As the healthcare industry continues to await the Medicare Access and CHIP Reauthorization Act (MACRA) final rule, expected to drop any day now, discussion around some of the proposed rule’s core health IT elements, and how they will affect participating physicians, have begun to heat up.
Last week, Healthcare Informatics published an interview with notable health IT figure Farzad Mostashari, M.D., who called out both electronic health record (EHR) vendors and hospitals for actively engaging in data blocking. (That interview can be read in full here, and a blog post written by Healthcare Informatics Managing Editor Rajiv Leventhal that took a deeper look into Mostashari’s comments on data blocking, can be read in full here).
What happened next was a Twitter discussion about the MACRA regulations, and about data blocking, led mostly by Niam Yaraghi, Ph.D., a fellow in the Brookings Institution's Center for Technology Innovation, with extensive background in the economics of health information technologies. Yaraghi disagreed with Mostashari’s perspective on data blocking, noting that it’s the Office of the National Coordinator for Health Information Technology (ONC) regulations that allow vendors to “game” the system. He unleased several tweets about it, including this one:
Yaraghi then took to the Health Affairs blog, penning a piece about why the MACRA proposed rule “creates more problems than it solves.” Following that, Healthcare Informatics caught up with Yaraghi to talk further about his perspectives on all of the above—MACRA, data blocking, EHR vendors’ moral and business obligations, and more. Below are excerpts of that interview.
You have been pretty outspoken about MACRA. Can you summarize your top gripes with the proposed rule?
My major criticism is with the MIPS [Merit-Based Incentive Payment System] part, which in the beginning is going to affect most physicians since very few will be in the advanced alternative payment model [APM] track. I believe it’s an open invitation to cheating. The MIPS composite score in the first year consists of different domains; three of those domains constitute 90 percent of the weight and are self-reported. If you ask a physician to self-report without having a mechanism to detect and deter cheating, then they will cheat. People don’t like to think of doctors being concerned about money or financial matters, but they are. That’s the truth. If you create a system that is so heavily based on self-reporting, and those self-reported measures could determine between a 4 percent bonus and a 4 percent penalty at a minimum—which is an 8 percent difference in Medicare payments—then people will cheat.
This is not a new idea. Medicare has been running something similar among nursing homes with a five-star rating system based on two self-reported measures of quality and staffing and one measure of on-site health inspections conducted by auditors. So the health inspections are done by the Centers for Medicare & Medicaid Services (CMS), but the other two measures are self-reported. There is already anecdotal evidence, as reported by the New York Times, about five-star nursing homes that tend to be very low quality, although those star ratings don’t reflect that. It’s primary because they cheat on self-reported measures. And cheating has a very negative connotation, but the data just looks very suspicious.
Even if people wanted to report the true measures [for MIPS], it’s really burdensome. What will happen is that the smaller and solo physician practices who do not have capability to report will have to stop seeing Medicare patients—and I see some people doing that—or they will have to join a larger practice that can support the overhead cost of all these reporting requirements.
The other interesting part to me is that one of the main criticisms about the Meaningful Use program was the fact that the measures were the same for everyone. So there was no flexibility, and people [complained] about that. For MIPS, they wanted to address that and provide more flexibility for doctors to report. For the quality component, for example, a family doctor has to report to six measures. But they can choose those six measures out of 38 that are available. In reality, there are more than 2.7 million possible combinations if you are choosing six out of 38. So at the end of the day, when Medicare receives these reports, they won’t be able to meaningfully compare one family doctor to another. If I choose six measures and you choose another six, comparing us together becomes meaningless. It’s like comparing your score in mathematics to mine in physics or comparing two SAT tests with different questions. Also, medical care is a much more complex process than just six measures. Being good in six measures doesn’t mean you are good in others as well.
Another problem is with the MIPS reporting feedback loops. [CMS] has proposed annual feedbacks at the middle of reporting periods, starting July 2017. I am suspicious about them even sticking to these dates. It’s like operating with your closed eyes—you don’t know how well you’re doing until at least halfway through the reporting period. Physicians are left with little time to see how they’re doing and change without real-time feedback. Also, the payment adjustments are two years after the reporting period, so the impact of that penalty or reward decreases with this distance.
Another thing that’s troubling is the field audits of EHR systems that allow ONC-certified people to go and check the interoperability [of a system] without patient consent, but using actual patient data. I couldn’t believe that this was part of the proposed law. Are you really going to send a team of people to a physician’s office and say, “Let me look at your computer to see if you can exchange data?” It’s very intrusive and it undermines patient privacy. There are ways to test interoperability without violating patient privacy.
In the piece you wrote, you talk about rewarding reporting capabilities instead of medical excellence. What could be done to better reward medical excellence?
I will be honest, I don’t have a clear answer. I don’t know how anyone could measure quality in the healthcare domain. We created this problem ourselves. Defining quality in any other area, like restaurants, in a structure in which the government will pay for it, is really difficult. Some obvious things you can measure in a restaurant, like getting food poisoning and if the staff treats you well, but when you want to pay for the food through the government based on those quality measures, things become complicated.
You tweeted that you don’t put the blame on EHR vendors gaming the system to meet CEHRT requirements. But would you agree that doing so could impact the quality of patient care in the end?
Of course it could. The system is designed in such a way that although the lives of people are at stake, the reality of the cruel world is that people don’t care. The only thing that speaks is the dollar. I am not defending the vendors; I’m [mad] at them. But I understand it. If what Farzad said about ‘hardcoding’ the system to past the CEHRT requirements is true, then somewhere along the line there should be a clause that says in the future you need to remain compliant. If not, there will be legal actions against you.
The other thought here is that data blocking is very difficult to define and measure, similar to racism. Sometimes there are obvious instances of hate crimes, such as if someone beat me up in the street for being Middle Eastern. But say I couldn’t get a job; that might be because of racism, but it might also be because I wasn’t qualified.
Data blocking is this magical enemy that ONC doesn’t have a clear definition of. How do you know someone is data blocking? And even if you do have clear definition of it, you can point to who is doing it, and there are penalties for it, this is not the reason for the lack of interoperability. There might be some instances where people are actively blocking the data, but it’s not the primary reason for lack of successful data exchange. The lack of business incentives for people to share data is the most important reason for not seeing information exchange.
What would you like to see changed in the forthcoming MACRA final rule, given that it stays pretty close to what was proposed?
I’d like to see a more solid and unified set of measures rather than allowing people to choose. Also, coming up with ways to independently measure those quality measures without interfering with physicians. How can we reduce the burden of data reporting and find ways to collect data so we can ensure the quality of data and at the same time foster interoperability? I’m not the technical guy, but if you look at how many patients receive a pneumonia vaccine as measure for a primary care physician, for example, it should be possible for EHR systems to report this to a data center at CMS. I don’t see why this is so difficult.
