Non-specialist interpersonal psychotherapy versus fluoxetine for major depression and post-traumatic stress disorder in Kenyan public-sector primary care: a sequential, multiple-assignment, randomised trial

<h2>Summary</h2><h3>Background</h3> Mental health conditions are leading causes of disability, driven substantially by inadequate treatment access in low-income and middle-income countries. WHO identifies psychosocial and pharmacological treatments as essential services. We aimed to compare the effectiveness of first-line treatments with interpersonal psychotherapy (IPT) versus fluoxetine delivered by non-specialists, for major depressive episode and post-traumatic stress disorder (PTSD) among public-sector primary care patients in Kenya, and to assess first-line and second-line treatment sequences. <h3>Methods</h3> We conducted a sequential, multiple-assignment, randomised trial (SMART) at Kisumu County Referral Hospital (Kisumu, Kenya) Eligible participants were outpatients aged 18 years or older, positive for major depression, PTSD, or both, on the Mini International Neuropsychiatric Interview 7.0.2, and able to attend IPT and fluoxetine treatment appointments. We randomly assigned (1:1) participants to first-line treatment (stage 1) with IPT (12 weekly sessions delivered by non-specialists) or fluoxetine (6 months, 20 mg starting dose) prescribed by a non-specialist (nurse or clinical officer). In stage 2, participants not in remission from both major depressive episodes (MDE) and PTSD were randomly assigned to second-line treatment consisting of switch (IPT to fluoxetine or vice versa), or combination (IPT and fluoxetine). The primary outcome was remission of MDE and PTSD at end of stage 1 (ie, after first-line treatment; 3 months for IPT and 6 months for fluoxetine) and end of stage 2 (ie, after second-line, crossover treatment; up to 12 months) using thresholds on the Beck Depression Inventory 2 and PTSD Checklist 5. We used intention-to-treat analyses to investigate first-line treatment superiority and treatment sequences, using relative risk (RR) and ratios of RRs for interactions. The completed trial is registered at ClinicalTrials.gov, NCT03466346. <h3>Findings</h3> Between Sept 1, 2020, and Oct 15, 2021, we screened 3864 participants for eligibility. Of these, we randomly assigned 2162 (56·0%) participants in stage 1 to IPT (1082 [50·0%]) or fluoxetine (1080 [50·0%]). 1958 (90·6%) were female and 204 (9·4%) were male. Treatment adherence was 98·7% for IPT and participants assigned to fluoxetine received a mean of 126 (SD 60; 70%) of the 180 doses for fluoxetine. At the end of treatment (stage 1), 127 (12·9%) of 986 IPT participants and 89 (10·1%) of 877 fluoxetine participants had MDE (RR 0·15 [95% CI 0·13–0·18] for IPT and 0·011 [0·090–0·13] for fluoxetine), favouring fluoxetine (ratio of RRs 0·73 [95% CI 0·57–0·94]; p=0·015). At the end of treatment (stage 1), 173 (17·5%) IPT participants and 108 (12·3%) fluoxetine participants had PTSD (RR 0·20 [0·18–0·23] for IPT and 0·14 [0·12–0·16] for fluoxetine), favouring fluoxetine (ratio of RRs 0·68 [0·54–0·84] p=0·0005). For stage 2, 104 people who had initially received IPT received second-line fluoxetine, 99 people who had initially received IPT received second-line fluoxetine and IPT, 68 people who had initially received fluoxetine received second-line fluoxetine, and 61 people who had initially received fluoxetine received second-line fluoxetine and IPT. As a result, stage 2 comparisons were underpowered. In stage 2, there were no between-sequence remission differences (between-group risk ratio for major depression in those who received first-line fluoxetine 1·18 [95% CI 0·99–1·39]; p=0·091; between-group RR for PTSD in those who received first-line fluoxetine 1·19 [0·96–1·46]; p=0·14; between-group RR for major depression in those who received first-line IPT 0·96 [0·86–1·07]; p=0·56; between-group risk ratio for PTSD in those who received first-line IPT 0·90 [0·77–1·06]; p=0·25). There were 43 adverse events in the fluoxetine group—the most common were gastrointestinal distress (16 [1·5%] of 1080), sweating (five [0·5%]), dizziness (three [0·3%]), anxiety or insomnia (three [0·3%]), sedation (two [0·2%])—one in the combined treatment group (thoughts of self-harm or suicide), none in the IPT group, and no deaths. <h3>Interpretation</h3> Although fluoxetine was superior as first-line treatment for remission of major depressive episode and PTSD, no differences were observed between the groups after treatment crossover at the end of second-line treatment. Our results suggest that non-specialist mental health services that adhere to global standards can be successfully delivered via primary care in a lower-middle-income east-African setting. <h3>Funding</h3> US National Institute of Mental Health and Global Alliance for Chronic Disease.