1. Bringing the Test to the Patient: Diagnosing and treating more infants, faster
Evaluating point-of-care (POC) testing for pediatric HIV in Eight sub-Saharan African Countries
Flavia Bianchi
AIDS
Amsterdam, NL
July 24, 2018
Good afternoon. I am so pleased to present the results from EGPAF’s pre and post intervention evaluation of POC EID for HIV in 8 sub-Saharan countries .

2. Overview
Background
Why POC EID?
Optimizing EID through strategic placement: EGPAF’s programmatic approach
Methods: a pre-and post-analysis comparing conventional and POC EID
Results
Key outcomes
Hub-and-spoke
High-yield entry points
Cost per test result returned
Conclusions
Today, I wanted to start with why now is the right time for POC, providing some details on EGPAF’s strategic approach to POC platform placement and project implementation. Then I will touch on our evaluation methods before moving on to our very exciting results. I’ll end with some final observations and remarks.

3. Why POC EID? Challenges with the EID Cascade
Turnaround time from blood sample collection to return of results to caregiver: days
When we consider the timing and potential of POC, we want to consider the three main challenges affecting early infant diagnosis for HIV in Sub-Saharan Africa.
The first challenge is poor access to and delays to testing – 2016 data from Unaids shows that only about 50% of HIV exposed infants are actually being tested.
The second challenge is delayed or no return of results. The literature we looked at shows that tests can take anywhere between 30 to 90 days to be returned to the caregiver. And again the data from Unaids shows that about 50% of results are actually reaching caregiver. This long TAT ultimately leads to our third challenge which is poor treatment initiation for HIV positive infants. About 50% of HIV-infected infants will be initiated on treatment. And poor treatment initiation will eventually lead to an increase in infant mortality and morbidity.
Source: On the Fast-Track to an AIDS-Free Generation, UNAIDS, 2016

4. Implementing POC EID in Routine Clinical Care: EGPAFs’ Approach
Bringing the test closer to the client in nine project countries
Pragmatic placement of POC platforms and implementation based on current resources and human resources
Phased approach – started with a 6 month pilot period
Maximize access to POC EID testing through hub-and-spoke models and multiple entry points
WHO guidelines from 2016 stated that POC could offer considerable advantages in dealing with delays in specimen handling and turnaround times by bringing the test closer to the patient. At the time, they requested more field evaluations of commercially available POC technologies to confirm the accuracy of results and to understand how POC could fit within national programs. They also recommended further research to understand the impact of POC EID on patient management, treatment and infant outcomes.
So here we are today. In 2015 EGPAF was awarded funding by Unitaid to implement POC EID across 9 countries highlighted in our map. Our selection of sites and placement of platform was strategic in nature – we considered human resources, facility preparedness, testing needs, location etc…always working in close collaboration with the ministries of health in each country. We implemented a phased approached, starting with a 6 month pilot phase involving close monitoring before scaling up. Because we wanted to reach as many kids as possible, we formed hub and spoke networks and accepted tests from multiple entry points.

5. Hub-and-Spoke Networks Multiple-Entry-Point Sites
Strategies Used to Increase Access to Testing
Stand-Alone Sites
Receive samples directly from clients and perform POC EID tests on site.
Hub-and-Spoke Networks
Hub sites provide testing for patients at that site and for spoke sites. Nearby spoke sites send samples to the hub sites for testing.
POC EID
POC EID
Hub
Spoke
Multiple-Entry-Point Sites
Stand-alone or hub testing sites receive samples from different units or wards within the same health facility.
Nutrition unit
POC EID
Pediatric ward
MCH clinic
To optimize access, we have both standalone testing sites, where tests are processed on-site and we created the hub and spoke network, where hub testing sites receive samples from surrounding spoke sites that alone could not support the placement of a platform but when combined with other sites to form the network, it increase platform throughput. Additionally, we also receive tests from multiple entry points beyond ANC/PMTCT clinics, including from maternity, pediatric inpatient wards, vaccination clinics, outpatient clinics.

