The content and delivery of psychological interventions for perinatal depression by non-specialist health workers in low and middle income countries: A systematic review

Psychological interventions delivered by non-specialist health workers are effective for the treatment of perinatal depression in low- and middle-income countries. In this systematic review, we describe the content and delivery of such interventions. Nine studies were identified. The interventions shared a number of key features, such as delivery provided within the context of routine maternal and child health care beginning in the antenatal period and extending postnatally; focus of the intervention beyond the mother to include the child and involving other family members; and attention to social problems and a focus on empowerment of women. All the interventions were adapted for contextual and cultural relevance; for example, in domains of language, metaphors and content. Although the competence and quality of non-specialist health workers delivered interventions was expected to be achieved through structured training and ongoing supervision, empirical evaluations of these were scarce. Scalability of these interventions also remains a challenge and needs further attention.

Psychological interventions delivered by non-specialist health workers are effective for the treatment of perinatal depression in low-and middle-income countries. In this systematic review, we describe the content and delivery of such interventions. Nine studies were identified. The interventions shared a number of key features, such as delivery provided within the context of routine maternal and child health care beginning in the antenatal period and extending postnatally; focus of the intervention beyond the mother to include the child and involving other family members;

Method
Studies were identified by a systematic literature search using the following strategies: (1) a database search of Ovid Medline, EMBASE and PsycINFO until December 31, 2012, was conducted to identify studies from LMIC describing interventions for perinatal depression delivered by NSHW. Search terms were adapted from another systematic review [17] and have been listed in Appendix 1. No start date was specified; and (2) cross-referencing of eligible articles to identify additional studies that met our inclusion criteria.

Inclusion criteria
Criteria for inclusion consisted of psychological treatments for perinatal depression in LMIC (according to the World Bank classification, July 2012) delivered by any type of NSHW. Studies involving women with perinatal depression, defined as a non-psychotic depressive episode or the presence of depressive symptoms that begins during pregnancy or in the early postnatal period (within 6 weeks of delivery) were included.

Exclusion criteria
Studies conducted with women with psychotic depression, depressive episode in a woman with bipolar disorder or other co-morbidities were excluded; studies on interventions involving provision by specialists (i.e psychiatrists, psychologists, psychiatric nurses, mental health social workers), and also studies conducted in high-income countries were excluded.

Data extraction
The titles and abstracts of each citation identified from the search were independently inspected by two reviewers (NC, NA) with reference to the inclusion and exclusion criteria. The potentially relevant full-text papers were accessed and independently reviewed by the two reviewers. Any disagreements were resolved by consensus and, when this could not be reached, a third reviewer (VP) adjudicated. Papers that referenced previous publications describing the details of the interventions and adaptations made were also retrieved. Data were summarised in a table based on the research questions identified for the review.

Data analysis
Thematic analysis was used to evaluate the strategies used in interventions, NSHW features and challenges encountered in intervention delivery. We followed the process of distillation [19], which is a method whereby interventions are conceptualised not as single units of analysis, but rather as composites of individual strategies, techniques, or components that can allow subsequent empirical grouping. Bernal's framework [20] was used for analysis of the nature of the cultural adaptations. The framework comprises eight dimensions that can be the targets of cultural adaptations: (1) language of the intervention; (2) therapist matching; (3) cultural symbols and sayings (metaphors); (4) cultural knowledge or content; (5) treatment conceptualisation; (6) treatment goals; (7) treatment methods; and (8) treatment context. Analysis was both deductive, consisting of pre-determined categories applied to data, and inductive (i.e. inferring themes from the coded data).

Description of the studies
After removing duplicates, the electronic search identified 1950 potential studies. The flow chart of studies from this starting point is shown in Fig. 1.
Nine studies were selected for final inclusion in this review [21][22][23][24][25][26][27][28][29]. The characteristics of the included studies are described in Table 1 , whereas the other seven studies were randomised-controlled trials. Of these, three studies used a cluster randomised-controlled design [22,27,29], whereas the remaining used individual level randomisation [21,25,26,30]. Although studies measured perinatal depressive symptoms as an outcome, in four studies this was the primary outcome whereas, in the remaining studies, the primary outcomes were the physical health of mother and infant, quality of mother-child interaction, infant weight and height, child development and HIV knowledge.

