IAS#: A-854-0221-03721 Evaluation of the Impact of the Accelerating Children’s HIV/AIDS Treatment (ACT) Initiative on Pediatric and Adolescent HIV Testing and Yield in Western Kenya 26 July 2017 N. Okoko1, A.R. Mocello2, J. Kadima1, J. Kulzer2, G. Nyanaro1, C. Blat2, M. Guzé2, E. Bukusi1, C.R. Cohen2, L. Abuogi3, S.B. Shade4 Kenya Medical Research Institute (KEMRI), Nairobi, Kenya Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco (UCSF), CA, USA Department of Pediatrics, University of Colorado, Aurora, CO, USA Department of Epidemiology and Biostatistics, UCSF, CA, USA Good morning ladies and gentlemen, I am Nicollate Okoko, working for Family AIDS Care & Education Services (FACES) program as a Pediatric Adolescent technical Lead On behalf of the co authors, I want to briefly share with you about Evaluation of the Impact of the (ACT) Initiative on Pediatric and Adolescent HIV Testing and Yield in Western Kenya

No conflicts of interest to declare Conflict of Interest No conflicts of interest to declare

Background Despite decreasing new HIV infections, pediatric HIV remains substantial 150,000 annual new HIV infections globally (<15 years) 1.8 million children living with HIV (<15 years) < 30% of children tested in Nyanza region of Kenya HIV testing - gateway to achieving 90-90-90 Despite declining new infections, pediatric HIV remains significant, with 150,000 new infections annually and 1.8 million children (<15 years old) living with HIV globally Less than one third of children have ever been tested for HIV in the Nyanza region of western Kenya Therefore, many children that are eligible for testing in this high HIV-prevalence region have not been reached

What was ACT? Accelerating Children’s HIV/AIDS Treatment (ACT) ACT is a public-private partnership between PEPFAR and CIFF Strategic response to treatment gap ACT is a public-private partnership between PEPFAR (the United States President’s Emergency Plan for AIDS Relief) and CIFF (the Children’s Investment Fund Foundation) It is a concerted, strategic and ambitious response to the ‘treatment gap’ for children 2 year project, launched at Africa Leaders Summit in August 2014 Aimed to initiate an additional 300,000 children living with HIV on treatment in 9 priority countries in Africa Initiate 300,000 with HIV on treatment in 9 priority countries in 2 years

Examine whether activities under the Accelerating Children’s HIV/AIDS Treatment (ACT) initiative increased testing and identification of children with HIV We examined whether activities under the Accelerating Children’s HIV/AIDS Treatment (ACT) initiative increased testing and identification of children with HIV  

Methods Family AIDS Care & Education Services (FACES) KEMRI & UCSF collaboration Comprehensive HIV prevention, care, and treatment program 144 health facilities supported Migori, Homa Bay, and Kisumu counties Nyanza region of Kenya Evaluation timeframe October 2015 – September 2016 Methods: (FACES) a family focused, comprehensive HIV prevention, care, and treatment program in Nyanza region of Kenya Launched services and support in 2004 144 health facilities supported by FACES and participated in the ACT initiative We implemented activities under the ACT initiative in 144 health facilities in 3 counties (Homabay, Kisumu and Migori) in western Kenya between October 2015 and September 2016

Health Facilities 85% rural 6% urban 26% comprehensive outpatient Characteristics 85% rural Peri-urban 8% 6% urban Health dispensaries 66% 26% comprehensive outpatient Sub county hospitals and county referral hospitals 8% Homabay, Kisumu and Migori counties are among the highest ranked HIV incidence Counties from where 65% of new HIV infections in Kenya occur The vast majority of sites were rural 2/3 (two thirds) were smaller dispensaries, only 8% of sites were larger county/sub-county hospitals

Intervention Steps for Pediatric/Adolescent Testing Family testing focus: Family Information Table (FIT) utilization FIT chart audits Integrated intervention steps Additional HIV counselors Create HTC space ACT consisted of a suite of strategies targeting each of the three elements of the HIV cascade of care: Testing and linkage of positives to care, ART, retention in care and viral suppression   Step 1 Family testing was a large initiative, we prioritized index patients to reach partners and children for testing We conducted retrospective FIT chart audits which involved retrieval and review of records of adult index patients who were enrolled in family-centered HIV care between May–July 2015 and followed family outcomes through October 2015 We improved and prospectively utilized the Family Testing Table and developed a register (FTCR) to track and follow up on family members in need of testing and care initiation Family member testing status, results, enrollment in care and ART initiation for those positive were abstracted; analysis done and intervention compared to other Step 2 We then identified service gaps and hired and trained 69 HTS counselors We renovated and allocated space for HTC to create privacy and richer counseling interactions Step 3 We Conducted community outreach testing targeting areas frequented by adolescents like video dens, discos, churches and major calendar events Caregivers also received text reminders to facilitate timely Early Infant Diagnosis (EID) Community outreach testing HIV-exposed infants’ caregiver text messages

Evaluation Methods Design Data Collection Analysis Convenience sample of clinics Sites assigned to intervention vs. control dependent on whether the intervention was actively being implemented in a given month This allowed determination of impact of individual intervention Data Collection Facility level Tracking logs Number tested Number HIV positive Infants <18 months Children 18 months – 9 years Adolescents 10 years – 14 years Analysis Intervention and control sites compared Negative binomial generalized estimating equations Adjusted for repeated measures, geographic location, health facility tier, and test kit stock-outs   We collected data on the implementation of each of the ACT interventions at all 144 FACES facilities throughout the initiative period In addition, we collected monthly aggregate data on the number of infants younger than 18 months, children aged 18 months to 9 years, and adolescents 10 to 14 years of age who were tested at each facility and on the number identified as HIV-positive. We then used negative binomial generalized estimating equation models to produce an incidence rate ratio for each intervention, comparing sites that were conducting a strategy in a given month to control sites that were not conducting that strategy Analyses adjusted for repeated measures, geographic location, level of services provided (tier), and whether rapid HIV testing kits were stocked out during each month.

Results: HIV Testing During the ACT initiative period, the mean number of children tested monthly increased significantly across all age groups   Among infants younger than 18 months, the mean number tested per facility increased from 2.8 in October 2015 to 7.2 in September 2016. (p < .0001) In children 18 months to 9 years of age, the mean number tested increased from 44.8 to 142.0 (p<.0001). And in adolescents aged 10 to 14 years, the number tested increased from 30.1 to 123.3 (p<.0001).

Results: Identification of HIV Positives Identification of HIV-positive children also increased significantly among infants, from a mean yield of 0.06 positive infants per month per facility at the beginning of ACT to 0.37 at the end of the initiative period (p<.0001) in infants   We also saw significant increases in children (from 0.34 to 0.62; p=.002) and more modest increases among adolescents (from 0.17 to 0.26; p=.03).

Results Age Group October 2015 September 2016 p-value Mean number tested per facility per month < 18 months 2.8 7.2 <.0001 18 months to 9 years 44.8 142.0 10-14 years 30.1 123.3 Mean number identified HIV positive per facility per month 0.06 0.37 0.34 0.62 0.002 0.17 0.26 0.03 Mean number of children tested monthly increased across all age groups: from 2.8 to 7.2 (p<.0001) in infants <18 months; from 44.8 to 142.0 (p<.0001) in children 18 months to 9 years; and from 30.1 to 123.3 (p<.0001) in adolescents 10-14 years   Identification of HIV-positive children increased: 0.06 to 0.37 (per month per facility; p<.0001) in infants; 0.34 to 0.62 (p=.002) in children; and 0.17 to 0.26 (p=.03) in adolescents.