6. EGPAF POC EID Project: Evaluation Methods
Pre-intervention data – Conventional EID
Retrospectively collected for 30 consecutive HIV-exposed infants who had a sample collected in 8 project countries
Data collected from facility registers from a subset of intervention sites
Purposive sampling of sites (up to 20 sites per country)
Tests conducted between March 2014 until May 2017.
Sample: 2,899 tests from 2,875 HIV-exposed infants from 96 sites
Post-intervention data – POC EID
Collected prospectively in all POC EID sites in 8 project countries
Sample: 19,071 tests from 18,220 HIV-exposed infants
Presenting data collected until December 31, 2017 from 339 sites, including 106 testing sites and 233 spoke sites.
When we started to implement the project, we planned for a pre and post intervention evaluation to compare POC with conventional EID with regards to several key outcomes. For pre-intervention, baseline data, we collected data for HIV-exposed infants who were tested with conventional EID. Data were collected from a subset of project sites, extracting data mostly from facility registers. In total, we collected data for 2875 kids from 8 project countries – one of our project countries was excluded because there were not yet implementing the project at the time.
For post intervention data collection – we are presenting data that were collected between December until December Data were collected once the platform was in place and for 7 out of the 8 countries, data were collected using a newly introduced POC EID test request form. In the country not using the form, we extracted data from existing sources. This analysis includes data for just over 18,000 kids from 339 sites.

7. Evaluation Results: Conventional vs. POC EID
Conventional EID
POC EID
p value
Median TAT from blood sample collection to result returned to caregiver (IQR)
55 days (31-77)
0 days (0-1 )
p<0.001
Results received by caregiver within 30 days
18.7 %
(542/2,899)
98.3% (18,737/19,058)
Percent of HIV-infected infants started on ART within 60 days of sample collection
43.3% (42/97) 
92.3% (639/692)
Median TAT from blood sample collection to ART initiation for HIV-infected infants (IQR)
49 days (30-68)
0 days (0-3)
Here is our first set of results – comparing conventional with POC EID across several key indicators. Our data show that POC did significantly better across all our indicators as compared to conventional EID. 5 times more results were returned to caregiver within 30days when kids were tested with POC – that TAT within 30 days is recommended by the WHO. The median TAT to results received was 0 days with POC as compared to 55 days with conventional. More than 90% of HIV-positive infants were initiated on treatment when tested with POC and if we look at the total turnaround time between sample collection and ART initiation, we note a considerable decrease from 49 days to 0 days.
The key message here is that HIV-infected infants can be started on treatment on the same day they are tested.

8. Evaluation Results: Testing vs. Spoke Sites
Testing Sites (n = 106)
Spoke Sites (n = 233 )
p value
Number of infants tested
13,070
5,155
 ---
Number of EID tests
13,673
5,398
Median TAT from blood sample collection to result returned to caregiver (IQR)
0 days (0-0)
2 days (1-7)
p<0.001
Results received by caregiver within 30 days
99.4% (13,591/13,667)
95.5% (5,146/5,391)
Percent of HIV-infected infants started on ART within 60 days
91.9% (488/532)
94.4% (151/160)
p=0.270
Median TAT from blood sample collection to antiretroviral therapy initiation for HIV-infected infants (IQR)
0 days (0-1)
3 days (1-5)
We also wanted to understand any difference for the same key indicators for kids tested at a testing site vs kids tested a spoke sites. Here are our results. Despite a statistically significant difference for most outcomes, there really is no clinical difference in key outcomes for children tested at either a spoke or testing site. And what is important is that there is no difference, statistically or clinically, for HIV positive infants who are initiated on treatment within 60 days, which is very encouraging.
Percent of HIV infected infants at testing sites: 4.10% (534/13,019) and spoke sites: (3.13% (161/5,136) – p=0.002.
Why 6- days – 20% of HIV-infected infants die if untreated by 2 months.

9. Evaluation Results: High-yield Entry Points
Infants Tested (% of all infants tested)
HIV-infected infant percent
Percent of HIV-infected infants started on antiretroviral therapy
PMCT
15,493 (85.4%)
3.2% (494/15,493)
95.1% (470/494)
Maternity
1,078 (5.9%)
1.1% (12/1,078)
66.7% (8/12)
Pediatric Inpatient
526 (2.9%)
15.2% (80/526)
86.3% (69/80)
Vaccination
412 (2.3%)
2.9%
(12/412)
83.3% (10/12)
Outpatient
265 (1.5%)
17.7% (47/265)
87.2% (41/47)
POC EID appears to be particularly beneficial for alternative entry point testing, such as pediatric inpatient and outpatient wards. Our results showed that these entry points had higher positivity rates than PMTCT services and children presenting at these entry points are often at an advanced stage of illness. Thus, time is of the essence because children do not stay in pediatric wards for a long duration of time and are likely to be discharged or die before results are returned from conventional EID and, as these children are ill, they are in urgent need of treatment.

