Unstructured clinical notes could be a key source for mental health detection and care in primary care settings. A growing use of LLMs to extract mental health indicators from unstructured clinical notes has the potential to aid in this effort. However, these technologies do not account for provider bias at the point of data entry. In this video vignette survey experiment study, we explore how mental health information in written clinical notes differ in frequency, content, and quality by patient race-gender presentation. Leveraging the strengths of an experimental design with 111 practicing healthcare providers in the US, we find substantive qualitative differences in clinical notes and significant quantitative differences in the assessment of patient mental health based on the race-gender presentation of the patient. Specifically, providers documented fewer mental health concerns and fewer requests or recommendations for mental health follow-up care for Black patients. They also assessed that White patients were in significantly worse mental health compared to their Black counterparts. Providers also included more justifying language for alcohol use and inquiries about drug use for male patients compared to female patients. These differences in the source data for LLM extraction have implications for health equity, making it so these new innovations do not avoid existing bias in mental health detection and care. Further use of these technologies should carefully consider how to account for and address this bias to improve upon existing mental health care disparities.

The widespread adoption of personal health devices has introduced a new source of health data for patients and providers to use in healthcare settings. Using patient-generated health data (PGHD) in healthcare settings has been found to improve patient-provider interactions and care outcomes. However, rates of PGHD sharing vary across the population. The decision to share one’s PGHD with providers is influenced by past healthcare encounters including quality of care, medical mistrust, and healthcare discrimination. Furthermore, those experiences and patients’ responses to them should be examined with an intersectional lens to acknowledge how gendered racism impacts health and healthcare. Using the 2022 Health Information National Trends Survey (HINTS) and logistic regression, this study shows how racial discrimination, healthcare quality, and medical mistrust are associated with patients’ willingness and decision to share PGHD with providers among different race-gender identities. Factors associated with patients’ perceived willingness to share data are not the same factors that are associated with actually sharing PGHD with providers. Perceived racial discrimination in healthcare settings was only significantly associated with willingness to share PGHD among Black men and white men. Meanwhile, poor overall healthcare quality was significantly associated with incidences of sharing data among Black and Hispanic women. These varied patterns across intent to share and choice to share, as well as across race-gender identity, reveal important insights into barriers to effective integration of PGHD in clinical spaces and the use of this technology to address healthcare inequities.

Inequality research has often used non-Hispanic Whites as the reference category in measuring U.S. racial and ethnic health disparities, with less attention paid to diversity among Whites. Immigration patterns over the last several decades have led to greater ethnic heterogeneity among Whites, which could be hidden by the aggregate category. Using data from the National Health Interview Survey (2000–2018), we disaggregate non-Hispanic Whites by nativity status (U.S.- and foreign-born) and foreign-born region of birth (Europe, Former Soviet Union, and the Middle East) to examine diversity in health among adults aged 30+ (n = 290,361). We find that foreign-born Whites do not have a consistent immigrant health advantage over U.S.-born Whites, and the presence of an advantage further varies by birth region. Immigrants from the Former Soviet Union (FSU) are particularly disadvantaged, reporting worse self-rated health and higher rates of hypertension (high blood pressure) than U.S.-born and European-born Whites. Middle Eastern immigrants also fare worse than U.S.-born Whites but have health outcomes more similar to European immigrants than to immigrants from the FSU. These findings highlight considerable diversity in health among White subgroups that is masked by the aggregate White category. Future research must continue to monitor growing heterogeneity among Whites and consider more carefully their use as an aggregate category for gauging racial inequality.