Healthcare Data Can Transform Benefits

A Johnson & Johnson employee filed a proposed class-action lawsuit in February claiming the company had cost its health plan beneficiaries millions of dollars by poorly managing pharmaceutical benefits.

Ann Lewandowski filed suit in federal court in New Jersey claiming the self-insured employer overpaid its third-party administrator, Express Scripts, for prescriptions. In doing this, the company failed to meet its responsibilities as a fiduciary for the plan under the Employee Retirement Income Security Act of 1974 (ERISA), she says.

In several instances, employees paid thousands of dollars more for prescription drugs through the plan than they could have paid with cash at pharmacies, the lawsuit says. In just one instance cited in the filing, plan participants’ cost for the generic multiple sclerosis medication teriflunomide (Aubagio) was $12,239.69 for a 90-day prescription; it was $40.55 at Wegmans.

Lewandowski’s lawsuit was, in part, built on the vast amounts of healthcare data now available to the public. Over the past five years or so, federal regulations have mandated data transparency by healthcare providers and insurers. And self-insured employers—like Johnson & Johnson—are expected to be good fiduciaries by using data to determine how to best spend their employees’ healthcare dollars.

The proliferation of data, in theory, could provide powerful insights into an employer’s health plan, save healthcare dollars, and help plan participants choose the best, low-cost providers in their networks. But potential uses of data don’t necessarily align with its availability. Even larger companies aren’t getting and parsing the data to its full potential. And most small businesses, which have little money for data analytics and even less human-resource heft, are at an even greater disadvantage. The market is responding, though, with vendors scurrying to offer usable, meaningful data on a small scale. 

Noise in the Data

The U.S. Congress in recent years has passed several rules and mandates on healthcare data transparency, notably measures contained in the Consolidated Appropriations Act of 2021 (CAA).

The spending bill prohibited insurers from using gag clauses that restrict carriers from providing healthcare cost information. It also required insurance companies and self-funded employer plans to make public on their websites machine-readable files with healthcare spending data. The No Surprises Act, included in the CAA, created a dispute resolution pathway for patients who receive “surprise” bills from emergency services and out-of-network providers. For example, a person may have surgery at an in-network hospital and then later receive a separate bill from the anesthesiologist on duty who was out of network.

The Hospital Price Transparency Rule, which also took effect in 2021, required hospitals to make their prices available online to consumers. These transparency regulations are intended to make healthcare pricing clear so people can shop for treatment like they would for a car or a home. Using their insurer’s website, health plan beneficiaries should be able to compare their cost-sharing responsibility among different providers in their network for treatments such as knee replacement surgery, prostate cancer screening, hysterectomy, or skin biopsy.

In the Johnson & Johnson lawsuit, the plaintiff compared data from the National Average Drug Acquisition Cost database, other pharmaceutical pricing information that was publicly available online, and the Johnson & Johnson estimation cost tool for her plan.

Having cost data in the market, in theory, would allow consumers to choose the best care at the lowest price. It would also enable employers and insurance companies to negotiate the best rates for providers in their networks.

But the data, from the beginning, was problematic.

“Health systems had to post rates online, and then health plans did,” says Keith LoMurray, vice president of platform and data at Kyruus Health, a healthcare software company based in Boston. “There was lots of noncompliance and lack of standardization. There was no shared schema on how to post it, and it was a couple of years before CMS [the U.S. Centers for Medicare & Medicaid Services] even considered any kind of enforcement.”

Machine-readable files, the format required for the data, are impossible for consumers to use and nearly so for employers or brokers in any meaningful way.

“If you are human resources or a small business, your computer would probably crash trying to download a machine-readable file because they are too big to open,” LoMurray says. “The end user can’t take it and run with it; it’s not structured to be used that way.”

The data are “intentionally obfuscated,” according to Todd Gottula, president and co-founder of Clarify Health Solutions, a health data analytics company out of San Francisco. He says the sea of information comes to about four terabytes and it requires data science experts to make sense of it.

That is just one of the issues that make the data challenging to use. Different payers are also complying with the regulations at different levels—some have very thorough information, while others have missing or patchy data.

There’s also a lot of noise in the data, says Rafiq Ahmed, co-founder and CEO of Serif Health, another health data analytics company based in San Francisco. This is largely due to what he calls ghost (or zombie) rates, which add volume but not clarity to the information. Healthcare providers are required to post information for every service available in a particular practice, and many report for services they don’t provide, which muddies the data. This means an orthopedic surgeon may list reimbursement rates for psychotherapy or there could be data from a cardiologist for an ACL repair. All those irrelevant codes must be filtered so the information isn’t skewed.

