The goal of this randomised controlled trial is to enhance the Friendship Bench intervention with antidepressants in adults with moderate to severe depression. The main questions it aims to answer are:

1. Is the combination of the Friendship Bench with nurse-led antidepressants prescribing superior to the Friendship Bench alone?
2. What are the barriers and enablers for the prescription of antidepressants by nurses in primary care?

Type of study: Randomized controlled superiority trial

Participants will be randomly selected and allocated into the control arm or intervention arm. Participants in the control arm will receive six sessions of the Friendship Bench Problem Solving Therapy while those in the intervention arm will receive the Friendship Bench intervention plus Fluoxetine (Sertraline for breastfeeding women).

International guidelines, including the World Health Organisation's mental health gap guidelines for treatment of mental disorders in primary care, propose antidepressants for patients with moderate to severe depression alongside psychological interventions. Despite these clear recommendations, there is a scarcity of trials on the effect of antidepressants in low-income countries such as Zimbabwe. Even though every fourth primary care attendee has depression up to 90% of them go untreated, in part due to scarcity of mental health care services. To relieve limited specialist care and yet respond to the high burden of disease, treatment of depression needs to be integrated into primary care and delivered by non-specialists. The Friendship Bench in Zimbabwe uses a brief psychological intervention that is delivered by lay health workers (LHW) in primary care. It has been implemented in 200 primary health centres, with over 200'000 patients having accessed treatment since 2016. During this scale-up only 40% of the patients showed good response. This warrants further development of the Friendship Bench with special attention to patients with moderate to severe depression.

Aims and objectives: The investigators aim to enhance the Friendship Bench with antidepressants in adults with moderate to severe depression and to inform Zimbabwe's Ministry of Health on the scale-up of nurse-led prescription of antidepressants in primary care. The primary objective is to evaluate the effect of the Friendship Bench Plus antidepressants on treatment response after 3 months as compared to usual care being the Friendship Bench stand-alone. The investigators hypothesise that the Friendship Bench Plus, which combines six sessions of problem-solving therapy delivered by LHW and nurse-led prescription of antidepressants is superior to Friendship Bench stand-alone consisting of six sessions of problem-solving therapy. The secondary objective is to identify barriers and enablers for the prescription of antidepressants by non-specialists in primary care.

Methods: To achieve the primary objective the investigators will conduct an individual randomized controlled superiority trial with 1:1 allocation to intervention and control arm based on stratified and blocked randomisation. A 3-month follow-up and an extended open label follow-up at 6 months will be conducted. A total of 296 adults with moderate to severe depression (PHQ-9 ≥11) will be recruited from 12 primary health care centres in Harare, Zimbabwe. The patients in the Friendship Bench Plus arm will receive Fluoxetine (Sertraline for breastfeeding women) prescribed by general nurses. The primary outcome will be treatment response (≥ 50% improvement in the PHQ-9 at 3 months) and the key secondary outcome will be remission (PHQ-9<5 at 3 months). An intention-to-treat analyses using mixed-effects logistic regression adjusted for baseline PHQ-9 and gender as fixed effects, and recruitment site as random effect will be performed. To achieve the secondary objective, the investigators will conduct three qualitative pilot studies to identify barriers and enablers for the prescription of antidepressants by general nurses in primary care. In-depth, face to face interviews with adults presenting with moderate to severe depression in primary care who meet the eligibility criteria for the trial, focus group discussions with general nurses and LHW as delivery agents of the FB+ intervention and a Delphi stakeholder meeting with experts from academia, primary care, the Friendship Bench, general nurses and LHW as the interventionists, the Ministry of Health of Zimbabwe, the WHO and service users in Zimbabwe will be conducted.

In order to ensure trial implementation according to highest scientific and ethical standards a pilot study of upto 40 patients (depending on theoretical saturation of the qualitative data collection) will be performed during a running phase in two primary health care centers. The aim of the pilot study is to test all study-related procedures such as recruitment, data collection, safety and referral SOP's, ensure that the trial complies with Good Storage and Distribution Practices and evaluate the safety protocol including probing of referral pathways due to safety signals, operationalize follow-up procedures and refine the FB+ intervention if necessary. Depending on the results of this pilot study the antidepressant prescribing guide might be adapted taking into account reported adverse events and non-adherence to the antidepressants, Fluoxetine and Sertraline. Therefore, all patients will be allocated to the FB+ intervention and followed-up until the primary outcome assessment at 3 months. After the 3 months follow-up, all patients will be referred to Sally Mugabe Hospital for the continuation of their medications and further specialist care outside of the study. Data collection of the pilot study will be done in a separate database and data of these pilot patients will not contribute to the analysis of the subsequent trial. To identify barriers and enablers for the prescription of antidepressants by general nurses during the pilot ,qualitative data will be collected. First, semi-structured interviews with all patients at their three months follow-up on the acceptability of antidepressants will be conducted. Second, the antidepressant tapering sessions of the nurses will be audio-recorded in order to adapt the intervention if necessary before beginning the trial.

Expected Results: The investigators aim to contribute to guideline development through development of a treatment manual on how to use antidepressants in primary care through nurse led-prescription that will inform the Ministry of Health of Zimbabwe on its scale-up. The study outcomes will be presented at international conferences and publish all related articles open access in high-ranking specialist journals. The study aligns with Zimbabwe's Ministry of Health's effort to scale up the provision of mental health care through primary health care. The study is also attentive to international mental health policy as depression is a priority condition covered by the WHO and Zimbabwe is one of the WHO's target countries through its Special Initiative for mental health. Contextualising nurse-led prescription of antidepressants into the current scale-up of the Friendship Bench and the relevant capacity building component of the investigators' research will translate to improved patient care beyond the lifecycle of the study.

Inclusion Criteria:

• Adults ≥18 years
• Moderate to severe depression defined as PHQ-9 ≥ 11
• Treatment naïve to the Friendship Bench intervention at the time of recruitment
• Speaks English or Shona (local language)
• Written informed consent

Exclusion Criteria:

• Mild depression defined as PHQ-9 <11
• Psychotic symptoms (SSQ-14 probing question 5 positive and confirmation by study coordinator)
• High risk of suicide according to P4 screener
• Patient has received FB in the past 12 months
• Patient is currently under treatment (counselling, antidepressants, followed up by a psychiatrist)
• History of or presenting with end-stage AIDS
• History of or presenting with kidney failure
• History of or presenting with liver failure
• History of or presenting with serious cardio-vascular disease (including previous heart attack, stroke or arrhythmias)
• History of or presenting with cancer
• Positive urine pregnancy test
• Clear intention of pregnancy during the study period
• Not willing to use effective contraception during study period
• Unable to comprehend the nature of the study in either English or Shona (local language)

Data will be deposited in the Bern Open Repository and Information System (BORIS), maintained by the University of Bern. This digital repository is conform with the FAIR Data Principle. BORIS allows searching and is indexed by search engines. All items are stored with a unique Digital Object Identifier (DOI) that can be referenced in respective publication. It should be noted that the investigators plan to register and deposit data that relates to individual, primary publications and not the whole study database as such.

Data will be available for sharing for at least 10 years after the final data analysis.

The Principal Investigators will review requests for data sharing.