Logo Logo
Help
Contact
Switch language to German
HIV Viral load measurement in the Public Health Approach to HIV/AIDS. developing a clinical score for patient management to identify patients at risk of failing HIV treatment
HIV Viral load measurement in the Public Health Approach to HIV/AIDS. developing a clinical score for patient management to identify patients at risk of failing HIV treatment
Background To end AIDS by 2030, the WHO 90-90-90 targets call for 90% virologic suppression in those on ART. In this context, it is crucial to understand which factors drive virologic suppression and how available resources can be targeted most effectively. This thesis evaluates a large HIV treatment program in Tanzania and explores performance and factors associated with treatment outcome on individual and on health system level. It then develops a clinical score to predict virologic failure and optimize patient management. Design This cross-sectional facility-based study assessed 702 patients stratified by time on ART at 7 study sites selected to represent regions of the study area and health care level. Methods Facility and patient-level information were collected during a single study visit. Logistic regression analysis and Generalized Boosted Model Technique derived Propensity Score Methods were used to explore health system and individual-level factors associated with virological failure. Predictive multilevel mixed logistic regression models were developed, externally validated and simplified into a normogram for the clinical score which was then tested against WHO recommended failure criteria using Decision Curve Analysis. Results Within the population on ART, 89% was virologically suppressed below 1000 copies/ml and 86% below 400 copies/ml. Differences could be found between health care levels but not regions. The study site had a direct impact on treatment outcome on the individual and health system level. Performance of the clinical scores was high with a ROC-AUC of 0.8 in the training, and ROC-AUC between 0.7 and 0.8 in the population and the geographic validation dataset. Decision Curve Analysis showed a net benefit against the WHO routine and targeted viral load monitoring strategies. Conclusion To fully reach the “the last 90” health system-level interventions should support sites. On individual level, the clinical score developed could be used to better identify and manage individuals at risk of treatment failure.
HIV, ART, Program outcome, Virologic Suppression, Predictive Clinical Score
Lennemann, Tessa-Suntje
2021
English
Universitätsbibliothek der Ludwig-Maximilians-Universität München
Lennemann, Tessa-Suntje (2021): HIV Viral load measurement in the Public Health Approach to HIV/AIDS: developing a clinical score for patient management to identify patients at risk of failing HIV treatment. Dissertation, LMU München: Faculty of Medicine
[thumbnail of Lennemann_Tessa_Suntje.pdf]
Preview
PDF
Lennemann_Tessa_Suntje.pdf

2MB

Abstract

Background To end AIDS by 2030, the WHO 90-90-90 targets call for 90% virologic suppression in those on ART. In this context, it is crucial to understand which factors drive virologic suppression and how available resources can be targeted most effectively. This thesis evaluates a large HIV treatment program in Tanzania and explores performance and factors associated with treatment outcome on individual and on health system level. It then develops a clinical score to predict virologic failure and optimize patient management. Design This cross-sectional facility-based study assessed 702 patients stratified by time on ART at 7 study sites selected to represent regions of the study area and health care level. Methods Facility and patient-level information were collected during a single study visit. Logistic regression analysis and Generalized Boosted Model Technique derived Propensity Score Methods were used to explore health system and individual-level factors associated with virological failure. Predictive multilevel mixed logistic regression models were developed, externally validated and simplified into a normogram for the clinical score which was then tested against WHO recommended failure criteria using Decision Curve Analysis. Results Within the population on ART, 89% was virologically suppressed below 1000 copies/ml and 86% below 400 copies/ml. Differences could be found between health care levels but not regions. The study site had a direct impact on treatment outcome on the individual and health system level. Performance of the clinical scores was high with a ROC-AUC of 0.8 in the training, and ROC-AUC between 0.7 and 0.8 in the population and the geographic validation dataset. Decision Curve Analysis showed a net benefit against the WHO routine and targeted viral load monitoring strategies. Conclusion To fully reach the “the last 90” health system-level interventions should support sites. On individual level, the clinical score developed could be used to better identify and manage individuals at risk of treatment failure.