by Kevin Schofield

---

About a year ago, I wrote about the “Hispanic health paradox”: the effort to understand why the country’s Hispanic population as a whole has better health outcomes than researchers predict based on their levels of household income, education, and insurance coverage. A study concluded that looking at the aggregate numbers for the entire population of Hispanic Americans was burying some disturbing health trends that show up when we look at subgroups.

This weekend’s read is a news article published by the Journal of the American Medical Association that makes largely the same point for the Asian American community. There are roughly 24 million Americans with Asian heritage. But “Asian” covers approximately 60% of the world’s population, including the two most populous countries in the world, India and China, as well as dozens of other smaller ones with diverse cultures, populations, climates, and degrees of social isolation (and thus genetic isolation) throughout their histories. For someone to mark their race or ethnicity as “Asian” on a form alone tells us very little about where they come from, what languages they speak, their religious or other cultural practices, their diet, or the DNA inside their cells. 

And yet the federal government’s Office of Management and Budget, which largely dictates the set of options for specifying one’s race on census forms and countless other government data-collection instruments, often specifies just that one option to cover Asian Americans — and sometimes lumps “Native Hawaiian or Pacific Islander” in there as well. 

The article begins by recalling how Asian Americans were often de-prioritized during the initial rollout of COVID-19 vaccines in 2021, due in part to the “model minority” myth, the belief that Asians have an extraordinary ability to overcome hardship to succeed in American society. But in reality, according to the article, in 2020, “the Asian American population experienced double the percentage of deaths due to COVID-19 than the White population and as much as a 53% higher case-fatality rate.”

There are two major risks created by over-aggregation of health data. The first is that important differences in sub-populations are buried and never noticed. The second is that in the cases where a specific sub-population (e.g., people of Japanese heritage) has been studied, the results will be extrapolated inaccurately to apply to everyone of Asian heritage.

Despite the barrier put in place by the government’s broad-brush categorization of the “Asian” population, researchers are working to disaggregate health data to understand where there are underlying trends that are going unnoticed, starting with states such as California, Hawai‘i, and New York that have diverse populations. So far, they have found that breast cancer contributes more deaths from cancer for Indian and Filipino women than for other Asian women (and for white women). Stomach and liver cancers, on the other hand, caused more deaths for Chinese, Korean, and Vietnamese people. Also, the Filipino population has the highest risk of cardiovascular disease, while Chinese Americans have the lowest comparative risk. A study looking at vaccination rates in New York City found that overall rates were high for Asian Americans, but the aggregate numbers were skewed up by high vaccination rates for Chinese Americans while the rates in the local Nepali community were low.

Disaggregating this kind of health data helps move us closer to what health care professionals have termed “personalized medicine,” where medical treatments can be customized for the individual — including potential medicines, as we’re starting to see now with cancer treatments that take a sample of a tumor and use it to create a one-off medicinal treatment that matches its DNA. But before we get there (and long before those kinds of treatments become affordable for everyone), health care professionals will look at each of us as the intersection of a collection of demographic subgroups and their respective health profiles. To that end, smaller demographics are better, up to a point: We want to know that we belong to a group that is small enough to have a differentiated health profile that can help doctors know how to care for us, but still large enough so that those differences are statistically significant, i.e., very likely to hold true for us too.

Defining those groups is still a big challenge, however. For example, if we identify a demographic group with a differentiated health outcome, how can we tell how much of that is genetics and how much can be attributed to the environment or to cultural practices? And for groups such as Asian Americans, how do we measure the extent of cultural adaptation? Traditionally, English language proficiency has been used as a proxy for an immigrant community’s level of adaptation into American culture, but researchers are now recognizing, in retrospect, that it’s a poor proxy that fails to represent all the different dimensions of cultural practices. In a related effort, the National Academies last year recommended that much of the population descriptors we frequently use to capture demographics, including “Black” and “Asian,” are sociopolitical in origin, rarely have a useful correlation with our genetics, and have limited use in health care research and practice. Further complicating things in the health care world is the country’s long-term trend toward a population that is multiracial, which makes it even more difficult to disaggregate health data.

Nevertheless, there is now a growing recognition that for the purposes of advancing health care we need much more disaggregation of demographic data on Asian Americans — in addition to Hispanic Americans, and probably Americans of all races and ethnicities. So look for a lot more choice in the boxes to check on those government forms in the near future.

Researchers Are Working to Disaggregate Asian American Health Data—Here’s Why It’s Long Overdue

---

📸 Featured image via Have a nice day Photo/Shutterstock.com.

---