Historical visit attendance as predictor of treatment interruption in South African HIV patients: Extension of a validated machine learning model.

Retention of antiretroviral (ART) patients is a priority for achieving HIV epidemic control in South Africa. While machine-learning methods are being increasingly utilised to identify high risk populations for suboptimal HIV service utilisation, they are limited in terms of explaining relationships between predictors. To further understand these relationships, we implemented machine learning methods optimised for predictive power and traditional statistical methods. We used routinely collected electronic medical record (EMR) data to evaluate longitudinal predictors of lost-to-follow up (LTFU) and temporal interruptions in treatment (IIT) in the first two years of treatment for ART patients in the Gauteng and North West provinces of South Africa. Of the 191,162 ART patients and 1,833,248 visits analysed, 49% experienced at least one IIT and 85% of those returned for a subsequent clinical visit. Patients iteratively transition in and out of treatment indicating that ART retention in South Africa is likely underestimated. Historical visit attendance is shown to be predictive of IIT using machine learning, log binomial regression and survival analyses. Using a previously developed categorical boosting (CatBoost) algorithm, we demonstrate that historical visit attendance alone is able to predict almost half of next missed visits. With the addition of baseline demographic and clinical features, this model is able to predict up to 60% of next missed ART visits with a sensitivity of 61.9% (95% CI: 61.5-62.3%), specificity of 66.5% (95% CI: 66.4-66.7%), and positive predictive value of 19.7% (95% CI: 19.5-19.9%). While the full usage of this model is relevant for settings where infrastructure exists to extract EMR data and run computations in real-time, historical visits attendance alone can be used to identify those at risk of disengaging from HIV care in the absence of other behavioural or observable risk factors.

Validation and Improvement of a Machine Learning Model to Predict Interruptions in Antiretroviral Treatment in South Africa.

INTRODUCTION: Machine learning algorithms are increasingly being used to inform HIV prevention and detection strategies. We validated and extended a previously developed machine learning model for patient retention on antiretroviral therapy in a new geographic catchment area in South Africa. METHODS: We compared the ability of an adaptive boosting algorithm to predict interruption in treatment (IIT) in 2 South African cohorts from the Free State and Mpumalanga and Gauteng and North West (GA/NW) provinces. We developed a novel set of predictive features for the GA/NW cohort using a categorical boosting model. We evaluated the ability of the model to predict IIT over all visits and across different periods within a patient's treatment trajectory. RESULTS: When predicting IIT, the GA/NW and Free State and Mpumalanga models demonstrated a sensitivity of 60% and 61%, respectively, able to correctly predict nearly two-thirds of all missed visits with a positive predictive value of 18% and 19%. Using predictive features generated from the GA/NW cohort, the categorical boosting model correctly predicted 22,119 of a total of 35,985 missed next visits, yielding a sensitivity of 62%, specificity of 67%, and positive predictive value of 20%. Model performance was highest when tested on visits within the first 6 months. CONCLUSIONS: Machine learning algorithms may be useful in informing tools to increase antiretroviral therapy patient retention and efficiency of HIV care interventions. This is particularly relevant in developing countries where health data systems are being strengthened to collect data on a scale that is large enough to apply novel analytical methods.

Outcomes of laboratory-confirmed SARS-CoV-2 infection during resurgence driven by Omicron lineages BA.4 and BA.5 compared with previous waves in the Western Cape Province, South Africa.

Objective: We aimed to compare clinical severity of Omicron BA.4/BA.5 infection with BA.1 and earlier variant infections among laboratory-confirmed SARS-CoV-2 cases in the Western Cape, South Africa, using timing of infection to infer the lineage/variant causing infection. Methods: We included public sector patients aged ≥20 years with laboratory-confirmed COVID-19 between 1-21 May 2022 (BA.4/BA.5 wave) and equivalent prior wave periods. We compared the risk between waves of (i) death and (ii) severe hospitalization/death (all within 21 days of diagnosis) using Cox regression adjusted for demographics, comorbidities, admission pressure, vaccination and prior infection. Results: Among 3,793 patients from the BA.4/BA.5 wave and 190,836 patients from previous waves the risk of severe hospitalization/death was similar in the BA.4/BA.5 and BA.1 waves (adjusted hazard ratio [aHR] 1.12; 95% confidence interval [CI] 0.93; 1.34). Both Omicron waves had lower risk of severe outcomes than previous waves. Prior infection (aHR 0.29, 95% CI 0.24; 0.36) and vaccination (aHR 0.17; 95% CI 0.07; 0.40 for boosted vs. no vaccine) were protective. Conclusion: Disease severity was similar amongst diagnosed COVID-19 cases in the BA.4/BA.5 and BA.1 periods in the context of growing immunity against SARS-CoV-2 due to prior infection and vaccination, both of which were strongly protective.

Outcomes of laboratory-confirmed SARS-CoV-2 infection in the Omicron-driven fourth wave compared with previous waves in the Western Cape Province, South Africa.

OBJECTIVES: The objective was to compare COVID-19 outcomes in the Omicron-driven fourth wave with prior waves in the Western Cape, assess the contribution of undiagnosed prior infection to differences in outcomes in a context of high seroprevalence due to prior infection and determine whether protection against severe disease conferred by prior infection and/or vaccination was maintained. METHODS: In this cohort study, we included public sector patients aged ≥20 years with a laboratory-confirmed COVID-19 diagnosis between 14 November and 11 December 2021 (wave four) and equivalent prior wave periods. We compared the risk between waves of the following outcomes using Cox regression: death, severe hospitalisation or death and any hospitalisation or death (all ≤14 days after diagnosis) adjusted for age, sex, comorbidities, geography, vaccination and prior infection. RESULTS: We included 5144 patients from wave four and 11,609 from prior waves. The risk of all outcomes was lower in wave four compared to the Delta-driven wave three (adjusted hazard ratio (aHR) [95% confidence interval (CI)] for death 0.27 [0.19; 0.38]. Risk reduction was lower when adjusting for vaccination and prior diagnosed infection (aHR: 0.41, 95% CI: 0.29; 0.59) and reduced further when accounting for unascertained prior infections (aHR: 0.72). Vaccine protection was maintained in wave four (aHR for outcome of death: 0.24; 95% CI: 0.10; 0.58). CONCLUSIONS: In the Omicron-driven wave, severe COVID-19 outcomes were reduced mostly due to protection conferred by prior infection and/or vaccination, but intrinsically reduced virulence may account for a modest reduction in risk of severe hospitalisation or death compared to the Delta-driven wave.