Elimination of Mother to Child Transmission of HIV:
Performance of Different Models of Care when Initiating Lifelong ART for Pregnant Women in Malawi
(Option B+)
Monique van Lettow, Richard Bedell, Isabell Mayuni, Gabriel Mateyu, Megan Landes, Adrienne Chan Vanessa van Schoor, Teferi Beyene, Anthony Harries, Stephen Chu, Andrew Mganga, Joep J van
Oosterhout
Malawi
• new PMTCT strategy in 2011Option B+
• Implemented in other countries
• No formal evidence base and concerns about losses to follow up
Option B+ in Malawi - Learning by Doing
• No guidelines for integrating Option B+ into the routine service
• Different approaches had to be considered for: Location, timing ART
Adherence counseling
Follow-up after delivery or breastfeeding
• Would this affect uptake, adherence, retention?
Chibwandira et al; MMWR 2013
• 750% increase ART pregnant & BF women 4
Nation wide HF level data:
• 6-month retention 83%
• Great variation between health facilities: 100 – 42%
Limited insight into factors determining uptake and retention
AIDS 2014
Study Aims
• Describe the diversity of approaches to service organization (models of care) for the delivery of Option B+ to pregnant women in health facilities in Malawi
• Explore associations between the model of care and program indicators at health facility level:
- uptake HIV testing in ANC- uptake ART - retention on ART
South East Health Zone
• 6 out of 28 districts 3.5 million inhabitants
• Integrated HIV care services in 153 facilities, out of 588 facilities nationwide
• 154,000 pregnancies annually 14.6% HIV-infected
Setting
1. Health Facility survey staff, ART initiation location, adherence counselling, follow-up and transfer identify models of care
2. Health facility cohort reports to ascertain:
- Uptake of HIV testing and ART initiation pregnant women in ANC (July 2012-June 2013)
- Cohort survival outcome data to evaluate 6-month outcomes women registered as having started ART under Option B+ (July 2012-Dec 2012)
Methods
3. Associations between identified models of care and:
I. Uptake of HIV testing and ART among pregnant women
II. 6- month treatment outcomes
descriptive analyses and logistic regression models
High HIV testing uptake >85%
High retention on ART >92%
• 141/153 health facilities included in the studyResults
Characteristics (median) District Hospital
(n=4)
Community
Hospitals (n=8)
Health Centers (n=120)
Private Clinics (n=9)
Medical Doctors 2 0 0 0
Other Clinical staffClinical Officers/Nurses/Medical Assistants
50 24 3 4
Support staffCounselors/Expert Patients/ART clerks
25 15 8 3
Total number of staff per facility 77 39 11 7
Women newly registered for ANC/quarter
1025 377 265 116
Women newly registered start ART Option B+ 6-month cohort: July-Dec 2012
184 105 50 14
A: Facilities where women are initiated and followed on ART
at ANC clinic until giving birth (n=75)
B: Facilities where women receive only the first dose of ART at ANC clinic, then follow up at ART clinic (n=38)
C: Facilities where women are referred from ANC to the ART clinic for ART initiation and follow-up (n=18)
D: Facilities serving as ART referral sites (n=9) (not providing ANC)
4 Models of Care identified
PMTCT service delivery
Model D
No ANC, ART clinic (n=9)
Facilities providing both ANC and ART services (n=131)
Model B
1st dose ANC then ART clinic (n=38)
Model A
All ART in ANC (n=75)
Model C
All ART in ART clinic (n=18)
At day of ART initiation or
transfer in (n=9)
Same day at ANC (n=29)Next visit at ART (n=2)Both at ANC & ART(n=7)
Timing and place of Adherence Counseling for ART initiation
Same day at ANC (n=47)Next visit at ANC (n=12)Both at ANC (n=16)
At day of ART initiation at ART clinic (n=18)
Timing of transfer to ART- or MIP clinic
No transfer (n=9)No transfer
(n=16)at 6 wks postnatal visit
to MIP clinic (n=2)
After receiving 1st dose at ANC to ART clinic
(n=38)
at 6 wks postnatal visit: to ART clinic (n=24)to MIP clinic (n=51)
ANC indicators of health facilities by Model of Care
MODEL A
MODEL B
MODEL C
MODEL D
P-value
Women newly registered for ANC/quarter (median)
252 316 224 0.07
Women newly tested positive/quarter (median)
11 12 10 0.4
Known HIV-infected women already on ART
84% 81% 85% 0.7
HIV-infected women started ART during ANC
82% 81% 80% 0.9
Women not tested for HIV during ANC 18% 32% 30% 0.001
6-Month ART outcomes of health facilities by Model of Care
MODEL A
MODEL B
MODEL C
MODEL D
P-value
Women in cohort (median) 52 55 45 26 0.1
Women retained on ART 80% 78% 89% 92% 0.008
Defaulted 18% 20% 10% 7% 0.02
Stopped <1% <1% <1% <1% 0.6
Died 1% <1% 1% <1% 0.6
Health Facility variables associated with high HIV testing uptake at
ANC
Variables ASSOCIATED Variables NOT associated
Low Client : HIV testing Staff ratio District
Client : Clinical staff ratio
Health Facility type
Number women registered in ANC
Low number of out of stock observations of HIV test kits
Model of Care applied aOR*
Model A 3.4 Model B (lowest HTC uptake rate) -Model C n s
*Controlled for all variables in the model
Health Facility variables associated with high 6-month retention
Variables ASSOCIATED Variables NOT associated
District Health Facility type
Client : Clinical staff ratio
Timing of Adherence Counseling
Availability ART/PMTCT services daily vs. not daily
Availability of ART / MIP clinic for follow up
Low number of women in Option B+ cohort
Model of Care applied aOR*
Model A n s Model B (lowest retention rate) -Model C 5.4Model D n s
*Controlled for all variables in the model
Discussion
18-32% of pregnant women not tested for HIV at ANC
HIV testing uptake associated with
o Client : HIV testing staff ratio
o Test kit stock outs
o Model of Care
7-20% of women defaulted Option B+ by 6 months
Retention associated with
o District location
o Patient volume
o Model of Care
Worse program indicators in Model BFacilities where women receive only first dose of ART at ANC
Strengths and Limitations
Strengths: • Operational research – high quality routine
government data: real world findings about Option B+ implementation
• Large dataset (141 facilities)
Limitations: • Results may not be representative for whole Malawi• Health facility level data only (not patient level)• Cross sectional study design– Residual confounding
• Attrition due to linkage of care not studied – Models of Care C and D
Conclusions
• Varieties in the way health facilities have integrated Option B+ care into routine service delivery
• Model of Care is associated with uptake of HIV testing in ANC and retention in care on ART
• Further patient-level research is needed to guide policy recommendations
USAID and The International Union Against TB and LD for funding
Dignitas International medical program, M&E and Data departments
Acknowledgements