Who's being insured?

Some insurance companies, pushed by federal regulators and accrediting bodies, say they can play a role in addressing health inequity by targeting outreach and resources to consumers who need support.

But pinpointing disparities in outcomes means first collecting race, ethnicity and other demographic data from members—who may be skeptical of sharing personal details. It’s a process many insurers are still fine-tuning.

 “We’re not going to be able to [reduce health disparities] if we don’t know more about where the disparities are,” said Abner Mason, CEO and founder of SameSky Health, a company that contracts with health insurers and provider organizations to connect patients with care navigation resources tailored to their cultural backgrounds. “We know they exist, but we need to be able to measure them.”

 Many insurers have started by collecting race and ethnicity data, with the goal of including sexual orientation and gender identity information down the line. While they say no single tactic is foolproof, they’ve tried a variety of strategies to figure out how their members identify themselves. The undertaking requires time and resources as organizations send out surveys and update their technology—but it can pay off for those that put the data to use.

“Once you’re able to identify a disparity, you can create programs or policies that may help eliminate that disparity, and can actually save the insurer money by getting care to a disparate patient sooner,” said Duane Reynolds, president of Chartis Just Health Collective, an equity-focused division of healthcare advisory firm Chartis Group.

Medicare and Medicaid plans have the most complete race and ethnicity data, while employer-based or individual plans are the furthest behind, according to Caroline Pearson, a senior vice president of NORC at the University of Chicago, a nonpartisan research organization.

Marketplace plans tend to be somewhere in the middle: During the 2022 open enrollment period for federal marketplace plans, race or ethnicity data was collected for about 70% of enrollees, according to researchers at the Center on Health Insurance Reforms at Georgetown University. In state-based marketplaces, race or ethnicity was captured for approximately 77% of enrollees.

The percentages may soon increase, as the federal government rolls out programs to include equity data in quality measures.

The Centers for Medicare and Medicaid Services will begin requiring the issuers of marketplace and Medicare Advantage plans to submit race and ethnicity data for members in January 2023. By 2024, marketplace plans will have to submit any available quality data, stratified by race, for measures such as colorectal cancer screening, high blood pressure management and prenatal care.

Some accrediting bodies are also adding enhanced data collection standards to their requirements for insurers, such as carriers of Medicaid, exchange and employer plans in some states.

In anticipation, insurers have launched surveys via mail, websites and mobile apps to collect race, ethnicity and other information on their members. They’re tapping data brokers and employers for help and directing case workers and call centers to ask about identity during customer interactions.

But there are kinks in each strategy, requiring insurers to take a multipronged approach, Reynolds said.

Surveys could have selection bias. The member may not fall into any of the provided categories or want to give the information at all, making the data incomplete.

The definition of race and ethnicity can differ among brokers, health plans and other entities, posing difficulties with verification and combined datasets. Adding optional race and ethnicity questions on enrollment forms for employer-based health insurance requires an extra layer of communication between insurers and employers.

And in cases where an insurance representative is involved, the worker may assume the race of the member by their appearance or the sound of their voice, which can cause inaccuracies.

“What is missing right now is there is no systematic or standardized collection,” Reynolds said.

Federal agencies provide some direction for insurers when it comes to sorting data. The Office of Management and Budget established broad demographic categories in 1997 for federal reporting; in June, agency officials announced a review of the standards, expected to be completed by 2024. The standards comprise five broad categories for race—American Indian/Alaska Native, Asian, Black/African American, Native Hawaiian/other Pacific Islander, and white—and one category, Hispanic/Latino, for ethnicity.

The Office of Minority Health in the Department of Health and Human Services could play a role in creating information standards for healthcare, Reynolds said. The agency offers guidance on data collection, including which categories to include.

For ethnicity, the office distinguishes several nationalities under the Hispanic designation, such as Mexican, Puerto Rican and Cuban. When it comes to race, it recommends health organizations offer white, Black/African American, American Indian/Alaska Native, Asian and Pacific Islander, with various subcategories. If a person identifies as Asian, for example, they could then select from a list of nationalities.

Critics say the guidance could be more expansive. A particular concern arises for people who identify as more than one race or ethnicity.

Historically, if a person selected two races, they would be counted as two different people. To keep the population count accurate, insurers and other health organizations sometimes instead group them in a “multiracial” or “other” category—excluding them from analysis.

“When you select more than one race, none of your race data gets reported,” said Dr. Siobhan Wescott, associate professor in the department of health promotion at the University of Nebraska Medical Center. “You are dumped into the category of two or more races and there is no more analysis.”

She described the collection process as a “push-pull between consistency and details.” Capturing the intricacies will require revising existing infrastructure and working across multiple departments to determine new data categories and subsets.

Given the shortcomings of each strategy on its own, insurers have implemented multiple approaches simultaneously to reach a wide swath of members.

Blue Cross Blue Shield of Massachusetts started ramping up its efforts to collect race and ethnicity data from its members a few years ago.

It had previously estimated race and ethnicity using members’ addresses and surnames—relying on a method developed by RAND, a policy think tank—to assess disparities internally. But the insurer decided to step up its efforts in summer 2020, prompted by inequities underscored by the COVID-19 pandemic and discussions about racism across industries in the wake of George Floyd’s murder by a Minneapolis police officer.

“That’s when we really became a lot more systematic,” said Dr. Mark Friedberg, senior vice president for performance measurement and improvement at BCBS of Massachusetts.

At the end of 2020, the company added fields to its app for members to self-report race and ethnicity data, using OMB’s categories.

The app also includes an “other” category for race—which the insurer says it includes in analyses—and members are allowed to select multiple categories. But only about one-third of the insurer’s members use the app, so the team found other ways to engage its customers.

