Imputing HIV treatment start dates from routine laboratory data in South Africa: a validation study

BMC Health Services Research, Jan 2017

Background Poor clinical record keeping hinders health systems monitoring and patient care in many low resource settings. We develop and validate a novel method to impute dates of antiretroviral treatment (ART) initiation from routine laboratory data in South Africa’s public sector HIV program. This method will enable monitoring of the national ART program using real-time laboratory data, avoiding the error potential of chart review. Methods We developed an algorithm to impute ART start dates based on the date of a patient’s “ART workup”, i.e. the laboratory tests used to determine treatment readiness in national guidelines, and the time from ART workup to initiation based on clinical protocols (21 days). To validate the algorithm, we analyzed data from two large clinical HIV cohorts: Hlabisa HIV Treatment and Care Programme in rural KwaZulu-Natal; and Right to Care Cohort in urban Gauteng. Both cohorts contain known ART initiation dates and laboratory results imported directly from the National Health Laboratory Service. We assessed median time from ART workup to ART initiation and calculated sensitivity (SE), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) of our imputed start date vs. the true start date within a 6 month window. Heterogeneity was assessed across individual clinics and over time. Results We analyzed data from over 80,000 HIV-positive adults. Among patients who had a workup and initiated ART, median time to initiation was 16 days (IQR 7,31) in Hlabisa and 21 (IQR 8,43) in RTC cohort. Among patients with known ART start dates, SE of the imputed start date was 83% in Hlabisa and 88% in RTC, indicating this method accurately predicts ART start dates for about 85% of all ART initiators. In Hlabisa, PPV was 95%, indicating that for patients with a lab workup, true start dates were predicted with high accuracy. SP (100%) and NPV (92%) were also very high. Conclusions Routine laboratory data can be used to infer ART initiation dates in South Africa’s public sector. Where care is provided based on protocols, laboratory data can be used to monitor health systems performance and improve accuracy and completeness of clinical records.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

http://www.biomedcentral.com/content/pdf/s12913-016-1940-2.pdf

Imputing HIV treatment start dates from routine laboratory data in South Africa: a validation study

Maskew et al. BMC Health Services Research Imputing HIV treatment start dates from routine laboratory data in South Africa: a validation study Mhairi Maskew 1 Jacob Bor 0 1 Cheryl Hendrickson 1 William MacLeod 0 1 Till Bärnighausen Deenan Pillay Ian Sanne 1 Sergio Carmona Wendy Stevens Matthew P Fox 1 0 Department of Global Health, Boston University School of Public Health , Boston, MA , USA 1 Heath Economics and Epidemiology Research Office, Department of Internal Medicine, School of Clinical Medicine, Faculty of Health Sciences, University of the Witwatersrand , Johannesburg , South Africa Background: Poor clinical record keeping hinders health systems monitoring and patient care in many low resource settings. We develop and validate a novel method to impute dates of antiretroviral treatment (ART) initiation from routine laboratory data in South Africa's public sector HIV program. This method will enable monitoring of the national ART program using real-time laboratory data, avoiding the error potential of chart review. Methods: We developed an algorithm to impute ART start dates based on the date of a patient's “ART workup”, i.e. the laboratory tests used to determine treatment readiness in national guidelines, and the time from ART workup to initiation based on clinical protocols (21 days). To validate the algorithm, we analyzed data from two large clinical HIV cohorts: Hlabisa HIV Treatment and Care Programme in rural KwaZulu-Natal; and Right to Care Cohort in urban Gauteng. Both cohorts contain known ART initiation dates and laboratory results imported directly from the National Health Laboratory Service. We assessed median time from ART workup to ART initiation and calculated sensitivity (SE), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) of our imputed start date vs. the true start date within a 6 month window. Heterogeneity was assessed across individual clinics and over time. Results: We analyzed data from over 80,000 HIV-positive adults. Among patients who had a workup and initiated ART, median time to initiation was 16 days (IQR 7,31) in Hlabisa and 21 (IQR 8,43) in RTC cohort. Among patients with known ART start dates, SE of the imputed start date was 83% in Hlabisa and 88% in RTC, indicating this method accurately predicts ART start dates for about 85% of all ART initiators. In Hlabisa, PPV was 95%, indicating that for patients with a lab workup, true start dates were predicted with high accuracy. SP (100%) and NPV (92%) were also very high. Conclusions: Routine laboratory data can be used to infer ART initiation dates in South Africa's public sector. Where care is provided based on protocols, laboratory data can be used to monitor health systems performance and improve accuracy and completeness of clinical records. Health systems; Monitoring and evaluation; Resource-limited settings; Missing data; Imputation; Validation; Laboratory; HIV/AIDS; Antiretroviral therapy; South Africa; Chronic disease management - Background In many low- and middle-resource settings, clinical records are often incomplete, missing, and/or poorly archived [1]. Where patient records are hand-written, implementation and maintenance of electronic records can be expensive and still not impervious to error [2–4]. Accurate clinical record keeping is increasingly important to clinical care and appropriate monitoring and evaluation in such settings [5]. Many developing country health systems were set up to provide maternal services, vaccinations, and acute care for injury and infections. However, the rising global burden of chronic disease requires health systems to manage patients’ health longitudinally, a challenge in the absence of accurate clinical information. Information gaps also make it difficult to monitor large health care programs and allocate resources optimally [6]. Improved record keeping from treatment sites or cohorts is particularly challenging in settings of high patient mobility across sites, providers, and sectors (public/private), and yet it is precisely in such settings that complete and accurate records are critical to the coordination of patient care [1, 3, 7]. For many diseases, laboratory testing is a meticulously specified element of clinical protocols, as it is the basis for many diagnostic and treatment decisions in lowresource settings. Laboratory tests are often conducted off-site at central laboratories and test results may be available directly from the laboratories, bypassing two steps prone to human error: manual entry of data into patient charts and manual transcription of chart information into electronic databases. Where data from routine laboratory tests are systematically available, such information may be used to impute missing data in clinical records. As with other chronic diseases, HIV/AIDS requires lifelong clinical management and decision-making guided by routine laboratory monitoring. South Africa has the largest number of peopl (...truncated)


This is a preview of a remote PDF: http://www.biomedcentral.com/content/pdf/s12913-016-1940-2.pdf

Mhairi Maskew, Jacob Bor, Cheryl Hendrickson, William MacLeod, Till Bärnighausen, Deenan Pillay, Ian Sanne, Sergio Carmona, Wendy Stevens, Matthew P Fox. Imputing HIV treatment start dates from routine laboratory data in South Africa: a validation study, BMC Health Services Research, 2017, pp. 41, 17, DOI: 10.1186/s12913-016-1940-2