Half of people with HIV can start treatment on same day

Half of those attending HIV clinics in South Africa and Kenya were eligible to start HIV treatment after simple checks that could be performed by any healthcare worker, and same-day treatment initiation modestly improved the uptake of antiretroviral therapy (ART), results of a randomised trial published in the journal PLOS Medicine show.

However, the study also found that same-day treatment initiation was not associated with better retention in care eight months after starting treatment. In both study arms, 43% of people who started treatment had been lost from care after eight months in Kenya, and outcomes were similarly poor in South Africa.

Same-day treatment initiation has been recommended wherever possible by the World Health Organization, but it may not be suitable for everyone. For example, people who have a history of prior antiretroviral treatment may require further investigation to avoid prescribing suboptimal treatment. Symptoms of TB such as cough and fever will require further investigation. Others may not be ready for treatment for psychological reasons.

Clinics need to be able to differentiate between people who can start treatment immediately and those who need further investigation, ideally without the need to use point-of-care tests that might increase the cost of treatment.

The Simplified Algorithm for Treatment Eligibility (SLATE) study was designed to test whether using a checklist of questions and physical symptoms could identify people who were ready to start treatment immediately, and whether same-day treatment initiation using this algorithm resulted in better uptake of treatment and retention in care.

The study was carried out in three primary care clinics in Johannesburg, South Africa, and three HIV outpatient clinics in western Kenya. People aged 18 and over were eligible to join the study if they presented at one of the clinics to have an HIV test, enrol in care or resume HIV treatment. The study excluded pregnant women as HIV treatment during pregnancy was managed through antenatal clinics.

Participants were randomised individually to standard-of-care treatment initiation according to national guidelines or to receive the SLATE intervention.

Participants assigned to the SLATE arm were assessed by a nurse or clinical officer for:

• Symptoms, including TB symptom screen
• Medical history, especially prior ART history, current TB treatment, concurrent medications or substance use
• Physical examination
• Assessment of readiness to start treatment.

If the screen did not identify any reason why the patient required further investigations or additional care, 28 days of medication was dispensed in South Africa and 14 days of medication in Kenya for an immediate start on treatment.

The study enrolled 600 participants in South Africa and 477 participants in Kenya. Sixty-three per cent of participants in South Africa and 58% in Kenya were female. In the intervention arm, the median CD4 cell count was 275 cells/mm³ in South Africa and 272 cells/mm³ in Kenya, around 60% in both South Africa and Kenya had CD4 counts below 350, and approximately 20%  in both settings had very advanced HIV disease (CD4 < 100) at the time of enrolment.

Around half of participants were attending the enrolment site to get an HIV test (50% in South Africa and 48% in Kenya). However, most of the remaining people had been diagnosed with HIV previously, and one-third of participants in the intervention arm had been told previously that they were eligible for ART. Sixty-four per cent attending in South Africa and 53% attending in Kenya expected to start ART that day or on the day they were enrolled into the study and almost all those randomised to both study arms wanted to start treatment on the same day.

Half of participants in the intervention arm were classified as eligible for same-day treatment initiation in South Africa; 73% of those deemed ineligible had one or more TB symptoms.

At Kenyan trial sites, 55% in the intervention arm were eligible for same-day treatment initiation; 85% of those deemed ineligible had one or more TB symptoms.

"The study was designed to test whether using a checklist of questions and physical symptoms could identify people who were ready to start treatment immediately."

At South African trial sites, 78% of the intervention arm started treatment within 28 days compared to 68% of the standard-of-care arm, and 54% in the intervention arm started treatment on the day of enrolment compared to 11% in the standard-of-care arm.

At Kenyan trial sites, 94% of the intervention arm started treatment within 28 days compared to 89% in the standard-of-care arm, and 70% in the enrolment arm started treatment on the day of enrolment compared to 54% in the standard-of-care arm.

Participants randomised to the intervention arm were almost five times more likely to start treatment on the same day in South Africa, but only 30% more likely in Kenya. Overall, the effect of the intervention on the uptake of antiretroviral treatment within 28 days was very small – treatment initiation within 28 days was 10% higher in South Africa and 6% higher in Kenya.

Furthermore, the study found no difference in retention in care eight months after starting treatment between the intervention group and the standard-of-care group. Eight-month retention was “abysmal”, the investigators comment. In South Africa, 54% of the intervention group were defined as in care at month 8, compared to 48% in the standard-of-care group. In Kenya, 57% in each study arm were in care at month 8.

“We speculate that, for some minority of patients, the offer of same-day ART initiation simply shifts the point of attrition from before to after starting ART," they say. "These patients simply do not wish to be treated, at least at the time of the offer; they may return to care later (and likely sicker) or not at all."

But they acknowledge that because they relied on routinely collected data, they don’t know what proportion of patients missing at eight months had transferred to other clinics.

The investigators say that, although the study was not powered to detect modifying factors, they found a big difference in effect between participating sites. The greatest improvement occurred at the least efficient clinics, where poor record-keeping and high staff turnover interfered with tracking of patients and correct implementation of protocols. They suggest that consistent application of the algorithm has greatest impact in these settings, so future roll-out of the algorithm should concentrate on clinics that are performing poorly.