Effect of a Multisectoral Agricultural Intervention on HIV Health Outcomes among Adults in Kenya: A Cluster Randomized Clinical Trial

Abstract

Importance: Food insecurity and HIV health outcomes are linked through nutritional, mental health, and health behavior pathways. Objective: To examine the effects of a multisectoral agriculture and livelihood intervention on HIV viral suppression and nutritional, mental health, and behavioral outcomes among HIV-positive adults prescribed antiretroviral therapy (ART). Design, Setting, and Participants: This cluster randomized clinical trial was performed in 8 pairs of health facilities in Kenya. Participants were 18 years or older, living with HIV, and receiving ART for longer than 6 months; had moderate to severe food insecurity; and had access to arable land and surface water and/or shallow aquifers. Participants were followed up every 6 months for 24 months. Data were collected from June 23, 2016, to June 13, 2017, with follow-up completed by December 16, 2019. Data were analyzed from June 25 to August 31, 2020, using intention-to-treat and per-protocol methods. Interventions: A loan to purchase a human-powered irrigation pump, fertilizer, seeds, and pesticides combined with the provision of training in sustainable agriculture and financial literacy. Main Outcomes and Measures: The primary outcome was the relative change from baseline to the end of follow-up in viral load suppression (≤200 copies/mL) compared between study groups using difference-in-differences analyses. Secondary outcomes included clinic attendance, ART adherence, food insecurity, depression, self-confidence, and social support. Results: A total of 720 participants were enrolled (396 women [55.0%]; mean [SD] age, 40.38 [9.12] years), including 366 in the intervention group and 354 in the control group. Retention included 677 (94.0%) at the 24-month visit. HIV viral suppression improved in both groups from baseline to end of follow-up from 314 of 366 (85.8%) to 327 of 344 (95.1%) in the intervention group and from 291 of 353 (82.4%) to 314 of 333 (94.3%) in the control group (P =.86). Food insecurity decreased more in the intervention than the control group (difference in linear trend, -3.54 [95% CI, -4.16 to -2.92]). Proportions of those with depression during the 24-month follow-up period declined more in the intervention group (from 169 of 365 [46.3%] to 36 of 344 [10.5%]) than the control group (106 of 354 [29.9%] to 41 of 333 [12.3%]; difference in trend, -0.83 [95% CI, -1.45 to -0.20]). Self-confidence improved more in the intervention than control group (difference in trend, -0.37 [95% CI, -0.59 to -0.15]; P =.001), as did social support (difference in trend, -3.63 [95% CI, -4.30 to -2.95]; P