This story was originally published by MinnPost.
When it comes to health equity, data collection is a powerful tool used to inform policy decisions to reduce gaps in health outcomes.
Data can be a way to form opinions and make change, said Tetyana Shippee, an associate professor at the University of Minnesota’s School of Public Health.
“If we don’t measure something, in many ways, it’s not real. If we can’t show that a phenomenon exists by measuring and reporting on it, then it’s very easy to say, regardless of people’s individual experiences, that’s not a system-wide issue, it’s not a larger issue that we need to invest any resources in,” she said.
Accurate data collection is the first step for useful findings, according to Dan Fernandez-Baca, the director of the Center for Health Statistics of the Minnesota Department of Health (MDH).
Fernandez-Baca works with base population data for MDH; things like vital statistics, births, deaths and population surveys. He said that while data can help discover disparities, the conclusions drawn strongly depend on the categories the researchers create.
Getting people to self-identify correctly is often a struggle, Fernandez-Baca said.
“If you are Middle Eastern, do you select white, or do you select other? There are all these issues at the point of data collection that get complicated,” he said. “It’s a bit of a process to clean all of that data and put them all into categories, but it’s work worth doing.”
Sometimes, results and reports will be published several years after the initial collection. As a result, there can be a lag between the present-day reality and the findings.
The Center for Health Statistics tries to use real-time data for things that are tracked that way, like life and birth, Fernandez-Baca said. He attributes the delay from collection to publishing because sometimes researchers are waiting for doctors to enter data into a system and also because cleaning the data can be time and resource intensive.
How are categories created?
Another difficulty in the data-sphere is collecting data on smaller population sizes, Shippee said.
Her work has focused on equity in access to quality of services for older people in long-term care facilities. She’s noticed that there are fewer non-white communities in those facilities, so often counting them in the data would cause the findings not to be statistically sound.
“If the sample size is not large enough, and we want to see if a difference is statistically significant, meaning if we were to do it in a different sample, we’d see the same results,” said Shippee. “Then the worry is that if a sample size is so small and you find it’s not statistically significant, it’s actually erroneous and biased findings because there were not enough people there to detect the difference, and so I’m always sort of mindful of this tug and pull.”
Shippee’s research found that Black people and American Indian people in nursing homes report lower quality of life than white people in the homes. In 2015, she shared her work with the Department of Human Services (DHS), she said.
Since then, two initiatives to combat those disparities have been implemented, including an emphasis on data collection for people who are non-native English speakers and a DHS initiative that pays facilities that show improvement in addressing racial disparities.
“From my standpoint, data is the like mandatory first step to be able to understand really any issues around us,” Shippee said.
Other times, groups get lumped together, despite their population size. MDHs analysis of children who have received the MMR vaccine, did not differentiate ethnic Africans from African Americans, Lynn Bahta, the immunization clinical consultant for MDH, said to MinnPost in June.
MDH used birth certificate data to identify ethnic Somali children and matched that data to the immunization data. But those same data markers weren’t available for African American children, Bahta said.
“We are trying to figure out ways that we can get data either from the department of human services or to figure out ways we can refine the birth certificate to better identify that denominator so that we can parse it out,” Bahta said. “We’re hoping that we can do that with our broader immunization rates. It’s in process, but we haven’t gotten to that specifically yet.”
Parsing out those differences are essential to get the complete picture, Shippee said.
“Stop lumping people from groups that are not meaningfully combined. What do your results even mean when you have a group, let’s say BIPOC, when they have such different lived experiences? I think researchers, myself included, are often guilty of these aggregations that then, in the end, are not as meaningful for policy change.”
Researchers also find it challenging to parse out populations when researching multi-racial individuals or immigrant populations. Fernandez-Baca explained how country of origin might not necessarily impact everyone the same way – and researchers should be aware of that.
“I was born in Mexico. I came to the U.S. at the age of one, so there’s not a lot of influence that country of origin had on my health outcomes. But if somebody was born in a certain country and lived there ‘til the age of 25 and then moved to the U.S., they will have different health outcomes,” he said. “Any one measure will not be a complete view of what a person had experienced, which is why data is just kind of intended to be a guidepost for the healthcare community.”
Determining identities
To combat the ambiguity of the country of origin category, Fernandez-Baca recommends parsing data by the “spaces” people exist in.
“Space: the places we live, work and play; they have different impacts on our health and our health outcomes. People who are not necessarily born in the United States come from an environment that has different food regulations, has different environmental health considerations to be made, and those things affect a group’s health outcomes,” Fernandez-Baca said. “It’s not about understanding what do immigrants do, but more of a what are the places and how do the spaces that we live in affect us.”
The Center for Health Statistics has a guide for Health Equity Data Analysis that aims to focus on people’s stories when working with data to avoid the stigmatization that numbers can often bring.
“That is intended to say, ‘These are the numbers for your population in your area, but what is the story? Is the particular health outcome we’re seeing in your population because of the people, or is it because you live in a food desert? Or is it because you have this external factor that is affecting you?'” Fernandez-Baca said. “As we get closer and closer to people’s identities, it becomes more and more important to get that context, that qualitative context of what the situation is.”
Where the data takes us
Data gives public health entities a place to start, from which that evidence can then refine the question at hand, Fernandez-Baca said.
“It is a process of giving clues, searching for context, refining the data collection, and then looking again,” he said. “We can refine what the data is collecting. And so we say, ‘It’s not necessarily a matter of obesity; it is a matter of food deserts.’ So now our refined question isn’t ‘What is the BMI of this population?’ It is ‘What is the level of access to healthy foods and what is the walkability factor of your neighborhood?’”
MDH makes recommendations to the governor’s office based on what areas data shows a need. The health commissioner then brings proposals to the governor, some of which are included inbiennial budget recommendations.
In the 2024-25 budget recommendations, there are more than 50 health proposals. One of those, for example, is a proposal for a Comprehensive Drug Overdose and Morbidity Prevention Act, which cited data that showed a state-wide increase in drug overdose deaths from 2020 to 2021.
This article first appeared on MinnPost and is republished here under a Creative Commons license.