Content of interventions
Content of interventions, and that of the adaptations (presented later), were extracted either from the study papers (n ¼ 9) or from their linked papers (n ¼ 3) [30][31][32]. In four studies, the interventions were adaptations of evidence-based psychological treatments; cognitive-behavioural therapy (CBT) (n ¼ 2) [24,27], interpersonal psychotherapy (IPT) (n ¼ 1) and problem Solving therapy(n ¼ 1) [28]. In  Table 1 Characteristics of the nine studies included in the review of perinatal depression interventions provided by non-specialist health workers in low-and middle-income countries.   one study, psychoeducation was adapted for relevance to postnatal care and delivered as part of a multicomponent stepped care intervention [26]. In another study, the intervention was an adaptation of an existing preventive mother-infant intervention programme [30], and, in three studies, the intervention was developed de novo for the study [21, 22,29]. The interventions as described in the studies were distilled into different strategies, and this has been presented in Table 2.
Most studies used interventions that consisted of various strategies targeting the mother, the mother-child dyad , the family or both. Strategies in which the mother was the main target were psychoeducation, cognitive restructuring, problem solving, behaviour activation, and befriending. Psychoeducation was the key component in one Chilean study [26], in which it was delivered as part of a multicomponent intervention that included structured pharmacotherapy if needed, systematic monitoring of clinical progress and treatment compliance, further training to doctors, and specialist supervision on a regular basis. Psychoeducation consisted of information about symptoms and the likely causes, offering hope and motivating women to seek appropriate treatment. This was also described in other studies where it was delivered either in the individual or group format [21,24,25,27,28].
Cognitive restructuring was defined as becoming aware of one's thoughts to identify and label those which are helpful and unhelpful, and modify the unhelpful ones into more helpful ones, thereby improving symptoms of depression. Low-intensity cognitive restructuring using culturally appropriate pictures was used in the 'Thinking Healthy Programme' in Pakistan, where it incorporated the additional techniques of active listening, collaboration with family, guided discovery, and home work added to the routine practice of mother and child health education [27]. The Mamekhaya programme in South Africa too was adapted from CBT for relevance to prevention of mother-to-child transmission( PMTCT) services to focus on four broad topics: healthy living; feeling happy and strong; partnering and preventing transmission; and parenting [24].
Problem solving consisted of the five general stages: problem orientation, problem definition and formulation, generation of alternatives, decision making, and verification. In a Turkish study [28], training in problem solving was conceptualised as a form of self-control training; that is, the women 'learns how to solve problems' and thus discovers for herself the most effective way of responding [28]. Problem solving was used as an important strategy in the Thinking Healthy Programme, the Chilean psychoeducation intervention, and the participatory women's group intervention in India, where it addressed problems faced by mothers and their families [26-28].
Behavioural activation (i.e. increasing behaviours that give the woman a sense of effectiveness and pleasure leading to improvements in thoughts and emotions) was used in three studies either as part of a cognitive behaviour intervention or independently [24,26,27]. Non-specialist health workers, who were often mothers themselves, hence peers, performed the additional role of 'befriending', developing positive, supportive relationships with the depressed mothers, reducing their sense of isolation and providing individual assistance [22,24].
Where the target was the mother-child dyad, the interventions focused on educating parents on the child's physical health, and also included healthcare practices for both mother and child, child nutrition and help seeking [21, 22,24,25,27,29,30]. Apart from physical health, 'psychostimulation', defined as the provision of affection and warmth, responsiveness to the child, and the encouragement of autonomy and exploration, is an important aspect of perinatal care, and this is reflected in its use as the next common strategy across interventions [21, 22,24,27,30]. Psychostimulation aimed to encourage the mother in sensitive, responsive interactions with her infant and thus sensitise the mother to her infant's individual capacities and needs. In one study [30], this was adapted from a preventive intervention programme by health visitors based on the principles contained in The Social Baby published by The Children's Project. This programme was adapted by incorporating the key principles of the World Health Organization's Improving the psychosocial development of children. Another study from Jamaica [22] developed an early stimulation home-visit programme, which focused on improving child development by improving mothers' knowledge and practices of child rearing and their parenting self-esteem. The NSHWs were trained to ensure that the mothers experienced success and feelings of competence. Some studies focused on discussion of parenting issues, including the importance of praise, attention, and responsiveness as well as appropriate discipline strategies.
Interventions that targeted the family and broader social milieu included strategies such as 'activating social networks' and 'addressing interpersonal issues'. Activating social networks Local credibility and acceptability, fluency in local dialect, shared experience in norms and events impacting community, and familiarity with local idioms of distress. Therapist-patient relationship Therapist attempted to develop friendly relationships with the mothers and to empathise with their expressed concerns [22].
To ensure patient engagement in the treatment process.
Use of non-mental health workers Use of Lady Health Worker within the primary care system, nurses, community workers. Role enhancement of the non-specialist health workers was highlighted most of whom were available in the clinics and were often closely connected to local neighbourhoods [26,27].
To reduce stigma and preserve patient's privacy (especially during home visits) from inquisitive neighbours and family members. Also to make best use of already available, low cost resources.
Metaphors: the symbols and concepts that are shared by a particular cultural group Use of material with cultural relevance Designation of a 'health corner' in each house, and a 'health calendar' provided to each mother to monitor homework and chart progress.  To increase access to care and reduce participant burden. Acknowledgement of the traditions and values allowed the therapy teams entry into these families and increased the possibility of followthrough. Concepts: the way in which the presenting problem of a woman is conceptualised and communicated Skill building Problem solving was conceptualised as a form of self-control training, that is, the women ''learns how to solve problems' and thus discovers for herself the most effective way of responding [28].
To preserve congruence with cultural beliefs and physical/ somatic belief models of illness causation.
Cultural norms surrounding the concept of infancy and child care practices were taken into account with the aim of sensitising the mother to her infant's individual capacities and needs [21, 22,24]. Goals: consideration of the specific values, customs, and tradition of the woman's culture when agreeing on treatment goals Client-derived goal Focus on mother and infant health rather than maternal depression and have an a priori agenda of achieving optimal infant development through the intervention [27].
Infant care was seen as a shared responsibility and this helped engage not only the mother, but the whole family in a supportive role for the mother.

