Related Experiment Video
Updated: Mar 8, 2026
23:56
Comprehensive & Cost Effective Laboratory Monitoring of HIV/AIDS: an African Role Model
Published on: October 31, 2010
17.4K
Imputing HIV treatment start dates from routine laboratory data in South Africa: a validation study
Mhairi Maskew1, Jacob Bor2,3,4,5, Cheryl Hendrickson1
1Heath Economics and Epidemiology Research Office, Department of Internal Medicine, School of Clinical Medicine, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa.
BMC Health Services Research
|January 19, 2017
Summary
Routine laboratory data can accurately infer antiretroviral treatment (ART) initiation dates in South Africa, improving HIV program monitoring. This method enhances health system performance and clinical record accuracy.
Area of Science:
- Public Health
- Health Informatics
- HIV/AIDS Research
Background:
- Poor clinical record keeping is a significant challenge for health systems monitoring and patient care in resource-limited settings.
- Accurate tracking of antiretroviral treatment (ART) initiation is crucial for effective HIV program management.
- Existing methods for monitoring ART initiation are prone to errors, particularly in large-scale public health programs.
Purpose of the Study:
- To develop and validate a novel method for imputing ART initiation dates using routine laboratory data.
- To enable real-time monitoring of South Africa's national ART program.
- To overcome the limitations of traditional chart review methods for data collection.
Main Methods:
- An algorithm was developed to impute ART start dates based on 'ART workup' laboratory test dates and established clinical protocols for time to initiation (21 days).
- Validation was performed using data from two large HIV cohorts in South Africa (Hlabisa and Right to Care), including over 80,000 HIV-positive adults.
- Key performance metrics including sensitivity, specificity, positive predictive value, and negative predictive value were calculated to assess the accuracy of the imputed dates against known ART initiation dates.
Main Results:
- The algorithm demonstrated high accuracy in imputing ART initiation dates, with a median sensitivity of 83-88% across cohorts.
- Positive predictive value was high (95% in Hlabisa), indicating reliable prediction of true start dates for patients with a documented lab workup.
- Specificity (100%) and negative predictive value (92%) were also strong, confirming the method's robustness.
Conclusions:
- Routine laboratory data provides a viable and accurate source for inferring ART initiation dates within South Africa's public sector HIV program.
- This imputation method can significantly enhance the monitoring of national ART programs by leveraging readily available real-time laboratory data.
- The findings support the use of laboratory data to improve the accuracy and completeness of clinical records and bolster health systems performance.