10. Cost Per Test Result Returned
Price of diagnostic technologies is a key consideration for national programs, implementers, and funders.
Total cost of ownership includes: reagents, controls and other consumables, costs of equipment, logistics, basic training, and service and maintenance costs.
Conventional
POC (current throughput)
POC (optimal throughput)
Total cost of ownership*
$24.25
($ )
$37.20
($ )
$26.75
($ )
Cost per result returned in 30 days (range)
$ USD ($96.26-$165.76)
$37.89 USD ($32.54-$43.25)
$27.24 USD ($21.39-$33.10)
Cost per result returned in 3 months (range)
$38.89 USD ($28.57-$49.21)
$37.51 USD ($32.21-$42.81)
$26.97 USD ($21.17-$32.76)
Finally, we want to understand how much POC costs because price is important to everyone.
When doing our cost per test analysis, we considered that the cost per test result received was actually a closer measure of the true value of a diagnostic. We can’t consider the cost of the test itself because if the test result is not received, it’s wasted resources and cannot impact clinical decision-making. Cost per result received was calculated using The Global Fund to Fight AIDS, Tuberculosis and Malaria’s total cost of ownership (TCO) estimates for both POC and conventional EID.
So the first row, shows the average TCO that was calculated for the most commonly used conventional EID products (Roche and Abbott conventional analyzers) and the POC EID products used in this project. As the TCO is dependent on instrument throughput, for POC EID, we calculated the TCO based on our current throughput (up to 3 EID/day) and an optimal throughput as estimated by the Global Fund (approximately 70% of platform capacity).
Then in the second and third row, we see the average TCO for conventional and POC EID that were then divided by the average percentage of results returned to caregiver for the respective EID testing modality.
If tests are returned within 30 days, POC is considerably more cost effective than conventional. If tests are returned within 3 months however, which most are regardless of testing modality, costs start to balance out a bit more but POC remain more cost effective. *

11. Conclusions
POC resulted in significantly improved EID outcomes when compared with conventional EID:
Four times more likely for results to be returned to caregiver within 30 days with POC
Twice as likely for HIV-positive infants to be initiated on treatment within 60 days using POC
HIV-positive infants more likely to be initiated on treatment on the same day as blood sample collection
Hub-and-spoke model successfully extends access to POC EID, without compromising care.
POC EID is cost-effective: a worthy investment
Important real-world evidence that POC is a critical tool to closing the diagnostic gap in ensuring quality and effective EID.
We believe that this is the first multi-country evaluation using routine testing data collected within MOH structures. Our real world evidence showed that POC resulted in significantly improved EID outcomes when compared to conventional testing. More caregiver received their testing results, sooner. More HIV-positive infants were initiated on treatment, sooner. The hub and spoke model successfully extends access to EID without compromising quality of care. Finally, our calculations showed POC EID to be cost effective.
So if we think back to those three challenges I mentioned at the start – we se that POC, which is both a worthy investment and valuable addition to existing EID networks, can help to close the diagnostic gap by ensuring quality and effective EID.

12. For more information visit:
Acknowledgements
Authors: Flavia Bianchi, Rhoderick Machekano, Jean-Francois Lemaire, Emma Sacks, Rebecca Bailey, Valery Nzima, Patricia Fassinou, Anafi Mataka, Collins Odhiambo, Addmore Chadambuka, Gcinile Nyoni, Manuel Carlos Sabonete, Gilles Francois Ndayisaba and Jennifer Cohn. The authors have no conflicts of interest to declare
The Elizabeth Glaser Pediatric AIDS Foundation and the POC EID Project Team acknowledge and thank the mothers, fathers and caregivers who brought their infants in for testing at our intervention clinics. Without them, this work would not be possible.
Thank You
For more information visit: