2. www.ias2015.org A TTRITION A MONG P REGNANT AND N ON - PREGNANT P ATIENTS I NITIATING ART IN H AITI F OLLOWING A DOPTION OF O PTION B+ Jean Wysler Domercant 1, Nancy Puttkammer 2, Lydia Lu 3, Kesner Francois 4, Dana Duncan 5, Reynold Grand’Pierre 4, David Lowrance 1, Michelle Adler 5 1 Centers for Disease Control and Prevention, Port au Prince Haiti; 2 International Training and Education Center for Health, Seattle WA, USA; 3 Eagle Applied Sciences, Atlanta, GA; 4 Ministry of Health of the Government of Haiti; 5 Centers for Disease Control and Prevention, Atlanta GA, USA.

3. www.ias2015.org Background  ART services started in 2003  Rapid scale-up with service access in all 42 Districts  High national coverage of HTC in ANC (80%) and ART among HIV infected pregnant women (87%)  Option B+ policy adopted in March 2012 and implementation started in October 2012  Limited data on level of attrition among ART patients following the adoption of Option B+ policy

4. www.ias2015.org Methods  Retrospective cohort analysis  Clinical data for 17,084 adult patients at 73 health facilities initiating ART from Oct. 2012 – Aug. 2014  Kaplan-Meier method and Cox proportional hazards regression  Option B+  ART initiation during or within 12 weeks of pregnancy  No breastfeeding data available  Attrition  90 day lapse after a missed clinical or pharmacy-dispensing appointment  Officially-recorded discontinuation due to death or patient preference  Assessment of “silent transfer” phenomenon

5. www.ias2015.org Attrition by Patient Group

6. www.ias2015.org Factors Associated with Attrition a p<0.10, b p<0.05, c p<0.01, d p<0.001 After controlling for health facility, marital status, ART regimen, calendar period of ART start, household size, having treatment buddy named, facility-based ART selection committee met to discuss patient case prior to ART start. Patient CharacteristicsHazard Ratio 95% Confidence Interval Patient group (reference = Non-pregnant women, WHO stage I or II) d Non-pregnant women, WHO stage III or IV1.14(1.03, 1.26) Option B+, WHO stage I or II1.63(1.44, 1.83) Option B+, WHO stage III or IV1.32(1.14, 1.53) Men, WHO stage I or II1.11(0.96, 1.29) Men, WHO stage III or IV1.35(1.21, 1.50) Age (each 10-year increase) d 0.86(0.84, 0.88) Residence in same commune as health facility b 0.93(0.88, 1.00) Any known HIV+ household members d 0.80(0.71, 0.90) Starting ART within 7 days of enrollment in care d 1.41(1.32, 1.51) Baseline BMI (each 1 unit increase) d 0.97(0.96, 0.98) Moderate or severe anemia d 1.31(1.22, 1.42) Pre-ART counseling provided d 0.84(0.78, 0.91) TB treatment or prophylaxis at baseline a 0.94(0.89, 1.00) Cotrimoxizole prophylaxis at baseline c 0.88(0.80, 0.97)

7. www.ias2015.org Role of “Silent Transfers”  5,904 attrition cases from iSanté matched to records in the HIV/AIDS Surveillance System (HASS) database  “Silent transfer” among attrition cases:  Overall: 418 of 5,904 (7.1%)  Option B+: 143 of 1,766 (8.1%)  Non-pregnant women: 151 of 2,440 (6.2%)  Men: 124 of 2,043 (6.1%)

8. www.ias2015.org When Option B+ Cases Drop Out of Care Relative to Delivery

9. www.ias2015.org Conclusion  Risk of attrition up to 63% higher for Option B+ group compared with non-pregnant of similar characteristics  Modifiable risk factors:  Initiating ART immediately after enrollment in HIV care  Initiating ART late in pregnancy  Key protective factors:  Pre-ART counseling  Living in the catchment area of the health facility  OI prophylaxis

10. www.ias2015.org Limitations  Routine data source, varying quality across sites  Potential misclassification of Option B+  Difficulty of assigning dates of pregnancy, lack of breastfeeding data  Modest overestimate of attrition due to “silent transfers” and individual-level data on transfers not available for this analysis  Observational data, so findings may reflect selection bias and confounding, or other types of bias

11. www.ias2015.org Recommendations  Effort to attract pregnant women into care early in pregnancy  Ensure access to appropriate social support  Intense adherence counseling  Importance of staying on medication  Benefit for patient in addition to infant and partners  Assessing risk of attrition and barriers to adherence  Proactivity in the retention strategy

12. www.ias2015.org Acknowledgements Ministère de la Santé Publique et de la Population US Centers for Disease Control and Prevention Paul Rashad Young, Barbara Marston, Patrice Joseph, Valerie Pelletier, Reginald Jean Louis University of Washington Department of Global Health Martine Myrtil University of Washington Center for AIDS Research Biometrics Core Krista Yuhas Implementing Partners MSPP-UGP, GHESKIO, PIH, CMMB, IHV, UM, I-TECH, CDS