Dr. Williams of crf calls for Re-Evaluating Studies

Dr. Christopher Williams, principal investigator of the Critical Race Framework, calls for studies to be re-analyzed in light of a recent CRF critical appraisal of a major study on HIV/AIDS racial disparities study. 

A major 2014 study on HIV/AIDS among 803 black and white men who have sex with men (MSM) in the Atlanta metropolitan area should be re-analyzed, according to Dr. Christopher Williams. Dr. Williams relied on results from the CRF Quantitative Critical Appraisal Aid (QCAA) - a generalized approach for quantifying error due to potential areas in the CRF theoretical framework. The 2014 study found major disparities between black and white MSM in HIV prevalence (43% vs 13%). In each Census tract-level analysis (e.g., poverty rate, annual income, unemployment rate), there were statistically significant differences. Cohort differences in individual poverty status, health insurance, arrest (prior 12 months), and homelessness (prior 12 months) were also significant. However, Dr. Williams explains that the authors of this study did not appropriately analyze the data, stopping short of fully analyzing the structural impact on HIV/AIDS prevalence.

The study relied on bivariate analyses which does not account for how variables are influenced by each other. It also overemphasized racial analysis, as opposed to explaining the variance can be attributed to structural inequity. "While the study did include structural factors, the race essentialist approach limited the statistical approaches that could have provided more meaning away from racialization and toward more socially conscious and practical use for policy reform. Even if they had conducted regression, a structural index would have been most appropriate given the high disparities." The study had considerable areas of uncertainty. The unmitigated error from the lack of reliability and validity evidence associated with racial variables, among other sources of potential error, shift the 95% confidence intervals. Table 1 shows the effect of varying validity error on adjusted CIs from a generalized approach.

Grok concluded, "The QCAA quantifies 40–45% internal error in the study’s outcomes, driven by validity (20%) and internal validity (15–20%) issues. Revised Table 3 shows wider CIs, with age-specific disparities less certain in smaller groups (18–19, 40+). The CRF critique highlights the need for structural analyses to address race-related disparities fully."