Successful Interventions on HIV Testing* Age Group Intervention IRR, 95%CI p-value Infants <18 months Family Information Table 2.89 (1.53, 5.49) <0.001 Children 18 months to 9 years FIT chart audits 2.15 (1.36, 3.40) Adolescents 10 to 14 years HTC space improvements 1.45 (1.09, 1.93) <0.01 Use of the Family Information Table was significantly associated with increased HIV testing in infants, with an incidence rate ratio of 2.89 (95% confidence interval [CI]=1.53,5.49; p<0.001) as compared to pre-intervention   Use of the Family Information Table was also associated with identification of HIV-positive infants, with an incidence rate ratio of 8.71 (95% CI=1.45,52.4; p<0.02) Among children, Family Information Table chart audits were significantly associated with increased testing, resulting in an incidence rate ratio of 2.15 (95% CI=1.36,3.40; p<0.001). And among adolescents, providing or renovating space for HIV testing and counseling was significantly associated with increased HIV testing, with an incidence rate ration of 1.45 (95% CI=1.09,1.93; p<0.01) *Adjusted for repeated measures, geographic location, health facility tier, and test kit stock-outs

Successful Intervention to Increase Identification of HIV Positives Age Group Intervention IRR, 95%CI p-value Infants <18 months Family Information Table 8.71 (1.45, 52.4) 0.02 *Adjusted for repeated measures, geographic location, health facility tier, and test kit stock-outs  Use of the Family Information Table was also associated with identification of HIV-positive infants, with an incidence rate ratio of 8.71 (95% CI=1.45,52.4; p<0.02)  

Creating HTC space boosts adolescent testing Family testing works Creating HTC space boosts adolescent testing ACT interventions -> Large testing gains & HIV+ yield Targeting family members of HIV positive adults increases both identification and HIV-positive yield One-time investment in improving HTC space may be an effective approach for increasing HIV-testing among adolescents in this context Significant increases in number of children tested resulted in only a modest number of new children identified with HIV, highlighting the need for multiple testing approaches

Recommendations Optimize the family unit to increase testing reach and care cascade entry Don’t let the untested slip away, track closely and conduct chart audits for follow up Consider structural improvements to facilitate testing, especially among adolescents Try multi-faceted approaches to test children and adolescents Three prior evaluations conducted at FACES that focused on family testing ( 2009, 2012, and 2015) corroborate these finding This initiative consistently leads to high identification, testing uptake among children, as well as good linkage to the care and ART initiation (2015 findings: 62% tested, 5.4% HIV positive, 87% linked, and 84% initiated on treatment) Therefore, it’s a promising strategy for 90.90.90 attainment for children and adolescents   There is a declining trend in yield through the years in 2009 the yield was 18% for children, 7.4% in 2012, and 5.4% in 2015 (significant drop). However, yield is was found to be higher with this approach (5.4%) than outpatient (<1%) and inpatient (1.5%) testing rates. (p<0.0001) Consistent updating of the family testing table and routine chart audits of children files kept us on out toes in terms of children follow for testing Consider structural interventions to facilitate testing, especially among adolescents This approach works in the facility but what about reaching children and adolescents of families outside of the health care setting? We should do all we can to find those not yet tested – both facility and community approaches Interestingly targeted community outreaches were also conducted by FACES as part of a different evaluation (and testing was reported at the facility), but the outreaches did not lead to increased testing, however a population-based community approach called the hybrid with the SEARCH study tested 86% of 10-14 years of age in a geographic region Combining strategies and continuously working to refine them will get us to 90 90 90

Acknowledgments Ministry of Health (MOH) Family AIDS Care and Education Services (FACES) Kenya Medical Research Institute (KEMRI) University of California, San Francisco (UCSF) Children’s Investment Fund Foundation (CIFF) FACES staff, clients and families We thank the MOH, FACES, KEMRI, UCSF, FACES staff, clients, and families and We greatly appreciate CIFF's support through the ACT Initiative for this evaluation and their support to reach and serve children and adolescents in need of HIV services. Learn more at: www.faces-kenya.org