Operationalizing racialized exposures in historical research on anti-Asian racism and health: a comparison of two methods

Background

Addressing contemporary anti-Asian racism and its impacts on health requires understanding its historical roots, including discriminatory restrictions on immigration, citizenship, and land ownership. Archival secondary data such as historical census records provide opportunities to quantitatively analyze structural dynamics that affect the health of Asian immigrants and Asian Americans. Census data overcome weaknesses of other data sources, such as small sample size and aggregation of Asian subgroups. This article explores the strengths and limitations of early twentieth-century census data for understanding Asian Americans and structural racism.

Methods

We used California census data from three decennial census spanning 1920-1940 to compare two criteria for identifying Asian Americans: census racial categories and Asian surname lists (Chinese, Indian, Japanese, Korean, and Filipino) that have been validated in contemporary population data. This paper examines the sensitivity and specificity of surname classification compared to census-designated "color or race" at the population level.

Results

Surname criteria were found to be highly specific, with each of the five surname lists having a specificity of over 99% for all three census years. The Chinese surname list had the highest sensitivity (ranging from 0.60-0.67 across census years), followed by the Indian (0.54-0.61) and Japanese (0.51-0.62) surname lists. Sensitivity was much lower for Korean (0.40-0.45) and Filipino (0.10-0.21) surnames. With the exception of Indian surnames, the sensitivity values of surname criteria were lower for the 1920-1940 census data than those reported for the 1990 census. The extent of the difference in sensitivity and trends across census years vary by subgroup.

Discussion

Surname criteria may have lower sensitivity in detecting Asian subgroups in historical data as opposed to contemporary data as enumeration procedures for Asians have changed across time. We examine how the conflation of race, ethnicity, and nationality in the census could contribute to low sensitivity of surname classification compared to census-designated "color or race." These results can guide decisions when operationalizing race in the context of specific research questions, thus promoting historical quantitative study of Asian American experiences. Furthermore, these results stress the need to situate measures of race and racism in their specific historical context.

Effect of Resident Gender and Surname Origin on Clinical Load: Observational Cohort Study in an Internal Medicine Continuity Clinic

BACKGROUND

Studies show patients may have gender or racial preferences for physicians.

OBJECTIVE

To determine the degree to which physicians' gender and name characteristics influenced physician clinical load in medical practice, including patient panel size and percent of slots filled.

DESIGN

Observational cohort study of a continuity clinic site in Rochester, MN, from July 1, 2015 to June 30, 2017 ("historical" period) and July 1, 2018 to January 30, 2020 ("contemporary" period).

PARTICIPANTS

Internal medicine resident physicians.

MAIN MEASURES

Resident gender, name, and race came from residency management system data. Panel size, percent of appointment slots filled ("slot fill"), panel percent female, and panel percent non-White came from the electronic health record. Multivariable linear regression models calculated beta estimates with 95% confidence intervals and R² for the impact of physician gender, surname origin, name character length, and name consonant-to-vowel ratio on each outcome, adjusting for race and year of residency.

KEY RESULTS

Of the 307 internal medicine residents, 122 (40%) were female and 197 (64%) were White. Their patient panels were 51% female (SD 16) and 74% White (SD 6). Female gender was associated with a 5.3 (95% CI 2.7-7.9) patient increase in panel size and a 1.5% (95% CI -0.6 to 3.7) increase in slot fill. European, non-Hispanic surname was associated with a 5.3 (95% CI 2.6-7.9) patient increase in panel size and a 4.3 percent (95% CI 2.1-6.4) increase in slot fill. Race and other name characteristics were not associated with physician clinical load. From the historical to contemporary period, the influence of name characteristics decreased from 9 to 4% for panel size and from 15 to 5% for slot fill.

CONCLUSIONS

Female gender and European, non-Hispanic surname origin are associated with increased physician clinical load-even more than race. While these disparities may have serious consequences, they are also addressable.

Nomen Est Omen: Medical and Pharmaceutical Occupations in Slovenian Family Names

The heritage of Slovenian house names and surnames reflects, among others, the former medicine and pharmaceutical occupations, midwifery, and folk medicine practices, and besides that, also health status and illnesses of people. Surnames, which are especially strongly intertwined with family, local and social history, are closely related to folk medicine and magic. Unlike house names (vulgo), which are the usual nicknames for physical and mental characteristics and abilities, surnames denote medical occupations and medicinal folk practice as such. According to the most recent data (as of January 1, 2020) of The Statistical Office of the Republic of Slovenia, at least 40 surnames reminiscent former medical or pharmaceutical professions. These newly discovered digital data in open access are precious for the history of medicine because they allow comparing surnames geographically, by frequency, and through the time.

A new method for estimating race/ethnicity and associated disparities where administrative records lack self-reported race/ethnicity

OBJECTIVE

To efficiently estimate race/ethnicity using administrative records to facilitate health care organizations' efforts to address disparities when self-reported race/ethnicity data are unavailable.

DATA SOURCE

Surname, geocoded residential address, and self-reported race/ethnicity from 1,973,362 enrollees of a national health plan.

STUDY DESIGN

We compare the accuracy of a Bayesian approach to combining surname and geocoded information to estimate race/ethnicity to two other indirect methods: a non-Bayesian method that combines surname and geocoded information and geocoded information alone. We assess accuracy with respect to estimating (1) individual race/ethnicity and (2) overall racial/ethnic prevalence in a population.

PRINCIPAL FINDINGS

The Bayesian approach was 74 percent more efficient than geocoding alone in estimating individual race/ethnicity and 56 percent more efficient in estimating the prevalence of racial/ethnic groups, outperforming the non-Bayesian hybrid on both measures. The non-Bayesian hybrid was more efficient than geocoding alone in estimating individual race/ethnicity but less efficient with respect to prevalence (p<.05 for all differences).

CONCLUSIONS

The Bayesian Surname and Geocoding (BSG) method presented here efficiently integrates administrative data, substantially improving upon what is possible with a single source or from other hybrid methods; it offers a powerful tool that can help health care organizations address disparities until self-reported race/ethnicity data are available.