The company began mailing paper surveys earlier this year to members who hadn’t responded on the app—a multimillion-dollar endeavor, according to Friedberg. So far, it’s mailed 700,000 surveys requesting information on about 1.4 million members. Each survey asked about all members on a policy and included a postage-paid return envelope, with the option to respond online.

BCBS of Massachusetts said it’s too early to determine the survey’s return rate.

The insurer is also planning to add race and ethnicity data collection to the online enrollment process and is outlining pilots for some of its partner-employers to share members’ self-reported race and ethnicity data directly.

As of July, the insurer had gathered race and ethnicity data from about 20% of its members. It aims to get to at least 80% in the next two to three years. In the meantime, the insurer is still using what it calls “statistically estimated” race and ethnicity data to supplement the information.

The company’s efforts are beginning to produce insights. Last year, it published a report breaking out measures such as high blood pressure management and well-child visits by race and ethnicity.

“We shared that with the community on our website, as part of our commitment that we’re making to hold ourselves accountable for reducing and ultimately eliminating inequities in care,” said Deanna Fulp, the insurer’s senior director of health equity.

BCBS of Massachusetts also uses the data it gathers to create reports for providers in its alternative quality contracts, a value-based payment model. The reports measure race and ethnicity disparities specifically in the provider’s patient population. The company is adding a component to its alternative quality contracts next year tying provider financial incentives to equity of care.

Friedberg said it was important to get started and not let “perfect” be the enemy of the good.

“People should just go ahead and do it,” he said of insurers collecting race and ethnicity data. “There’s no reason any health plan can’t start right now.”

Health Net, a Centene subsidiary in California, has used its own member data to shape outreach strategies, such as those concerning COVID-19 vaccines.

For years, the insurer has collected race and ethnicity information through forms members fill out when applying for health insurance. The company also asks members about such demographics when they call customer service, and the data might be shared by healthcare providers too.

After crunching the numbers on who had been vaccinated for COVID-19 and stratifying the results by race and geography, Health Net realized Black members had lower rates of vaccination than other racial groups in certain areas.

It responded by hosting events with Black-led community organizations where residents could get vaccinated. The company’s goal was to improve access by holding events on the weekends and other off-hours, along with providing a space where members would feel comfortable, said Dr. Pooja Mittal, chief health equity officer at Health Net.

In total, Health Net has hosted more than 210 mobile vaccination clinics and administered upwards of 22,000 vaccines.

“Use every opportunity to get data,” Mittal advised. “It really allows us to inform how we develop our programs and targeted initiatives.”

Download Modern Healthcare’s app to stay informed when industry news breaks.

Elevance Health, formerly Anthem, is using its member information to implement programs aimed at addressing disparities in maternal health and mortality. The insurer found Black members who gave birth were two times more likely to experience extreme preterm births compared with their white counterparts. Black patients in the insurer’s commercial membership were also 25% more likely to have a C-section. Elevance responded by forming partnerships with local organizations, and has awarded grants to not-for-profits that support women’s health.

The insurance company gets most of its members’ race and ethnicity data from self-reported information submitted on enrollment and eligibility forms, especially for Medicare and Medicaid. It plans on adding questions to its member-facing app, as well as striking partnerships with employers to share data on workers participating in sponsored plans.

But hurdles still exist, leading Elevance to fund a report conducted by the think tank Urban Institute examining the barriers to collecting race and ethnicity data for insurers. The insurer didn’t respond to a request for comment about the current percentage of members for whom it has race or ethnicity data.

Additional demographic factors such as sexual orientation and gender identity are also important, but “we just felt that we had to start somewhere,” said Dr. Shantanu Agrawal, chief health officer at Elevance.

The report, released in July, identified member trust as a core component of data collection. If individuals aren’t comfortable providing the information, it won’t matter if the insurer reaches them through apps, paper, the phone or another way.

“We have to do a better job talking about the value of the data, and how we would use the data to drive health outcomes,” Agrawal said. “We’ve got to be much better about crystalizing that value proposition.”

Many consumers share concerns about how their data may be used, according to a Deloitte survey cited in the Urban Institute report. Some worry about being denied insurance coverage, not having claims covered or experiencing discrimination because of their race or ethnicity. Employers may also be less inclined to provide additional information about their workforce to insurers out of fear it will affect their own cost of coverage.

Federal laws prohibit programs or activities that accept certain federal funds—such as insurance plans established or administered as part of the Affordable Care Act—from discriminating based on race and national origin, among other identities. Still, such consumer uneasiness has insurers weighing the best time to collect the data and the best way to ask.

“Do they collect it as they are writing policies?” Reynolds said. “That can create a bit of angst for consumers, who may feel that they’re going to use that information against them in terms of rating them from risk scoring of their actual premiums.” Individuals may also not want their employers to share their data.

Some health insurers have sought to be transparent with members about what race and ethnicity information is being used for, leading to increased response rates.

New York State of Health, the state’s health insurance marketplace, updated its enrollment process for 2022 specifically to enhance race and ethnicity data collection for state-based marketplace plans. The changes included categories to capture a more diverse range of racial and ethnic identities. The agency also added detailed explanations about why the exchange was collecting race and ethnicity information and trained customer service representatives to answer questions.

After implementing the changes, the exchange saw its response rate rise from 81% to about 90%. A spokesperson said the improved data on race and ethnicity will help the marketplace identify and address health disparities as part of its quality improvement efforts.

“We need to actively work for the social outcomes that we desire,” NORC's Pearson said. “At the end of the day, better data is going to drive better decision-making and better behavior.”