Talking to the media last summer, Victor Echevarria sounded confident about Remedy, his startup helping people correct mistakes and get refunds on their medical bills. The company had been testing its service–a mixture of algorithmic software and human billing expertise–for months, and the results seemed both impressive and promising. Beta customers had saved thousands of dollars after Echevarria’s team had questioned bills on their behalf.
Jaime Chiang, a San Francisco branding executive who was badly injured after being pushed down some stairs in a Whole Foods store, had escaped the lion’s share of her $22,000 in medical costs, as you can see in the video below:
Another man had reversed a charge for a pregnancy test (bills for male pregnancy tests are more common than you might think, Echevarria says). And Echevarria, a former executive at TaskRabbit, had even seen Remedy work for his own family, beating back bills when his son had a febrile seizure and had to be rushed to hospital. Remedy worked in finding mistakes in the bill, and, what’s more, the stressful episode gave Echevarria the ideal sort of origin story to tell in interviews.
Eight months later, Remedy was closed. Echevarria and his cofounders John Schulte and Marija Ringwelski shut the business down, overnight, citing a lack of cooperation from hospitals, doctors, and insurers in accessing billing records. Eighteen months of struggling with hospital systems and clinics had worn everyone down. The money had run out.
Echevarria wrote in a goodbye post: “Until the industry begins respecting the rights people have to their own information, it will remain difficult for individuals and their agents to be vigilant against baseless medical bill overcharges.”
Remedy is just one in a pile of failed startups around the healthcare industry, but it’s especially notable for what it says about access to our healthcare billing data. Remedy failed because it couldn’t get information efficiently enough. There was no pipe of data it could tap into. It was stymied by endless paper forms and roadblocks from back office staff at the hospital systems and clinics it worked with. That it couldn’t get hold of the necessary information from the healthcare industry on our behalf raises the question: Can anyone ever corral this system to help patients?
Finding errors (and sometimes outright fraud)
Mistakes are more common in healthcare invoices than any other industry. With their impenetrable CPT and ICD-10 codes, medical bills are a reflection of our opaque, complex, and circular healthcare system, a system that speaks its own language, one only insiders can understand.
Estimates of how many medical bills contain errors go as high as 90%, from mismatched procedure codes to problems in the “balance billing” process–where insurers and health providers decide who is going to pay what. (Described as the “ugly beast” of medical insurance, it allows a health provider to collect the entire balance of what the insurer feels disinclined to pay.)
In an interview in September, Echevarria said the incident with his son–with its bill of $12,000 in all–had turned him into a billing hawk with a “personal vendetta.” He wanted to fight for all the people who are overcharged for their medical interactions. Remedy would “offer a layer of protection between you and the broken billing system,” he said.
By rooting out mistakes, Remedy was going to build a business from the industry’s fat margins, while helping people at the same time. The San Francisco startup received $1.9 million in funding, including a personal investment from Marc Benioff, the CEO of Salesforce. And it joined the Financial Solutions Lab, a New York incubator for fintech aimed at low- or medium-income consumers.
When users signed up for Remedy, they granted electronic access to their insurance bills. Remedy would then go to healthcare providers if they found a mistake and try and correct it. Remedy positioned itself like Mint, the automated financial manager that links to your bank account and keeps track of your bills and credit score. It charged 20% of what it saved customers when it spotted overcharges, up to a limit of $99.
Perhaps Remedy–and its investors–should have thought about the problems with this information pipeline before building a healthcare data analysis business. But it’s still notable how much Remedy was stymied from the start. When Echevarria’s team of billing subcontractors contacted the back offices of hospital and doctor groups, they were frequently met with slow-walking. Officials would require forms to be filled for each doctor, multiplying paperwork if patients saw several health professionals. They would lose request forms and ask Remedy to resubmit new ones. They would contact patients and tell them that working through Remedy was illegal (it wasn’t).
“The main problem wasn’t the analysis of the data,” Echevarria says in a recent interview. “It was accessing the data. Getting doctors and providers to cooperate with us as agents of the patients was just a mind-blowingly difficult process. They will cooperate if you pay them lots of money, but they’re not in the business of helping customers get their data.”
HIPAA, the 1996 Health Insurance Portability and Accountability Act, gives patients the right to access medical information on request. But it’s a double-edged sword. As well as smoothing data delivery, it’s meant to protect it from falling into the wrong hands. Echevarria says that gives healthcare operators a ready-made excuse not to cooperate with a startup like Remedy.
Under HIPAA, healthcare providers are allowed to charge administrative fees for data access, including fees for copying data, the media that information is presented in (like a CD or thumb drive), and postage. That all increased Remedy’s cost of doing business. “Every time we made a request, it was like [it] took three months and cost $100 worth of labor,” Echevarria says.
