Depression in the perinatal period carries a significant burden of disease and can have negative impacts on foetuses and infants of mothers suffering from the illness. Risk factors for perinatal depression are particularly high in Low- and Middle-Income Countries (LMICs), and include partner abuse, economic insecurity, HIV, unwanted pregnancy, and food insecurity. Despite the substantial burden, there is a considerable 'treatment gap' between the need for treatment and the provision of services for perinatal depression in LMICs. Task sharing using non-specialist health workers has been recommended as a costeffective means to address this treatment gap and reduce the burden on mental health specialists in public health services. Evidence has shown moderate effects of task-shared treatments on the reduction of perinatal depression, but little is known about the processes, mechanisms and elements that lead these treatments to be effective. This thesis is nested within the Africa Focus on Intervention Research for Mental Health - South Africa (AFFIRM-SA) randomised controlled trial (RCT), which aimed to test a task-shared psychological treatment for perinatal depression in Khayelitsha, a low-income township outside of Cape Town in South Africa. The aim of this thesis was to explore the mechanisms of implementation and change of this intervention through a process evaluation. Before implementation of the intervention, qualitative research was employed to explore the idioms, symptoms and perceived causes of depression particular to perinatal women living in Khayelitsha, using semistructured interviews and a framework analysis approach. This was conducted with 12 depressed and nine non-depressed pregnant women and mothers of young babies, and 13 health care providers. These idioms and symptoms were also compared with the ICD 10 and DSM-5 criteria for major depression. The research found that local idioms used to describe depression included 'stress', 'thinking too much', being sad or unhappy, and being scared. Some of the common symptoms of depression were expressed as withdrawal and not wanting to talk, crying or sadness, poor concentration, thinking too much, fear and anxiety, stress, sleep problems, headaches, and body pain. The primary causes that women attributed to these depressive symptoms were lack of support, having an unwanted pregnancy, death of a loved one, poverty, unemployment, thinking too much, coping with a new baby, and stress. These were exacerbated by the extreme risk factors the women faced in Khayelitsha such as low income levels, poverty, partner abuse, low education levels, poor housing and living conditions, and poor health care. The findings from this research were recommended for inclusion in the development of the counselling intervention manual for the RCT. Following implementation of the AFFIRM-SA RCT counselling intervention, the trial outcome assessments found non-significant effects in the reduction of depressive symptoms on the Hamilton Depression Rating scale (HDRS) at three and 12 months post-partum, but also found significant improvements on the Edinburgh Postnatal Depression Scale (EPDS) at both time points for the intervention group, compared to the control group. The process evaluation for this thesis subsequently examined mechanisms and contextual factors that may have influenced the intervention outcome. This involved reviewing the counselling manual and conducting a grounded theory analysis of a sample of the counselling session transcripts from the intervention. The review of the counselling manual found that the structure, layout, instructions and grammar in the manual may have led to some difficulties in its interpretation and use for counsellors and participants. The grounded theory analysis included 39 participants who had completed all six sessions of the intervention (totalling 234 sessions). The use of grounded theory allowed for findings to emerge which had not been prespecified before analysis. This process began with the identification of 'open codes', which was anything that 'stood out' from the data. Following this, a secondary 'axial coding' of the data then identified four themes that encompassed all of the open codes. The themes were: therapeutic breakdowns in the counselling sessions, the adverse influence of socio-economic context on therapeutic effectiveness, reported positive outcomes, and attributes given for the reported changes. In turn, these themes could be represented by one of two 'core concepts' that characterised the processes that occurred during the counselling sessions. These were deviations from the intended counselling protocol (when things fall apart), and effectiveness of the counselling sessions (when things come together). The third level of coding, termed 'selective coding', examined the potential reasons for the deviations from protocol and the mechanisms or elements behind the attributions of the reported outcomes. Possible reasons for deviations include the original context of the development of the intervention, not fully incorporating the formative research and pilot findings, the limited skill base of the counsellors, limited training and supervision, the structure and design of the intervention, ownership by the counsellors of the intervention, the role of advice in this context, and contextually related need from the participants. This also explained potential reasons for the non-significant effects of the intervention on the HDRS. In terms of the attributions that the participants gave for their outcomes of change, many of these acted as 'mechanisms' or therapeutic elements of the counselling, and these elements were similar to previous research on common or 'non-specific' elements in the therapeutic space. These elements played an important role in participants' feelings of connection and reduction of distress, despite evidence of deviations from the counselling protocol. This was in keeping with the significant effects of the intervention on the EPDS outcomes. The thesis presents two models of processes that occurred in the intervention. The first posits that the intervention did not sufficiently disrupt the mechanisms or context that creates and perpetuates depression to enable long term shifts or significant changes in clinical depressive symptoms. The second suggests that the intervention provided a sense of connection and a subsequent 'buffer' of resilience to handle every-day stressors, but that this buffer was short-term and could not provide longer-term resilience against the extreme context of poverty, unemployment, abuse and trauma. Through a process evaluation of the design and implementation of the AFFIRM-SA intervention, this thesis presents a wide range of contextual considerations and therapeutic elements relevant to designing and implementing more acceptable and responsive public mental health interventions that aim to bring about real and sustainable change for perinatal depression in South Africa and other LMICs.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:uct/oai:localhost:11427/36846 |
| Date | 19 October 2022 |
| Creators | Davies, Claire Thandiwe |
| Contributors | Schneider, Marguerite, Lund, Crick |
| Publisher | Faculty of Health Sciences, Department of Psychiatry and Mental Health |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Doctoral Thesis, Doctoral, PhD |
| Format | application/pdf |
Page generated in 0.003 seconds