Water insecurity nearly quadruples odds of poor HIV outcomes, while hunger and partner violence compound risk in East Africa

Adults on HIV treatment in western Kenya who are experiencing water insecurity have nearly four times the odds of poor treatment outcomes, regardless of gender, researchers reported at the 26th International AIDS Conference (AIDS 2026) in Rio de Janeiro. In a separate Ugandan study presented in the same session, delegates heard that women facing intimate partner violence, food insecurity and water insecurity together acquired HIV at eighteen times the rate of women facing none of them.

"We know that HIV treatment success is not contingent only on the clinical services that people receive, but also on the conditions in which they live," said Jacquelyne Odak of Maseno University, who presented the Kenyan findings. "Yet we continue to focus primarily on biomedical determinants of success, with less attention on structural barriers. One of the strongest barriers lies outside the clinic."

Kisumu County carries one of the heaviest HIV burdens in Kenya, with one in 10 people living with HIV. Among those on treatment, 17% are not virally suppressed. It is also a water-stressed county as only 42% of the population has access to an ‘improved’ water source – in other words, one which is protected from contamination.

“We see a situation whereby availability, access, and quality constraints affect the ability or the capability of people with HIV to take medication, to hydrate, to observe hygiene, to prepare food and prevent infections, all with implications for HIV treatment outcomes,” Odak added.

Kenya study

Odak and colleagues carried out a matched case-control study among 356 adults receiving treatment at three HIV care clinics. A total of 181 cases each had at least one poor treatment outcome (an unsuppressed viral load, low CD4 count at or below 200, an opportunistic infection, or missed clinic visits). They were matched against 175 controls from the same underlying cohort but with none of the poor outcomes. Each case was matched one-to-one with a control on age, sex and time on treatment. Participants had a median age of 45 years and had been on treatment for an average of 14 years. The majority (60%) were women and over 70% were water insecure.

Water insecurity more than tripled the odds of a poor treatment outcome (adjusted odds ratio 3.66, 95% CI 2.13–6.29). Compared with people reporting no or marginal water insecurity, the odds of a poor outcome were around seven times higher at both low and moderate levels of water insecurity and 61 times higher among those with high water insecurity (OR 61.31, 95% CI 19.41-198.8).

Women had a higher predicted probability of a poor outcome than men when water secure (37% versus 25%), but once water insecurity was present, the predicted risk rose to 59% in both genders.

The session also provided evidence that structural conditions do not only shape HIV treatment outcomes, but also HIV acquisition.

Uganda study

In the second study, researchers followed almost 3,500 women aged 15 to 49 who were HIV negative in the AMBSO Population Health Surveillance (APHS) cohort, which runs across six rural, semi-urban and urban communities in Uganda’s Wakiso and Hoima districts. The findings were presented by Dr Amanda Miller of San Diego State University on behalf of lead author Alex Daama.

Almost 70% of participants were under 34 years. At baseline, 18% reported physical and/or sexual violence in the past year, 7% reported food insecurity, and 5% reported water insecurity. Four consecutive annual survey rounds allowed the team to count new acquisitions rather than prevalent ones.

There were 52 new HIV infections. Recent intimate partner violence was the strongest single predictor: women who reported it in the past year acquired HIV at more than three times the rate of women who did not (adjusted incidence rate ratio 3.29, 95% CI 1.08-9.98). Water insecurity (aIRR 2.33, 95% CI 0.54-9.97) and food insecurity (aIRR 1.97, 95% CI 0.57-6.75) were also associated with higher incidence, although neither association reached statistical significance, probably because of the relatively small number of HIV acquisitions observed.

HIV incidence rose when these vulnerabilities occurred together. It was three times higher with one vulnerability (aIRR 3.08, 95% CI 0.94-10.09), five times with two (aIRR 4.57, 95% CI 0.72-28.95) and 18 times with all three vulnerabilities (aIRR 17.71, 95% CI 1.86-168.75). Modelled continuously, each additional barrier was associated with a 2.4-fold increase in incidence (aIRR 2.39, 95% CI 1.33-4.29).

“HIV prevention strategies really need to move beyond screening for individual risk factors alone,” Dr Miller said. “Screening for several vulnerabilities at once, and prioritising women who report more than one, could be a better use of limited prevention resources, including PrEP.”

These studies suggest that addressing structural problems could improve both HIV treatment outcomes and HIV prevention, alongside biomedical interventions.