Sometimes information is missing from the data, including physician fees and some out-of-network rates. Some data are also simply inaccurate, Ahmed says. “A lot of our job is to tell clients when the data is good and when it’s not so good. There are some plans where we don’t trust any of their data.”

Small employers (those with 50 or fewer employees) that get insurance through state exchanges are particularly challenged by the regulations and data consistency issues. These marketplace plans are community-rated, which aims to keep high-cost beneficiaries from paying exorbitant premiums. Community rating prohibits insurers from increasing the premiums of people in a geographic area based on their gender, health status, or history of claims.

Because premiums are held steady, community-rated plan carriers typically don’t perform the extensive underwriting that risk-rated plan carriers do, because they don’t need to justify their renewal rates (age and tobacco use—in some states—are the only things that can increase plan beneficiary premiums). Large group markets, on the other hand, are risk rated. With risk rating, carriers can change premiums based on past or future risk of health costs. It’s in the carriers’ interests to track and analyze as much data as possible to create and justify premium increases to covered employers.

The CMS provides some information in public-use machine-readable files for small business plans on state exchanges. The files include some premium rates and the essential benefits required to be covered by the plans. But this information is limited and inconsistent, according to David Schulman, director of marketing and sales at the benefits management company Ascela (a PCF company).

“For small clients to get that data, they would have to put up a fight to get it,” Schulman says. “If you can give a small client data they haven’t had before, it would give them back some control over their plans. Participation and buy-in at every level helps create conditions to impress upon the carriers that they need to update these older plans.”

Outside Assistance

About 98% of companies in the United States have fewer than 100 employees. Those businesses are looking to their vendors to make sense of this new influx of healthcare data. In its 2024 benefits broker survey, employee benefits provider Optavise found 94% of broker respondents say their employer clients rely on them for compliance and reporting of healthcare benefits.

Smaller companies—especially those with 50 or fewer employees—will have a hard time getting meaningful data using only their plan participants, LoMurray says. Their employee sample is so small that the amount of data they have from carrier contracts is not particularly helpful. With such a small number of claims, it’s difficult to tell which hospitals may be overpriced or if an employer could negotiate a better rate with a different insurer.

That said, all of the information an employer needs is now, technically, available. The next phase of data evolution will be off-the-shelf tools smaller businesses can use to benchmark and provide healthcare cost insights, LoMurray says. Several companies are working on this, but many are in the early stages, figuring out how to offer the data on a small scale without being cost-prohibitive.

 “Some vendors are trying to help small businesses make more informed decisions and allow them to do things like consult with payers to say, ‘You should renegotiate these costs,’” LoMurray says.

Schulman says a handful of vendors work with smaller employers to use available data to get important insights.

One vendor in this space is insurance carrier Self Fund Health, which works with small, self-funded employers in Wisconsin and some other midwestern states. The company crunches machine-readable files to determine what providers charge in an area. When possible, it then steers plan beneficiaries to identified top-quality, low-cost providers.

Self Fund Health co-founder and CEO Jonathan Baran says the company can use available data to get as much as 300% differentials in healthcare costs. The company is relatively young—it began operations in January 2023— and Baran says its job would be “considerably harder” without the data transparency to which the industry is now subject.

Gottula says employers are now also accountable for analyzing data and ensuring their plans are good. Because insurance carriers must answer to their shareholders, employers shouldn’t assume that their network of providers is best suited to their needs or that providers are the highest-quality, lowest-cost available.

“Self-funded employers need to take responsibility to be educated on cost and quality data,” he says. “And don’t try it for yourself. There are vendors to help you. And brokers can serve this need, becoming a trusted advisor to their customers. The data is all here; let’s just do a good job with it.”

Ahmed says forward-thinking brokers can use plan data to compare claims for different plans in an area to see which are negotiating better rates with providers. Brokers can look at different procedures and create a “basket of CPT codes” that are high-volume or high-cost. A broker could then use that data to compare reimbursement benchmarks to find average prices for those services to help with provider negotiations. The information could also be used to compare plans from the same carrier to determine which might be the most cost-effective.

Steering Patients with Data

Using data well can make a difference in an organization’s bottom line. Gottula says it can provide insight into the cost of healthcare providers in a company’s network. If a company is covered by UnitedHealthcare in a large market like Dallas, it may be flush with providers. Brokers or employers can look at claims data from a range of doctors to see if they have the most cost-effective network.

