Comment from Heather Duncan

As an epidemiologist, I am concerned that the proposed restrictions on federal grant funding related to “DEI or DEIA policies or practices,” “gender ideology,” and related topics described in §200.300 are overly broad, lack operational definitions, and risk introducing substantial methodological limitations into biomedical and public health research. The rule prohibits federal funds from being used to “fund, promote, encourage, subsidize, or facilitate” a wide range of activities, including research related to gender identity, sex-based differences, and gender transition-related care for individuals under 19. These prohibitions are embedded as mandatory conditions across federal agencies and programs, and may result in enforcement actions or grant termination even when the relevant activities are conducted using separate, non-federal funding sources. From an epidemiologic perspective, a central concern is the impact of these restrictions on construct validity and measurement practices. Terms such as “cisgender” and “transgender” are widely used in peer-reviewed medical and public health literature as standardized constructs for defining exposure and population subgroups. These constructs are not ideological in nature; rather, they are operational definitions developed to enable consistent classification of gender identity in epidemiologic studies. Removing or discouraging the use of such terms would compromise construct validity by forcing researchers to rely on inconsistent, non-standard, or proxy measures that do not accurately capture the underlying exposure of interest. Relatedly, these restrictions increase the risk of differential misclassification bias. If researchers are constrained in how they define or measure gender identity, there is a heightened likelihood that individuals will be misclassified into inappropriate analytic categories. Such misclassification is particularly concerning if it is systematic rather than random, as it can bias effect estimates toward or away from the null in unpredictable ways, ultimately distorting observed associations between gender identity and health outcomes. The proposed limitations may also increase measurement error across studies by preventing the use of validated survey instruments and standardized classification frameworks commonly used in population health research. Measurement error of this type can reduce statistical power, attenuate true associations, and undermine the reproducibility and comparability of findings across datasets and study populations. More broadly, restricting the use of established scientific terminology and constructs risks impairing the ability of researchers to accurately characterize health disparities. Sexual and gender minority populations experience well-documented differences in mental health outcomes, access to care, and disease burden. High-quality epidemiologic surveillance and analytic validity depend on the ability to reliably define and measure these populations using consistent constructs across studies and time. Additionally, the provision that allows enforcement actions based on research activities conducted with non-federal funds raises serious concerns regarding academic freedom and the integrity of scientific inquiry. In practice, this will create a chilling effect, leading institutions to avoid entire domains of legitimate public health research due to uncertainty about compliance risk. This type of anticipatory restriction can further exacerbate measurement limitations by discouraging the collection of sensitive but essential data. Over time, these constraints may reduce the quality and validity of epidemiologic evidence available to inform public health policy. When construct validity is compromised and misclassification bias is introduced at scale, the resulting literature base becomes less reliable for guiding interventions, allocating resources, and addressing health disparities. For these reasons, the proposed restrictions risk introducing systematic methodological bias into public health research and weakening the empirical foundation needed for evidence-based policy making. -Heather Duncan, MPH, PhD