Adaptations to the interventions
Details of the cultural and contextual adaptations made to the interventions were categorised using Bernal's framework [20] and are presented in Table 3. Adaptations for language went beyond the literal translation to incorporate the use of colloquial expressions to replace technical terms, for example, using 'stress' instead of 'depression' and 'thinking healthy' instead of 'cognitivebehaviour therapy' [27]. Therapist adaptations, apart from using NSHWs most of whom were already available in the clinics and were often closely connected to local neighbourhoods, also used peers (i.e. mothers with experience in child rearing). These adaptations focused on therapist-patient matching to enhance the acceptability and credibility of the counsellor by emphasising shared experiences and awareness of local customs. The NSHWs attempted to develop friendly relationships with the mothers and to empathise with their expressed concerns [22,26,27]. The use of metaphors to increase cultural relevance took the form of using material that was culturally appropriate; for example, a health calendar to monitor homework, the use of local stories and examples with characters resembling the patient's situation and background, and the use of idioms and symbols such as feeling cups to identify and quantify the intensity of feelings [24,26,27,29]. These enabled the simplification of abstract concepts into more concrete, easy to understand terms. Cultural considerations were integrated into the content of the psychological treatment by focusing on pressing social concerns in the woman's life and addressing local customs; for example, issues related to Chinese postpartum practice 'Zou Yue Zi' ie. 'doing the month', which refers to the traditional Chinese custom of having new mothers rest for a month at home, often under the care of their mother-in-law [25]; and, in Pakistan, not expecting outdoor activities during the chilla (40day confinement of mothers after delivery) when mothers do not go out of the house [27]. Adaptations in the dimension of concepts involved addressing cultural norms surrounding the concept of infancy and childcare practices, and focusing on relevant skill building techniques such as problem solving [24,27,28]. Adaptations of goals involved development of client-derived treatment goals that were personally and culturally relevant, such as focusing on the health of the child and family unit rather than the mother. Goals were also extended beyond depression treatment; for Contextual stressors were seen as one of the major contributors to depression.
Where other caregivers (for example, fathers, grandparents) were present, they were encouraged to take part in the intervention. Focus on improving relationship and reducing conflict with husbands as well as mothers in law [22,25,27].
Acknowledges the central role of the family in the treatment process Home-made toys and books and materials in the home were used to keep the intervention low cost [22].  example, by enhancing roles of self-help group members into community advocates and focusing on women's empowerment [24,27]. Adaptations to methods such as reducing the focus on tasks requiring literacy (i.e. reading and writing) were important for ensuring applicability to lowliteracy populations. Delivery of sessions at home or over the telephone and integrating with routine healthcare visits helped to increase acceptability and feasibility of intervention delivery as well as adherence [25,26]. Adaptations to ensure that the psychological treatment fits into the patient's broader social context consisted of involving other family members in the intervention, focusing on interpersonal conflicts that may occur in joint family settings, and addressing issues related to the baby's gender (e.g. women attributed responsibility for the baby's gender to themselves especially in cultures that show preference for male children [22,25,27]) (Table 3).

