CROI2025

Transcript

1. Clinic Switching and HIV Care Engagement in Africa: A Multi-Country Population-Based Study Joseph G. Rosen1,2,3,4, M. Kate Grabowski5, Robert Ssekubugu6, Caitlin E. Kennedy4, Larry W. Chang4,5, Gertrude Nakigozi6, Fred Nalugoda6, Ronald M. Galiwango6, Godfrey Kigozi6, Steven J. Reynolds4,5,6,7, Joseph Kagaayi6 1 Rhode Island Hospital, Providence, Rhode Island, USA 2 The Warren Alpert Medical School, Providence, RI, USA 3 School of Public Health, Brown University, Providence, RI, USA 4 Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD, USA 5 School of Medicine, Johns Hopkins University, Baltimore, MD, USA 6 Rakai Health Sciences Program, Kalisizo, Uganda 7 National Institutes of Health, Bethesda, MD, USA 32nd Conference on Retroviruses and Opportunistic Infections (CROI) San Francisco, CA, USA Wednesday, March 12th, 2025 Gaps and Solutions in the HIV Response
2. Clinic switching can destabilize HIV care continuity Limited population-level evidence of the contexts under which clinic switching drives HIV care disengagement
3. 3 Quantify population-level correlates and consequences of HIV clinic switching 2020 – 2023 Multistage cluster sampling
4. 4 Quantify population-level correlates and consequences of HIV clinic switching Interviewer-administered structured survey 2020 – 2023
5. 5 Quantify population-level correlates and consequences of HIV clinic switching Household-based HIV testing and counseling 2020 – 2023
6. 6 Quantify population-level correlates and consequences of HIV clinic switching Laboratory-based HIV viral load quantification 2020 – 2023
7. 7 Biological and self-reported measures of HIV treatment discontinuity Detectable HIV viremia Unsuppressed HIV viral load High-level HIV viremia ≥1,000 RNA copies/ml ≥200 RNA copies/ml ≥50 RNA copies/ml ART non-adherence Missing ≥2 ART doses (>5%) in the past month
8. Past-year HIV clinic switching “In the past year, did you change the clinic where you receive HIV care?”
9. 9 Statistical analyses Pooling country-level data files into aggregate multi-country dataset • Restricted sample to PLWH self-reporting HIV care enrollment with available viral load results • Implemented multi-country selection/non-response weights with jackknife standard errors Survey-weighted descriptive sample statistics • Identified univariate correlates of clinic switching using design-based F tests • Estimated prevalence of clinic switching by age, sex, residence, mobility, and country Multivariable logistic regression (adjOR with 95% CI) • Adjusted models first identified socio-behavioral and structural correlates of clinic switching • Models for HIV clinical outcomes were stratified by age, sex, residence, mobility, and country
10. 10 Statistical analyses Pooling country-level datasets into single multi-country data file • Recalculated replicate weights to correct for data aggregation across independent surveys • Implemented corrected selection/non-response weights with jackknife standard errors Survey-weighted descriptive sample statistics • Identified univariate correlates of clinic switching using design-based F tests • Estimated prevalence of clinic switching by age, sex, residence, mobility, and country Multivariable logistic regression (adjOR with 95% CI) • Adjusted models first identified socio-behavioral and structural correlates of clinic switching • Models for HIV clinical outcomes were stratified by age, sex, residence, mobility, and country
11. 11 Statistical analyses Pooling country-level datasets into single multi-country data file • Recalculated replicate weights to correct for data aggregation across independent surveys • Implemented corrected selection/non-response weights with jackknife standard errors Survey-weighted descriptive sample statistics • Identified univariate correlates of clinic switching using design-based F tests • Estimated prevalence of clinic switching by age, sex, residence, mobility, and country Multivariable logistic regression (adjusted odds ratios with 95% CI) • Adjusted models first identified socio-behavioral and structural correlates of clinic switching • Models for HIV clinical outcomes were stratified by age, sex, residence, mobility, and country