“Businesses should not just take it as a foregone conclusion that the cost of healthcare is the cost of healthcare for their employee base,” Gottula says. “There is enough de-identified claims data available that they can find costs on a majority of the care that is delivered. An insurer should give data showing they offer a high-performance network, and the employer has to analyze that.”

Some vendors are marrying cost and quality data so they can “stitch through the technological patient journey,” Gottula says. For instance, if an employee has back pain, a business can find the cost for treatment to compare with like providers. And that business should be able to go a step beyond to first determine which providers use less expensive, conservative treatment, like physical therapy, instead of rushing to surgery. One can survey the data for patients’ healthcare experience to provide insight into which providers offer more preferable treatment pathways or reach better outcomes. Brokers, he says, can help educate employers so they can then talk with their employees about the best healthcare options.

“There is value not just to know the prices but to marry cost data with performance,” Gottula says. “You may pay a higher rate for the in-patient portion at Cleveland Clinic, for instance, but an organization might then have post-acute care that is demonstrably less and with better outcomes.”

Most employers aim to steer patients toward low-cost, high-quality providers. Data can help them make that shift—even small businesses that don’t have much data for comparison. Public data sets, like Medicare, can be used to provide cost information for providers. Using artificial intelligence, a vendor can pluck certain codes that a group might use, such as for cardiologists, to compare them to other groups in an area, Ahmed says.

When cost and quality data are combined, they can be used to steer patients toward preferred providers in their network. Once the best providers in an area are sussed out, carveouts can be built for high-cost services like back pain that encourage people to try physical therapy first. Or nurse navigators may recommend an ambulatory care center that might cost less than a health system. Some companies even contract directly with organizations for bundled services, like joint replacement.

Steering patients is the backbone of Self Fund Health’s business. The company helps employers design plans according to the cost of care that is being delivered. After analyzing claims costs, Self Fund Health aligns the insurance plan to guide participants to preferred doctors. One way to do this is by having zero-cost or low-cost deductibles when patients go to particular doctors.

“The employer will pay $500 to save someone from going to a place where the cost is $5,000 more,” Baran says.

Self Fund Health also uses primary care to help direct employees. Instead of starting people at a primary care doctor in a health system—who will then refer people to high-cost specialists within that system—the company uses employer-based clinics or works with independent primary care doctors. The referrals, then, can align with care that benefits the employer and not a large system, Baran says.

In addition, the company identifies ways, in real time, to steer patients through nurse navigators. If someone has a question about care, they can talk with a navigator who will recommend preferred providers for treatment.

“The only way employers can fight back is to stop thinking about buying insurance and be better buyers of healthcare,” according to Baran. “Every year, they go through renewal and look at ways to reduce costs like switching networks or changing out-of-pocket costs and deductibles. But if you want to spend less on healthcare, you have to go where you get an MRI for $500 instead of $5,000. You just have to spend less on healthcare.”

Changing the way employees use healthcare doesn’t happen overnight, Baran acknowledged. In year one, as many as 50% of a workforce will start using direct primary care providers, he says. This tends to grow as people become accustomed to the switch.

Another use for claims data is to understand if you have a good-quality network. An employer can ask a broker or other vendor for employer-specific claims and use AI to see what those claims would look like on a different network.

“They can use transparency data to reforecast claims costs,” says Ahmed Marmoush, CEO of Handl Health, an AI platform that analyzes publicly available healthcare data.

Healthcare data can also be used to see if there is a gap between what providers say they are charging and what the employer is paying out in claims. If this is the case, it may be time to renegotiate fees. The data can also be used for negotiating reimbursements for out-of-network providers.

“It can be used for democratizing some of that repricing,” Ahmed says. “Instead of just throwing out a rate for out-of-network physicians to take, employers can reference prices with existing payer agreements or use the claims information to negotiate when they receive out-of-network claims.”

Beyond traditional healthcare services, use of data can cross into the workers compensation realm. Claims data analytics can be applied to move “dubious” health claims over to workers compensation. Employers can also have workers compensation claims analyzed to help predict future claims—and help people lower their risk for conditions like repetitive strain injuries. Workers compensation claims can be monitored for potential workplace injuries. For instance, falls are among the most common types of workers compensation claims for teachers—often due to crowded classrooms, hallways, or spilled food and drinks on the floors. An organization then can manage risks that contribute to injury.