Intervention delivery
Issues related to intervention delivery are presented in Table 4, including the location, duration, format of intervention delivery, and the delivery agent. As described, most of the interventions for perinatal depression were integrated into existing health programmes, such as nutrition, child health, and development programmes, positive parenting programs, PMTCT programmes, routine childbirth education sessions, and community health programmes. Most of the interventions were delivered at home (n ¼ 5) [21,22,27,28,30]. In three studies, the intervention was delivered in the clinic (though one study included an additional telephone session) [24-26] and, in one study, the intervention was delivered in a community setting other than the woman's home [29]. Five studies delivered interventions in the individual format [21,22,27,28,30], whereas four used the group format [24-26,29]. Number and frequency of sessions varied widely among the studies from three to 20 sessions delivered at weekly to monthly intervals over 6 weeks to 20 months.
To ensure that the intervention is provided competently and that it can be generalised to other situations and settings, the nature of the intervention needs to be clearly described and documented and the accuracy of implementation verified. All the studies incorporated mechanisms to promote fidelity, but these varied in form and application. Five mechanisms were used to maximise fidelity in the studies reviewed: use of a manualised intervention; attention to NSHW recruitment; NSHW training; regular supervision; and assessment of therapy quality. All the interventions were structured, involving specific content and a prescribed number of sessions or duration of programme. In three interventions, these details were documented in a programme-specific manual [21,23,27]. The remaining studies do not specifically mention the use of an intervention manual.
Five studies used paraprofessionals already working within the health system, such as community health workers, midwives and nurses, three studies recruited women who were mothers themselves [21,23,24], and hence could be considered peers, and one study used lay women from the community without specifying they were mothers [21,23,24,29]. Two studies described the characteristics of the NSHW: these were women from the local community who had no formal training, apart from that received from the study team for delivery of the intervention. In addition, they had a focused task (rather than responsibility for comprehensive community health), they were given appropriate support and supervision, and they had strong community support, all of which are regarded as essential for effective community health worker programmes. Furthermore, they were selected in consultation with the local community council [23]. The NSHW understood the sociocultural context of the women's problems and, as many were trusted 'health educators' within their community, they were able to take on the therapist's role and access the families with relative ease [27].
Non-specialist health worker training varied in length from 12 hours [26] to 4 months [23]. All studies report supervision of NSHW by specialists throughout the duration of the study, ranging in frequency and intensity from weekly [21,23,26] to monthly supervision [27], either in the individual or group format. The effectiveness of the intervention depends on the therapy quality and competence of the counselor [33], which can be assessed by various means: the evaluation of patient outcomes (e.g. symptom reduction); the assessment of individual sessions; and evaluation of standardised role plays.
None of the studies describe assessment of NSHW competence to confirm acquisition of knowledge or skills. Only one study reported assessing therapy quality [23]. In this study, written records were reviewed weekly as part of group supervision. In addition, during the pilot, the NSHW tape-recorded all sessions of their last two cases, and a random selection of these were transcribed and subjected to content analysis. The transcripts were coded by an independent rater to quantify the presence of essential counselling skills and strategies required by the programme. On the basis of ordinal ratings of global categories, the study reports that all NSHW showed at least moderate to good performance, with two of the four NSHW being rated as excellent on all dimensions, confirming therapy quality was satisfactory.