12. 12 Descriptive sample characteristics 14,017 11,642 54.5% 66.6% 61.6% 12.4% 88.4% 77.1%
13. Prevalence of past-year HIV clinic switching
14. 14 Socio-behavioral and structural correlates of clinic switching <40 years Age group (ref. 40+ years) 30+ consecutive days away from home Mobility, past year (ref. <30 days) 12+ months Duration on ART (ref. <12 months) Any travel-related barriers to care Travel-related barriers to HIV care (ref. None) 3-5 months ART prescription amount (ref. <3 months) 6+ months ART prescription amount (ref. <3 months)
15. 15 Clinic switching, HIV viremia, and ART non-adherence Outcomes Detectable HIV viremia (Ref. <50 RNA copies/ml) Unsuppressed HIV viral load (Ref. <200 RNA copies/ml) High-level HIV viremia (Ref. <1,000 RNA copies/ml) ART non-adherence (Ref. Missed <2 ART doses) Clinic switching adjOR (95% CI) p-value adjOR (95% CI) p-value adjOR (95% CI) p-value adjOR (95% CI) p-value Overall 1.05 (0.87-1.27) 0.563 0.90 (0.71-1.12) 0.325 0.83 (0.59-1.17) 0.282 1.29 (1.03-1.61) 0.028 Age group <40 years 1.12 (0.85-1.46) 0.411 0.80 (0.57-1.14) 0.208 0.73 (0.46-1.14) 0.157 1.37 (1.02-1.86) 0.040 40+ years 0.98 (0.77-1.26) 0.893 1.05 (0.75-1.49) 0.757 1.09 (0.67-1.78) 0.710 1.18 (0.83-1.67) 0.342 Sex Male 0.79 (0.58-1.08) 0.135 0.84 (0.57-1.25) 0.375 0.85 (0.48-1.51) 0.561 1.24 (0.81-1.91) 0.307 Female 1.23 (1.01-1.52) 0.047 0.93 (0.70-1.23) 0.579 0.82 (0.56-1.19) 0.278 1.32 (1.02-1.72) 0.038 Residence Urban 1.04 (0.77-1.40) 0.806 0.87 (0.60-1.27) 0.457 0.63 (0.36-1.10) 0.101 1.21 (0.85-1.72) 0.279 Rural 1.07 (0.85-1.34) 0.558 0.91 (0.68-1.22) 0.527 0.97 (0.63-1.49) 0.895 1.35 (1.02-1.79) 0.037 Multivariable models adjusted for age, sex, country, duration on ART, and ART prescription amount
16. 16 Clinic switching, HIV viremia, and ART non-adherence Outcomes Detectable HIV viremia (Ref. <50 RNA copies/ml) Unsuppressed HIV viral load (Ref. <200 RNA copies/ml) High-level HIV viremia (Ref. <1,000 RNA copies/ml) ART non-adherence (Ref. Missed <2 ART doses) Clinic switching adjOR (95% CI) p-value adjOR (95% CI) p-value adjOR (95% CI) p-value adjOR (95% CI) p-value Overall 1.05 (0.87-1.27) 0.563 0.90 (0.71-1.12) 0.325 0.83 (0.59-1.17) 0.282 1.29 (1.03-1.61) 0.028 Age group <40 years 1.12 (0.85-1.46) 0.411 0.80 (0.57-1.14) 0.208 0.73 (0.46-1.14) 0.157 1.37 (1.02-1.86) 0.040 40+ years 0.98 (0.77-1.26) 0.893 1.05 (0.75-1.49) 0.757 1.09 (0.67-1.78) 0.710 1.18 (0.83-1.67) 0.342 Sex Male 0.79 (0.58-1.08) 0.135 0.84 (0.57-1.25) 0.375 0.85 (0.48-1.51) 0.561 1.24 (0.81-1.91) 0.307 Female 1.23 (1.01-1.52) 0.047 0.93 (0.70-1.23) 0.579 0.82 (0.56-1.19) 0.278 1.32 (1.02-1.72) 0.038 Residence Urban 1.04 (0.77-1.40) 0.806 0.87 (0.60-1.27) 0.457 0.63 (0.36-1.10) 0.101 1.21 (0.85-1.72) 0.279 Rural 1.07 (0.85-1.34) 0.558 0.91 (0.68-1.22) 0.527 0.97 (0.63-1.49) 0.895 1.35 (1.02-1.79) 0.037 Multivariable models adjusted for age, sex, country, duration on ART, and ART prescription amount
17. 17 Clinic switching, HIV viremia, and ART non-adherence Outcomes Detectable HIV viremia (Ref. <50 RNA copies/ml) Unsuppressed HIV viral load (Ref. <200 RNA copies/ml) High-level HIV viremia (Ref. <1,000 RNA copies/ml) ART non-adherence (Ref. Missed <2 ART doses) Clinic switching adjOR (95% CI) p-value adjOR (95% CI) p-value adjOR (95% CI) p-value adjOR (95% CI) p-value Overall 1.05 (0.87-1.27) 0.563 0.90 (0.71-1.12) 0.325 0.83 (0.59-1.17) 0.282 1.29 (1.03-1.61) 0.028 Age group <40 years 1.12 (0.85-1.46) 0.411 0.80 (0.57-1.14) 0.208 0.73 (0.46-1.14) 0.157 1.37 (1.02-1.86) 0.040 40+ years 0.98 (0.77-1.26) 0.893 1.05 (0.75-1.49) 0.757 1.09 (0.67-1.78) 0.710 1.18 (0.83-1.67) 0.342 Sex Male 0.79 (0.58-1.08) 0.135 0.84 (0.57-1.25) 0.375 0.85 (0.48-1.51) 0.561 1.24 (0.81-1.91) 0.307 Female 1.23 (1.01-1.52) 0.047 0.93 (0.70-1.23) 0.579 0.82 (0.56-1.19) 0.278 1.32 (1.02-1.72) 0.038 Residence Urban 1.04 (0.77-1.40) 0.806 0.87 (0.60-1.27) 0.457 0.63 (0.36-1.10) 0.101 1.21 (0.85-1.72) 0.279 Rural 1.07 (0.85-1.34) 0.558 0.91 (0.68-1.22) 0.527 0.97 (0.63-1.49) 0.895 1.35 (1.02-1.79) 0.037 Multivariable models adjusted for age, sex, country, duration on ART, and ART prescription amount
