Travis Whitacre

@traviswhitacre.bsky.social

1147 Followers 1336 Following

Yale Postdoc | SMU PhD | Fields: Health Economics, Labor Economics, Applied Microeconometrics | Studying child development, mortality, and health inequities | he/him | WT alum Website: https://traviswhitacre.github.io/

Cherry blossom
Panda walking toward some trees
A table showing the effect on future employment, income, educational attainment, and educational achievement of likely receiving an ADHD diagnosis for men and women.
Table of ADHD, Undiagnosed ADHD, and Conditional ADHD Underdiagnosis risk for the full sample, male vs. female sample, Black vs. Non-Black sample, Hispanic vs. Non-Hispanic sample, Parental Income Quartile, Parental Health Insurance Status, and Observed Diagnosis Status.
Ai is latent determined by fi, a vector of symptoms and genetic covariates which correlate with true ADHD status Ai = 1(fi*rhoi + ni >= 0) di represents a correct diagnosis by a physician, determined by wi a vector of covariates which correlate with a correct diagnosis: di = 1(wi*gamma + UI >= 0) The reported diagnosis in the data is defined as: Di = Ai*di = 1(fi*rho + ni >= 0, wi*gamma + UI >= 0)
A DAG showing the following causal mechanism: Hereditary/Genetic Traits -> ADHD ADHD -> Patient Selection & Educational and Labor Market Outcomes Socioeconomic and Demographic Variables -> Patient Selection & a Diagnosis Patient Selection -> a Diagnosis Diagnosis -> Educational Attainment and Labor Market Outcomes
Title: Hidden in Plain Sight: The Prevalence and Impact of Undiagnosed ADHD Abstract: While diagnostic errors are prevalent throughout health and mental health fields, little is known about their long-term consequences. In this paper, I assess the long-term education and labor market effects of missed Attention-Deficit and Hyperactivity Disorder (ADHD) diagnoses. Doing so is challenging for two reasons. First, a person's true ADHD status is unobserved; only their diagnosed status is known. Second, even if diagnostic errors exist, they are likely akin to non-classical misclassification errors and therefore endogenous. To overcome these empirical challenges, I use a partial observability model to estimate an individual's true ADHD status and their probability of underdiagnosis. Then, I estimate their effect on educational attainment and adult labor market outcomes. I find that ADHD is underdiagnosed -- with large differences by race, gender, and socioeconomic backgrounds -- and underdiagnosis reduces educational attainment and worsens labor market outcomes. I also find evidence of vicious intergenerational cycles in which inequity in mental health care drives and is driven by socioeconomic inequities. JEL: I14, I15, I24, I25, J13 Keywords: ADHD, Underdiagnosis, Child Development, Mental Health, Measurement Error, Partial Observability, Inequity