Challenges encountered
Numerous practical and cultural barriers were experienced in the delivery of the interventions. Common practical barriers were poor adherence, economic cost of home visits, and lack of private space for delivery of the interventions. Increased work pressure and low motivation of NSHW was also reported. The cultural barriers consisted chiefly of low acceptability of 'talking treatment', stigma of mental health interventions, and salience of social problems that demand appropriate attention over and above the counselling interventions. Most of these challenges were addressed by the various adaptations described above, although some challenges remain unaddressed. For example, although the use of NSHW is a potentially low-cost strategy to increase the coverage of evidence based care in LMIC, a limitation of this approach is its sustainability and feasibility when taken to scale. Continuous work with depressed and psychosocially deprived women could lead to burnout or a drop in efficiency of NSHW already carrying a number of other responsibilities. It is also important to ensure optimum quality of the intervention being delivered. If such programmes were to be implemented at a larger scale, it would be necessary to have scalable training, supervision and monitoring mechanisms to ensure NSHW competence and treatment quality. The lack of assessment methods in all but one of the reviewed studies is a major challenge in determining NSHW competence and quality of intervention delivery. The other potential challenge is the utility of this approach in settings where there are no NSHWs within the health system. In such circumstances, it is important to ensure the intervention is simple enough to be taught to volunteer peer-workers and family members. The interventions can be further adapted for telephone delivery and, with literate women, it may even be possible to have selfguided versions of the intervention. This, in addition to reducing costs, could be important in enhancing treatment adherence.

Conclusion
In this review, we included nine studies demonstrating that NSHW delivered psychosocial interventions for perinatal depression are feasible in LMIC. The interventions incorporated important features that had particular relevance to LMIC. These are community-based delivery provided within the context of maternal and child health beginning in the antenatal period and extending postnatally, focus of the intervention beyond the mother to include the child and involving other family members, attention to social problems such as domestic violence, substance abuse and HIV, and a focus on empowerment of women. The interventions had active psychotherapeutic components derived from CBT and interpersonal therapy, in addition to general supportive measures, such as empathic listening.
The key components of the interventions were two-fold: information giving and skill building. The information-giving components included parenting skills, basic healthcare practices, and information about perinatal depression and care-seeking, whereas the skill-building components included communication skills, behavioural activation, and problem solving. All the interventions were adapted for contextual and cultural relevance, and to be deliverable by NSHW. Delivery of the interventions was aided by keeping them simple and structured, for example, the low intensity cognitive restructuring delivered as part of the 'Thinking healthy program'. Ensuring adequate training of NSHW was important but not sufficient in itself to ensure optimum intervention delivery. The training was, in every instance, followed up with structured supervision either provided in the individual or group format.