2. Evaluation of SAMBA Viral Load point-of-care test operation by trained non-health workers in rural Health Centers, Chiradzulu District, Malawi Birgit Schramm 1, Johanna Wapling², Liselotte Wolters², Tamika Munyenyembe², Charlie Willie Masiku², Davie Baxter Zolowere 3, Eustice Mhango 4,James Raphael Kandulu 4, Sarala Nicholas 1, Jean-François Etard 1,5, Martine Peeters 5, Isabel Amoros², Elisabeth Szumilin 6, Monique Gueguen 6. 1 Epicenter,Paris; ²Médecins Sans Frontières, Malawi; 3 Ministry of Health, Chiradzulu District, Malawi; 4 Ministry of Health, Lilongwe, Malawi; 5 UMI 233 Institut de recherche pour le développement-Université de Montpellier; 6 Médecins Sans Frontières, France. ICASA 2015, Harare, Zimbabwe. Abstract WEAE0605

3. Context Chiradzulu District, Southern Region, Malawi Population 280,000 District hospital & 10 rural health centers Cohort of 35,000 PLHIV

4. Decentralisation of viral load testing HIV Viral Load (VL) testing using the SAMBA HIV-1 Semi-Q test implemented at four health centers from August 2013 – November 2014 Decentralized access to VL monitoring for 13,700 PLHIV on ART Availability of healthcare and laboratory staff is a major challenge

5. SAMBA HIV-1 Semi-Q Test Semi-quantitative detection 1000 copies HIV RNA /mL cut-off Two step process: Sample preparation - automated extraction process Amplification and detection - isothermal amplification in closed system with visual dipstick visual readout <1000 copies/mL ≥ 1000 copies/mL S imple AM plification B ased A ssay

6. To evaluate the performance of trained non-health community worker operation of the SAMBA HIV-1 Semi-Q test for POC viral load as compared to lab technicians. Objective

7. Methods Study sites: Namitambo Health center, Bilal Health center Operators: 1 lab technician (LT), 2 trained community workers (TCW) at each site Secondary school certificate, residing in the health center catchment area 12.5 days structured training Sample size: n = 271 Sample type:plasma samples from EDTA whole blood for routine VL test Tested by:LT on SAMBA TCW on SAMBA Quantitative viral load Study period: 2.5 months Average 2.5 samples/day (range 0 – 6)

8. Characteristics of study samples SAMBA Samples tested on SAMBA N (%) Total tests271 (100) Total invalid tests0 Tests at Namitambo health center141 (52.0) Tests at Bilal health center130 (48.0) Detectable tests Namitambo health center23 (10.3) Detectable tests Bilal health center28 (13.4)

9. Good agreement between SAMBA operators Agreement : 95.2% (CI95%: 91.9-97.4) Kappa Coefficient: 0.85 (95%CI : 0.76-0.93) Discordance: 4.8% (CI95%: 2.6-8.1) LT TCW detectableundetectableTotal detectable46854 undetectable5212217 Total51220271

10. VL (copies/mL)N (%) not detected/below threshold220 (82.2) 390 -10004 (6.3) 1,000 – 10,00019 (7.0) 10,000 – 100,00016 (5.9) > 100,00012 (4.4) <1000224 (82.7) ≥100047 (17.3) Characteristics of study samples quantitative viral load Median VL: 12, 799 copies/mL IQR: 3,656 – 97,136 copies/mL

11. Good agreement of SAMBA operators with quantitative VL SAMBA Biocentric VL (cps/ml) <500 500 - 9991000-1999 >2000 LT detectable72637 undetectable214114 TCW detectable82639 undetectable213112 95.9% agreement (92.8%-97.9%) 96.3% agreement (93.3%-98.2%)

12. TCW SAMBA operation feedback SAMBA operation rated ‘easy’ or ‘very easy’ Community worker-cadre can operate SAMBA with sufficient training LT not required Identified potential gaps for TCW operation of SAMBA Community ownership of access to viral load

13. Conclusion  Adequately trained community workers delivered SAMBA VL POC results equivalent to lab technicians  Task shifting of simplified VL POC-technologies to non-health workers can be a key strategy to overcome health worker shortage and to ensure sustainable access to routine VL testing in decentralized settings

14. Acknowledgements Study team Study Participants MSF field team Namitambo & Bilal Health Center Ministry of Health Malawi Sponsors UNITAID Médecins Sans Frontières France

16. Training Schedule DayActivity 1Welcome and Introductions Observation of the health center and mini lab Basic HIV knowledge 2HIV care and laboratory tests SAMBA-1 test – introduction & what’s in the box 3Observation SAMBA test Pipetting and setting up your lab bench 4A day in the mini lab Study protocols Development of basic study SOPs 5Orientation at health center for Study protocol Hands on familiarisation of SAMBA test 6Observation at health center & mini-lab SAMBA Demo run DayActivity 7Observation at health center & mini-lab SAMBA familiarisation run 1, TCW 2 8Observation at health center & mini-lab SAMBA familiarisation run 2, TCW1 9Refresh and revise; HIV, lab practices, study protocols Meeting at the Health Center for Study commencement 10Observation at health center & mini-lab SAMBA familiarisation run 2, TCW 1 11Observation at health center & mini-lab SAMBA Proficiency test, TCW 1 12Observation at health center & mini-lab SAMBA Proficiency test, TCW 2

17. Table of operator discordant discordant results with quantitative VL 13 LT/TCW discordant results 2 within accuracy range LT: 3 false positive, 3 false negative TCW: 4 false positive, 1 false negative VL copies/mLSAMBA–LTSAMBA -TCW NDdetectableundetectable NDdetectableundetectable NDdetectableundetectable NDundetectabledetectable NDundetectabledetectable NDundetectabledetectable <390undetectabledetectable 619undetectabledetectable 677detectableundetectable 3996undetectabledetectable 5215detectableundetectable 9526undetectabledetectable 18620undetectabledetectable

18. Table of VL test discordant results LT N=11 discordant 7 false-positive, 4 false-negative TCW N=10 discordant 8 false-positive, 2 false-negative N=5 LT and TC same discordant result: 4False positive, 1 false-negative