Adaptation and evaluation of an intervention to reduce loneliness amongst adults with depression

Loneliness is highly prevalent amongst adults with mental disorders and tackling loneliness is a policy priority in the United Kingdom. This thesis builds the evidence base for an intervention to reduce loneliness amongst adults accessing NHS secondary care services for depression. Loneliness was targeted using a psychosocial intervention to increase social connectedness. The aims of this thesis were to adapt an existing intervention called Groups4Health for adults (18-65) with depression, and to evaluate the feasibility of delivering this adapted intervention in a future clinical trial. The thesis was conducted in line with the Medical Research Council (MRC) framework for designing and evaluating complex interventions and a mixed methods multiphase design was used. Four sequential studies were conducted. First, a systematic review of measures of social connectedness was conducted to identify measures for use in the feasibility study evaluation and to develop a conceptual framework of social connectedness. The systematic review found 27 studies and 21 different measures. The Interpersonal Needs Questionnaire-Thwarted Belonging measure was identified as having the strongest psychometric properties. The measures were thematically analysed to develop a conceptual framework of social connectedness. This framework comprised five dimensions: Closeness, Identity and common bond, Valued relationships, Involvement, and Cared for and accepted (giving the acronym, CIVIC). The CIVIC framework provided a theory base to guide the adaptation of Groups4Health by identifying key targets for intervention and indicating potential mechanisms of change. Second, experiences of social connectedness were explored in a qualitative interview study involving 13 adults with mental disorders and nine mental health professionals. The aim was to explore the validity of the dimensions within the CIVIC framework, and to highlight any areas of difference between the conceptual framework and the themes generated from the data. Inductive and deductive thematic analysis was used. Both the inductive and deductive analyses provided validation of the conceptual framework, with all five themes from CIVIC of Closeness, Identity, Valued relationships, Involvement, and Cared for and accepted identified. In addition, themes of Shared mental health experiences, Stigma, and Connectedness beyond social relationships were identified in the inductive analysis. Shared mental health experiences and Stigma were added to the five original dimensions of the CIVIC framework, which led to the seven dimension CIVIC-Mental Health framework. Connectedness beyond social relationships was not included as the focus in this thesis was upon human relationships. Third, three sub-studies were conducted to adapt the Groups4Health intervention manual for use with people experiencing depression, and to generate a logic model of the intervention to guide evaluation. These studies comprised: (a) a pilot intervention study of Groups4Health with stakeholders (n=4); (b) one-to-one interviews with adults with mental disorders and staff (n=22); and (c) an international expert consultation survey (n=12). Ten recommended adaptations to Groups4Health were adopted, such as having one-to-one meetings with intervention recipients before the first group session. Eight recommendations were not adopted, such as providing a support worker to accompany participants to social activities and extending the length of the intervention. The findings were used to develop a logic model, facilitator manual, and participant workbook for the modified intervention, named Groups4Health-Depression. Fourth, Groups4Health-Depression was delivered and evaluated in a feasibility study involving adults accessing secondary care services for depression (n=29). This intervention involves a one-to-one meeting followed by five group sessions delivered over a six week period. All participants completed pre-post standardised measures of loneliness, social connectedness, depression and social anxiety, and 13 participants took part in process evaluation interviews. Five participants per month were recruited into the study on average, which is an acceptable recruitment rate given limited study resources. There was a 31% dropout rate, which is comparable to related interventions. The largest effect size among the candidate outcome measures was found for the Patient Health Questionnaire-9 depression measure (d=-0.79, p