Pushed Out of Network for Behavioral Healthcare

Why did you undertake this report?

HARBIN: We wanted to follow up on a report the Bowman Family Foundation commissioned from Milliman that was issued in 2019. It covered similar measurements of almost 30 million commercially insured lives from 2013 to 2017. It showed significant disparities in access between behavioral health and medical/surgical care, and we approached RTI to provide additional data from 2019 to 2021. This report was from a large commercial database with over 22 million people.

The second purpose was to see if there had been any improvement—we found there [was] hardly any—and to broaden the types of measures and analyses. We wanted to use different ways to measure provider reimbursement, not just looking at the averages but at the higher ends of reimbursement. And we wanted to compare subspecialty providers on both the medical and mental health side. We also looked at tele-mental health versus medical telehealth and in-person care.

You found that little has changed in recent years when it comes to the availability of in-network mental health providers in consumer plans.

MARK: I’ve been researching behavioral health policy issues for over 30 years and doing research on the parity issue for most of that time. I did a similar analysis over a decade ago, so it was very interesting to update and expand on those prior analyses and disappointing to see that we had made little progress on the issue of network adequacy. The mental health parity law [The Paul Wellstone and Pete Domenici Mental Health Parity and Addiction Equity Act of 2008] did reduce very obvious disparities like limiting psychotherapy sessions or inpatient stays or having higher co-insurance for mental health visits. But these less transparent disparities that have become really important for people being able to access their insurance benefits have not improved.

If the mental health parity law was supposed to improve access, how are provider networks still insufficient?

HARBIN: I think there has been improvement since the law passed on quantitative limits, in part because those are transparent: it’s clear insurers are violating the law if they have a 30-day limit for inpatient mental health, which used to be standard. But medical management is more complicated to measure. You can reduce the benefit you are paying by using stricter prior authorization or review [for treatment] or if you pay low reimbursement levels or you exclude certain services. If you don’t have a robust network, you reduce access, and you have also reduced your costs.

You found dramatic differences in in-network provider access between mental health treatment versus medical and surgical care. Why should this matter to employers?

MARK: Let’s say you hire a broker to help you find insurance for your employees, and you think you’re giving them mental health and substance use disorder treatment. Then your employees start saying they can’t find anyone who will treat their child at a residential facility or can’t find a psychiatrist and couldn’t get care. Or they did get care but had to pay thousands of dollars out of pocket. As an employer, you’re not really providing the benefit that you thought you were or that your employees thought they were getting.

HARBIN: This leaves the burden on the consumer, the patient who has got mental health or substance use problems. When they try to access insurance, the financial burden on them is significantly higher than it is for a medical/surgical patient getting care. It’s a significant problem, particularly if you look at inpatient care. We found that almost 30% of behavioral health patients had to go out of network for residential treatment. The bills for that one episode might be $20,000 to $40,000, depending on how long you stay. What person could possibly pay 50% of that cost? Then, out-of-network use on the medical side for inpatient care is like 3%.

The results we found underestimate the problem. This study included HMOs and PPOs, but some HMOs don’t have any out-of-network benefits. So, when you’re measuring claims, there are people that are going out of network and paying 100% out of pocket, but it wouldn’t show up in this study.

Has network access not changed since parity mainly because it’s more difficult to measure than quantitative things like days allowed for inpatient care?

MARK: It’s harder to measure, but it didn’t take a supercomputer for us to do this. There’s no reason the state can’t go to a plan and get this information. We think it’s important to look at the outcomes. A plan can say, “We have 50 psychiatrists in our network, and they’re all within 10 miles of people.” But we know when we do secret shopper studies people can’t access all of those providers. Instead of asking what a plan’s network looks like, we’re seeing what people are actually experiencing. Even though plans might be claiming their networks are adequate, they’re not, because people are going out of network and they’re paying out of pocket. What we hope is that this report will help to solve that problem by giving a path to the transparency that regulators haven’t had in the past to enforce the laws.

You were thinking about how regulators can better understand network access with this study, but can employers use this information to gauge their networks, too?

HARBIN: We have seen a few employers that do require this [network access] data and act on it, and they can make a big change in access. But they have to ask for it and specify the methodology, or they won’t necessarily get the full picture. Employers have a unique role to play in this because self-insured employers are liable for the behavior of their third-party administrator. They have a lot of power, and I think part of the value of this report is you can translate this into specific quantitative templates.

MARK: We tried to make it very clear in this report. You can hand the methods section to the broker and ask for your percentage [of patients who go] out of network and reimbursement methodology. You don’t just want to see a book that shows all of the psychiatrists that participate in your network; you need to see how many of your employees and their families are going out of network.

The report shows that plan participants were even going out of network for telehealth services more on the behavioral health side. Were you surprised to find that after the increase in remote mental healthcare during the pandemic?

MARK: We hoped that the disparity in in-network use would have been reduced when we expanded telehealth, but we don’t see that. We see huge disparities. Insurers just don’t have large enough networks, so telehealth is not a panacea.

HARBIN: Tele-behavioral health has been an important addition for access, but we were shocked. I know a lot of plans were trying to lean over backwards during COVID to make more behavioral telehealth in-network. But there were actually worse disparities compared to in-person care in 2021.

You also contend that the issue can’t be attributed to a shortage of mental health providers but to a lack of in-network providers.

MARK: The federal HRSA [Health Resources and Services Administration] data is very telling about that. HRSA has found there are more provider shortage areas for primary care physicians than for mental health clinicians. Yet people leave the network only 2% of the time to see a primary care physician and about 16% of the time for psychiatrist visits. If it was really driven by a shortage area, you’d see people going out of network for primary care, but we don’t. We really think that issue is overplayed in terms of explaining the difference. One of the takeaways for me was that plans have done a great job with medical/surgical, even in specialty care; there is almost no out-of-network care. Plans know how to do this. They’re just not doing it for mental health.

Do you think one of the main components of getting providers into a network is higher reimbursement?

MARK: We found, on average, that the medical/surgical side was reimbursed 22% higher than behavioral health clinicians. You can slice and dice the data in a lot of different ways, but one comparison that we thought was particularly striking was the reimbursement for common codes. Physician assistants and psychiatrists use E&M [evaluation and management] codes when they do a brief evaluation and provide a psychiatric medication. We found for that code, on average, physician assistants were paid more than psychiatrists. So a midlevel-trained medical professional is getting paid more than an M.D.-trained specialist.

HARBIN: This study also looked at the higher levels of reimbursement, which hasn’t been reported before. What it showed is that, at the 75th percentile of reimbursements, medical/surgical is paid 48% more than behavioral health providers and, at the 90th percentile of reimbursement, they are paid 70% more.

What do insurers need to do to get more behavioral health providers on their panel of practitioners?

MARK: Fast-tracking credentialing and making prior authorizations less of a hassle. Right now, behavioral health providers have to fax and call to get all of these things approved. Plans have to recognize that more mental health and substance use providers are smaller groups. The plans need to use different techniques to get them. Some plans say that big medical systems have more bargaining power, so they pay them more. But that’s not an excuse for parity noncompliance. And there is so much demand for mental health providers, they know they can be busy without being on insurance panels. There are demand and market forces that mental health professionals have now that they never used to. We did a survey of residency programs [for behavioral health professionals] and asked if any insurers had tried to recruit them during training. Almost none of them had been approached by any major plans to get them to join their network. These plans usually do aggressive outreach to providers, and not a lot of that is going on in the behavioral health space.