Are medical records a more reliable and valid source of CD4 count, viral load, and outpatient visit data than self-reports? A comparison with electronic medical records as the gold standard

CD4 cell count and viral load are the most widely used markers of Human Immunodeficiency Virus (HIV). Obtaining accurate measures of these laboratory markers can be challenging. The purpose of this study was to assess the reliability and validity of self-reported and abstracted paper medical record CD4 counts, viral load, and outpatient visits using electronic data as the gold standard. We sampled a subgroup of 114 HIV-infected individuals in Los Angeles from a larger multi-site study for whom we could confirm outpatient visits and CD4 count and viral load data through electronic data. Agreement between electronic and paper record abstractions was approximately 90% for CD4 count (kappa = 0.85; P < 0.001), 78% for viral load (kappa = 0.56; P < 0.001), and 62% for visits (kappa = 0.02; P = 0.36). Agreement between electronic and self-reported measures was 80% for CD4 count (kappa = 0.71; P < 0.001), 77% for viral load (kappa = 0.54; P < 0.001), and 62% for visits (kappa = 0.13; P < 0.05). Using electronic data as the gold standard, we found that the accuracy of self-report was as good as paper record abstraction regarding viral load levels and visits, and only slightly lower for CD4 count. Despite limited generalizability, these findings are important because they suggest that the accuracy of clinical information collected by self-report and paper medical records are similar and raise questions about whether conducting abstractions of paper medical records is worthwhile, especially if it increases the costs of data collection.