Anticipated Impacts on Veteran's Healthcare. This pilot study is designed to inform a larger study, which will provide insight into how specific characteristics of the clinical setting may contribute to unintentional provider bias and result in healthcare disparities. This research is designed to guide the development of strategies within and outside the VA to reduce the documented provider contribution to inappropriate variation in care across patient subgroups, including but not limited to non-white patients. Background. Previous research indicates that VA and non-VA clinicians' diagnostic and therapeutic decision- making vary as a result of clinically irrelevant characteristics, such as patient race, and that this variation may contribute to healthcare disparities. However, there is little understanding of the particular features of the healthcare setting under which clinicians are most likely to be inappropriately influenced by these characteristics. Social cognitive research suggests that providers will be more likely to contribute to disparities under high levels of cognitive load--the amount of mental activity imposed on working memory--which may come from competing mental tasks, environmental factors, our own psychological or physiological state (e.g., fatigue), as well as from the demands inherent in the task at hand. This is particularly problematic because there is evidence that providers who work in settings where the majority of racial minorities receive care are likely to experience higher levels of cognitive load. However, the hypothesis that racial bias will be more likely under high levels of cognitive load has not yet been systematically tested. Objectives. The intent of this project is to pilot an innovative experimental study, to be administered to a national sample of VA primary care providers (PCPs) over the internet. The project will be designed to test the hypothesis that patient race will be more likely to inappropriately influence clinical decision-making under high levels of cognitive load. The objectives of this pilot study are to provide specific information about the feasibility of the larger study Including estimated response rate; degree and type of selection bias, effect sizes, and effectiveness of our experimental manipulation of cognitive load, all of which will inform the primary study (to be submitted to VA HSR&D). Methods. A national sample of VA PCPs will be identified from payroll records (the PAID database) and will be invited, by email, to participate in a study of the treatment of chronic noncancer pain (an area in which there is a great deal of evidence of racial disparities in care). We will recruit participants in phases until we achieve a sample of 160 (40 per cell). Participants will log into a secure website and will be randomly assigned to view a clinical vignette of a chronic pain patient that will differ in terms of patient race (white vs. black) under two conditions of cognitive load (low vs. high). Cognitive load will be operationalized by having participants engage in a concurrent computer task. Participants will then be asked to make a series of treatment decisions (the primary dependent measure) and will complete the Mental Effort scale, used as a manipulation check to test whether which we successfully manipulated cognitive load. Participants will complete a series of questions to assess pain-related beliefs and experience (covariates) and will also be asked a series of questions designed to capture any difficulties they may have had completing the study. Participants will receive a $10 gift certificate to the VA Canteen for their participation. The primary analyses will calculate: (1) the response rate; (2) the extent of and sources of selection bias (i.e., whether response to the survey varied by age, gender, site), and 3) the effect sizes of race, cognitive load, and the race X cognitive load interaction. We will use ANCOVA (analysis of covariance) to determine the extent to which scores on Mental Effort scale were highest in the 'high cognitive load' condition and lowest in the 'low cognitive load condition,' controlling for relevant covariates.
Racial disparities in healthcare remain a serious problem within and outside of VA. Although there is evidence that patients receive poorer care in facilities with a higher percentage of racial minority patients, there is little research on the underlying processes by which characteristics of healthcare settings may contribute to racial disparities in care. This is particularly important since the majority of African Americans in the U.S. currently receive care within a relatively small percentage of facilities. The ultimate anticipated contribution of this line of research is to obtain information about how the clinical setting may contribute to unintentional provider bias (including but not limited to decisions about treatment of chronic noncancer pain), in order to guide the development of strategies to reduce the documented provider contribution to inappropriate variation in care across patient subgroups in VA, including but not limited to non-white patients. Thus, this research will help inform interventions within the VA to eliminate system-level sources of disparities.