Remedy automated much of the bill analysis process, building machine learning and predictive tools. But it spent too much gathering data and putting it into a machine-readable format, Echevarria says. Each healthcare company would have its own data request forms. (Remedy couldn’t use a standard form across providers.) And the information would come back in a way that wasn’t friendly to automation–like as a PDF document. Remedy would have to strip out the data from the form before it even begin to do its work–a laborious process.
At the end of 2016, realizing it couldn’t make money from consumers, Remedy switched to a business-to-business model. It started offering its error-identifying software to insurers, hoping to correct mistakes in the information flows between providers and payers. Echevarria says that approach was gaining steam by this summer. Insurers have a strong incentive to see that claims are correct. They already pay for “payment integrity” services on large bills. And they seemed willing to fork up money for Remedy’s automated approach. Most importantly, they also have access to consumer healthcare data, allowing Remedy to do its job.
“We felt we had to get hold of the data,” Echevarria says. “Payers are in a good place to do that because they have longitudinal authority on every individual. They know every doctor. We were about to do deals. But we just ran out time. Eight months was not long enough to complete a full sales cycle.”
The limits of technology
Tech startups are adept are combining software and data to make our lives better. From books to taxicabs to consumer finance, Silicon Valley has made accessing goods and services vastly more convenient, and solved many social problems, all the while advancing the interests of individuals over incumbent operators. But the healthcare industry represents a particularly hard nut to crack even for brilliant computer engineers and brash CEOs. Companies like Remedy can arm us with powerful, easy-to-use apps. But ultimately they’re dependent on the industry playing ball, particularly in the release of consumer data.
A lot of startups fail because they’re fundamentally bad ideas that people don’t need. They have poor business models or over-expensive products. They have no reason for being. That’s not necessarily the case with Remedy, according to several well-placed people.
Ryan Falvey, managing director of the Financial Solutions Lab at the nonprofit Center for Financial Services Innovation, says he has no regrets about the incubator’s $250,000 grant. He sees Remedy closing as being more about the lack of cooperation from healthcare companies than it was about a poorly executed startup.
“This was a good idea,” he tells Fast Company. “This is a multibillion dollar industry, where there’s a lot of waste, and inflated expenses. I don’t think anyone who’s ever encountered healthcare in this country would come away thinking anything other that there are tremendous opportunities to improve the consumer experience and how we pay for it.”
Lucia Savage, who was chief privacy officer in the U.S. Department of Health and Human Services during the Obama Administration, also thinks the basic concept behind Remedy was sound. But it required a better data pipe–an API–if it were to automate the processing of bills and make money.
“I loved what Victor and the team at Remedy were doing. They were ahead of their time–the DeLorean of medical billing information. Unfortunately the underlying open API and related data standards don’t exist yet like they do for clinical data,” she says in an interview.
The healthcare industry is now adopting a new data-sharing standard called FHIR, or Fast Healthcare Interoperability Resources. FHIR allows developers to build apps on top of data from electronic healthcare records. That will mean patients will be able to manage prescriptions from their phones, for example, or for doctors to monitor test results electronically. But it’s only for clinical data; as yet, no equivalent standard exists for billing information of the type that might have allowed Remedy to do its job more easily and cheaply
Savage says we don’t necessarily need new laws to improve access to billing information. Healthcare providers and insurers could develop their own standards, much like we have industrial agreements in place for, say, internet browsers. But she says they would be costly to develop and implement and the industry is not incentivized to open itself up to scrutiny.
In theory, a billing information API could do more than reduce mistakes. It could also reduce underlying prices in the medical system, which of course are some of the most expensive in the world. U.S. healthcare is notorious not only for high out-of-pocket costs, but also for the unpredictability of those costs. An MRI scan that a family believes might be billed at $1,000 ends up costing $25,000. Because of a lack of price transparency, the same procedures can vary widely state to state, city to city, seemingly for no rhyme or reason.
Aneesh Chopra, chief technology officer in the Obama White House, argues that the adoption of FHIR could create a consumer groundswell for other types of data and apps, including the billing kind.
“I believe that the complexity of the system is such that our cultural and technology investments have not focused on the problem of making [medical billing] easier for patients,” he says in an interview. “It’s not been seen as part of the responsibility or the requirement of the institutions and that cultural shift is the most important to tackle.”
Despite Remedy’s problems, Chopra and Savage don’t blame healthcare providers for failing to collaborate. The back offices of hospitals and doctors are culturally unaligned with modern expectations of consumers for access to information, they say. Mostly they focus on getting paid by insurance companies.
Echevarria says if he was starting Remedy again, he would build a product for smaller payers, then, as revenues started to build up, he would create a consumer service on top of that. But he’d prefer to have an open data standard from the start.
“Helping insurance companies was a great monetization model, and it got us access to the data. But it was really a workaround,” he says. “We really have to wait until the pipes are laid–when there’s a Facebook Connect or Yodlee for your medical records. At the moment, we’re a long way from that because there are a lot of powerful companies standing in the way.”