While using point-of-care assays for early infant diagnosis eliminates diagnostic delays and results in a rapid antiretroviral start, a randomised trial in Zambia found that this did not translate into treatment success at 12 months. In contrast, better clinical outcomes and better value in a Zimbabwean cost-effectiveness analysis led researchers to support the scale-up of point-of-care early infant diagnosis there.
The two studies were presented at the Conference on Retroviruses and Opportunistic Infections (CROI 2020) this month.
Infants are usually diagnosed with HIV testing performed at central laboratories. The challenges of this conventional form of early infant diagnosis (EID) are well documented, notably the significant delays in getting results to clinicians and caregivers. There are many steps in the process, including sample transport to a central laboratory, a return visit for result delivery and linkage to care or repeat testing after breastfeeding.
Devices for EID used at the point of care (POC) are simple to operate, can be used in a wide variety of settings and deliver results within 90 minutes. POC is costlier but allows for same-day results making early antiretroviral therapy (ART) start possible.
Dr Carla Chibwesha from the University of North Carolina at Chapel Hill said despite the success of Zambia’s prevention of mother-to-child transmission programme, adverse outcomes were common. She undertook a randomised trial at six district health clinics in Lusaka, Zambia to see whether a POC EID strategy improved 12-month outcomes of children living with HIV.
HIV-exposed infants were assigned between 6 and 12 weeks of age to either POC or lab-based testing (standard of care). In the lab-based arm, the trial provided a safety net of testing an archived sample should the public sector lab not return the results within four weeks.
Between 2016 and 2018, a total of 4000 HIV-exposed infants were randomised to the standard-of-care (2011) or POC (1989) arms, at a median age of 6.4 (4-12) weeks.
Almost all mothers (96%) had taken ART during pregnancy and 97% of infants took ART as prophylaxis. There were 81 infants identified with HIV (37 and 44 in the standard of care and POC arms, respectively). The rate of mother-to-child transmission was low (decreasing from 3% to 2% over the study period).
Health facilities received a same-day result for every infant in the POC arm, compared to after a median of 27 days in the standard-of-care arm. Without the safety net 40% would not have received a result. Parents and guardians received the result within two days in the POC arm, compared to a median of 32 days in the standard-of-care arm.
Overall 89% of infants with HIV started ART, 81% and 91% within six months in the standard-of-care and POC arms, respectively. The median time was 36 days in the standard-of-care arm, while almost all infants in the POC arm started the same day.
However, 15 infants died and 15 were lost to follow-up. At 12 months, 51 of the 81 infants with HIV (63%) were alive and in care, yet only 20 (25%) were alive, in care and virally suppressed. As with the preceding results, this differed little by study arm, 7 (19%) and 13 (30%) in the standard-of-care and POC arms, respectively, risk ratio: 1.56 (95% CI: 0.7-3.5). The POC strategy improved clinical outcomes by only 11% despite a greater proportion on ART.
Dr Chibwesha concluded that decisions to replace existing systems with POC EID must be site specific and guided by cost-effectiveness analyses. Investing in POC EID means weighing the costs of the new technology against improving existing EID systems. Most importantly, substantial investments are needed to strengthen paediatric HIV treatment programmes, she said.
A cost-effectiveness analysis of different strategies for EID in Zimbabwe was presented by Dr Nicole McCann of Massachusetts General Hospital, Boston. It found that integrating POC EID into current programmes would result in better clinical outcomes and be of better value than strengthening existing central lab-based EID systems.
Three strategies were modelled:
- Lab-based EID, which is the current standard of care.
- Strengthened lab-based EID, with improved sample transport, two additional lab staff and increased lab maintenance, and
- Point-of-care (POC) EID.
Programmatic and published data from Kenya and Zimbabwe were used to inform the assumptions that the model was based on.
Model inputs included the POC assay being less sensitive (able to correctly identify those with HIV) than the laboratory assay (96.9% and 100% respectively, while the assays had comparable specificity (ability to correctly identify those without HIV).
Other inputs for the lab model, strengthened lab model and POC model included:
- Probability of caregiver receiving the results: 79%, 91% and 98%, respectively.
- Result return time: 61, 53 and 0 days, respectively.
- Linkage to care among those testing positive: 52%, 71% and 86%, respectively.
- Cost per test kit: $17, $30 and $31, respectively.
- Cost per person tested (including staff and logistical expenses): $200, $220 and $240, respectively.
The researchers found that the POC model improves short-term survival and life expectancy and is a more efficient use of resources. One-year survival among children living with HIV was 67%, 70% and 76% for the lab-based model, strengthened lab-based model and POC model, respectively. Life expectancy was 21.8, 22.8 and 24.5 years respectively.
The authors calculated incremental cost-effectiveness ratios (ICER) for each strategy (the difference in life expectancy divided by the difference in lifetime costs between strategies). Judged against the lab-based standard of care, the strengthened lab-based model was ‘weakly dominated’, reflecting an inefficient use of health resources. The ICER of the POC model compared to the lab model was $870 per year of life saved, meeting the threshold for cost-effectiveness in this context (under $1600 per year of life saved).
The strengthened lab-based model returned results approximately 50 days later than POC, gave results to 7% fewer people, started 15% fewer children on ART and cost $1.50 less for each test.
The authors conclude that unless the return time for results and probability of results being returned for the strengthened lab model can match those for POC at a lower cost, programmes should invest in point-of-care early infant diagnoses rather than strengthening existing lab-based EID systems.
Full image credit: Dry blood spot test (for HIV) on an infant. PNG 2005. Photo: AusAID. DFAT photo library. Available at www.flickr.com/photos/106853342@N04/10693642694 under a Creative Commons licence CC BY 2.0.
Chibwesha C J et al. A randomized trial of point-of-care early infant diagnosis in Zambia. Conference on Retroviruses and Opportunistic Infections, abstract 133, March 2020.
McCann NC et al. Early infant diagnosis: strengthen existing systems or invest in point-of-care? Conference on Retroviruses and Opportunistic Infections, abstract 785, March 2020.