feasibility of the design and the acceptability of the intervention. The qualitative accounts suggest there were cultural barriers to attending for treatment and that these may work against referral to services. Insufficient time to talk with GPs was also proposed as a deterrent to seeking help, but stigma among families and concerns about seeking individual treatment for depression were also barriers.

The main findings were that the social intervention alone and the combined intervention (the social intervention and the antidepressant protocol), were superior at 3 months and not at 9 months compared with the antidepressant protocol alone, albeit the findings were not statistically significant. The antidepressant protocol was important and could itself have been culturally adapted to improve take up; the social intervention effects appear to be short lived for depressive symptoms although greater satisfaction was sustained at 9 months. Adherence to antidepressant prescribing at 3 months was highest for the combined intervention (71%) compared with the antidepressant protocol (54%) and with the social intervention only (18%). The social intervention may therefore be conceptualised as a mechanism to boost access to existing care by overcoming stigma and isolation, with less emphasis on producing long-term gains. Pakistani women who benefit from the social intervention then return to their usual lives. Investigating the longevity of the social networks and support beyond the trial might be important. This effect may be especially marked given the research that suggests that depression in Pakistani women is experienced as an interpersonal and social problem, rather than an individualised and embodied experience of a depression. ⁹ The literature on cultural variations in the personal identity and sense of self argues that group therapy techniques may be helpful and should be tested in the future.

The authors searched for effects at the individual level, taking account of attrition, perhaps because more definitive findings were not discerned. The trial was underpowered for a clustered design. However, pilot studies should not necessarily be identifying definitive results on which to base further study questions or provide definitive answers; ¹⁰ rather these should test appropriateness and feasibility of methods, and may give some indication of the effect size so that a future trial can be adequately powered. Similarly, assumptions about cost-effectiveness in the absence of more formal economic analyses in an adequately powered design are premature, as are assertions about which treatments will prove to be more effective in a definitive trial in the future.

Cultural adaptation of complex interventions

The challenge facing researchers in this area of study is to operationalise the terms cultural sensitivity, appropriateness and adaptation and to focus on which components of the complex intervention address which cultural barriers to access and effectiveness. These barriers can be individual beliefs and attitudes, family based, culture based and society based. Measurement of effects at family, cultural group and society levels may have captured wider benefits.

Adaptations of some interventions, for example, cognitive–behavioural therapy, may improve the attractiveness and acceptability, and hence referrals and uptake, as well as the effectiveness; there will be multiple influences on the pathway to recovery. Alternatively, some effective interventions may not be accessed because of stigma or different narratives about the cause and meaning of illness and remedies. These barriers require more public health approaches and health promotion work.

More in-depth qualitative work throughout the pilot phase may distinguish the mechanisms of action that were universally relevant from those that can only be relevant to specific cultural groups. It is possible that, over time, some cultural groups will adapt and adopt health beliefs and expectations found in society in general and in health promotion literature. Although this argues for less cultural adaptation, it is well established that acculturation itself is a stressor, and may lead to a number of outcomes, some of which include retaining a traditional and distinct cultural identity, and that these influences are patterned by gender, age and education. Future interventions should therefore be designed to remedy intersectional inequalities and address the needs of marginalised groups defined by a number of characteristics.

Trials of culturally adapted interventions risk being of value only for people from the cultures under study, and would require a series of separate trials, each as costly and adequately powered, and specifically adapted to the cultural group of interest; thus stratification of trials is not always recommended. Paradoxically, although developed as an intervention for Pakistani women, this intervention may benefit socially isolated women with depression in general. The study provides lessons for how to culturally adapt existing interventions. An important lesson for researchers is to consider the extent to which an intervention should be adapted. Practitioners will wish to scrutinise the strength of evidence of effectiveness for diverse cultural groups, including an account of the cultural adaptations and the results of field testing.