18. 18 Clinic switching, HIV viremia, and ART non-adherence Outcomes Detectable HIV viremia (Ref. <50 RNA copies/ml) Unsuppressed HIV viral load (Ref. <200 RNA copies/ml) High-level HIV viremia (Ref. <1,000 RNA copies/ml) ART non-adherence (Ref. Missed <2 ART doses) Clinic switching adjOR (95% CI) p-value adjOR (95% CI) p-value adjOR (95% CI) p-value adjOR (95% CI) p-value Overall 1.05 (0.87-1.27) 0.563 0.90 (0.71-1.12) 0.325 0.83 (0.59-1.17) 0.282 1.29 (1.03-1.61) 0.028 Age group <40 years 1.12 (0.85-1.46) 0.411 0.80 (0.57-1.14) 0.208 0.73 (0.46-1.14) 0.157 1.37 (1.02-1.86) 0.040 40+ years 0.98 (0.77-1.26) 0.893 1.05 (0.75-1.49) 0.757 1.09 (0.67-1.78) 0.710 1.18 (0.83-1.67) 0.342 Sex Male 0.79 (0.58-1.08) 0.135 0.84 (0.57-1.25) 0.375 0.85 (0.48-1.51) 0.561 1.24 (0.81-1.91) 0.307 Female 1.23 (1.01-1.52) 0.047 0.93 (0.70-1.23) 0.579 0.82 (0.56-1.19) 0.278 1.32 (1.02-1.72) 0.038 Residence Urban 1.04 (0.77-1.40) 0.806 0.87 (0.60-1.27) 0.457 0.63 (0.36-1.10) 0.101 1.21 (0.85-1.72) 0.279 Rural 1.07 (0.85-1.34) 0.558 0.91 (0.68-1.22) 0.527 0.97 (0.63-1.49) 0.895 1.35 (1.02-1.79) 0.037 Multivariable models adjusted for age, sex, country, duration on ART, and ART prescription amount
19. 19 Clinic switching, HIV viremia, and ART non-adherence Outcomes Detectable HIV viremia (Ref. <50 RNA copies/ml) Unsuppressed HIV viral load (Ref. <200 RNA copies/ml) High-level HIV viremia (Ref. <1,000 RNA copies/ml) ART non-adherence (Ref. Missed <2 ART doses) Clinic switching adjOR (95% CI) p-value adjOR (95% CI) p-value adjOR (95% CI) p-value adjOR (95% CI) p-value Overall 1.05 (0.87-1.27) 0.563 0.90 (0.71-1.12) 0.325 0.83 (0.59-1.17) 0.282 1.29 (1.03-1.61) 0.028 Age group <40 years 1.12 (0.85-1.46) 0.411 0.80 (0.57-1.14) 0.208 0.73 (0.46-1.14) 0.157 1.37 (1.02-1.86) 0.040 40+ years 0.98 (0.77-1.26) 0.893 1.05 (0.75-1.49) 0.757 1.09 (0.67-1.78) 0.710 1.18 (0.83-1.67) 0.342 Sex Male 0.79 (0.58-1.08) 0.135 0.84 (0.57-1.25) 0.375 0.85 (0.48-1.51) 0.561 1.24 (0.81-1.91) 0.307 Female 1.23 (1.01-1.52) 0.047 0.93 (0.70-1.23) 0.579 0.82 (0.56-1.19) 0.278 1.32 (1.02-1.72) 0.038 Residence Urban 1.04 (0.77-1.40) 0.806 0.87 (0.60-1.27) 0.457 0.63 (0.36-1.10) 0.101 1.21 (0.85-1.72) 0.279 Rural 1.07 (0.85-1.34) 0.558 0.91 (0.68-1.22) 0.527 0.97 (0.63-1.49) 0.895 1.35 (1.02-1.79) 0.037 Multivariable models adjusted for age, sex, country, duration on ART, and ART prescription amount
20. 20 Conclusions and implications Clinic switching highest in younger, mobile groups with structural barriers to care
21. 21 Conclusions and implications Multi-month ART dispensing may be protective against care discontinuity during transfers
22. 22 Conclusions and implications Unobserved temporality: clinic switching 🡪🡪 HIV care discontinuity?
23. 23 Conclusions and implications Strategies enhancing the safety and efficiency of clinic transfers are needed
24. 24 Acknowledgments Funding • U.S. Centers for Disease Control and Prevention (U2GGH001226) • National Institute of Allergy and Infectious Diseases (R01AI155080) • National Institute of Mental Health (R25MH083620) • Johns Hopkins University Center for AIDS Research (P30AI094189) Assistant Professor (Research) Department of Medicine The Warren Alpert Medical School Brown University greg_rosen@brown.edu Greg Rosen, PhD, MSPH • PHIA participants • Survey enumerators • Study managers and personnel • ICAP at Columbia University • Co-authors / thought partners Thank you!