“Looking at data over time allows us to predict the bigger picture,” Schulman says. “We can create plans that address problems early on and use interventions that create conditions for better outcomes. This is different than looking at renewal and knocking down our costs a couple of points. This is on a large scale, and we are in the nascent stages of what’s possible.”

This kind of data collection and analysis was historically available only through large consulting firms or for major companies with the resources to do it in-house. In theory, price transparency is levelling the playing field for smaller employers. Consulting work that used to be done for large, individual employers is becoming scaled for smaller businesses as well. And employers should expect their brokers to be able to come to them with solutions in this space, Marmoush says.

Small amounts of data may not be as useful as what an employer would get with 5,000 employees. But that can be counteracted by including several years of claims or by joining with other small employers in the same industry and geographic area through data-sharing agreements that aggregate claims. Employers may turn to employer forums to assist with those projects.

Fulfilling Fiduciary Duties

The CAA aims to increase transparency so health plan beneficiaries can choose the best prices for needed services and payers and so employers can be better positioned to negotiate prices with insurance carriers and providers. Because they now have more data, health plans and self-insured employers have also been given more responsibility. Insurers have to make cost information available to beneficiaries. And self-funded employers, like Johnson & Johnson, now have fiduciary responsibility to use that information and get the most bang out of plan beneficiaries’ bucks.

Many self-insured employers are checking compliance off their to-do list by signing contracts saying their third-party administrators or insurance carriers are complying with the CAA on their behalf. But that doesn’t mean employers can wash their hands of all responsibility for the plan.

 “Lawsuits are showing who is ultimately liable for compliance [with the CAA],” Marmoush says. “They can have a contract, but pleading ignorance is no longer sufficient. They can’t say they didn’t know what they needed to do, because it is now their duty per the legislation to know. The buck stops there.”

Marmoush says it is every employer’s “responsible act” to confirm that a carrier’s machine-readable files are up to date, accurate, and complete. He says he consistently sees empty data sets or ones that are terrible quality among smaller groups and TPAs that use regional plans. This is an issue, because, if checked, they would be out of compliance, he says.

Employers should also ensure their plans have no gag clauses and that they can access claims data in order to evaluate the cost-effectiveness of their plans. If an employer’s network is paying 20% more for care to its providers than the competition, it can be sued for not taking fiduciary responsibility for the plan. Employers also must ensure their plan design is adequate. If a covered parent has to pay out-of-network prices for behavioral health services for a child because of an insufficient network, the employer could be held liable.

“The biggest issue employers face is that the legislation addresses some, but not all, parts of what is needed for employers to make these decisions,” Marmoush says. “They have access to the data but not to some contracts which providers can use for cost shifting. For instance, fees can’t be hidden in claims, but employers don’t have access to that data, so they have to ask to ensure all fees are exposed and there are none hidden.”

To help with compliance, he recommends companies have a fiduciary board with responsibilities and accountability beyond the company’s human resources department. This group would analyze the adequacy of the plan; if the employer is spending the right amount for the plan; and whether it is being judicious with beneficiaries’ dollars. The group could include legal advisors, the company’s CFO or chief compliance officer, and potentially a broker.

“Brokers should know what it means to be a good fiduciary,” Marmoush says. “They should be empowered and should know what type of questions to ask and how to uphold the employer’s responsibility.”

Employers can write into their vendor contracts deliverables like monthly reports showing network adequacy, evidence that the vendor is meeting benchmarks, and attestations that the plans have no hidden fees or gag clauses.

There was a lot of excitement in the industry when these transparency rules were passed, Ahmed says. There was interest in exploring the multitude of ways the data could be used. Then came frustration when analytics teams were determining if it could even be used at all.

“Last year people didn’t really trust the data, but there were lots of early movers in the space, and they are seeing results,” he says. “Excitement is re-emerging, and I hope this momentum stays strong.”

Increased interest in ways to use the data tends to lead to greater enforcement of the regulations, Ahmed says. If insurers continue to flesh out the data, it can only become more valuable to employers and plan beneficiaries. This July, changes to the regulations are supposed to make the data more manageable and clearer. More pharmaceutical data could be on the horizon along with disclosures of out-of-network pricing, which are currently nearly nonexistent.

“I hope the data helps us understand the cost of healthcare more clearly,” Ahmed says. “It is a universally recognized issue that healthcare costs are high, and the employer bears the brunt. If enforcement stepped up in areas where there are gaps in knowledge, it would be very powerful.”