Last week, there was an important story out of California. Insurance regulators fined the two largest insurers for having craptastic directories. Additionally, the regulators, as empowered by California law, are holding individuals harmless by requiring the insurers to attribute cost sharing as if the providers were in-network as the individuals had a rational and reasonable belief that the providers were in network.
The state’s Department of Managed Health Care levied fines of $350,000 against Blue Shield of California and $250,000 for Anthem Blue Cross.
At issue were the companies’ error-riddled provider directories that frustrated many consumers statewide as they tried to find doctors during the rollout of the Affordable Care Act in 2014. As a result, some patients incurred big unforeseen medical bills because they unwittingly went out of network for care….
The state said Blue Shield has already reimbursed more than $38 million to enrollees who incurred out-of-network costs. Officials said they didn’t have a reimbursement figure for Anthem yet.
I had a very enjoyable back and forth with Claire McAndrews of Families USA concerning directory accuracy and she pointed to an interesting secret shopping study on the Maryland market:
— Claire McAndrew (@claire_mcandrew) November 5, 2015
Most of the insurers (page 9) had a large number of psychiatrists in the directory but a large number of false positives, non-psych specialities, and even fewer had psychiatrists who could make an appointment for a new patient in a reasonable time frame. However one insurer showed a much smaller network of psychiatrists but that very small network had a very high percentage of useful information.
This study raises an interesting set of questions below the fold:
I’ve been told that I have an unusual introduction to my pregame when I am working with referees that I’ve not worked with before. I’ll introduce myself, suggest that we walk the field and then the opening line of my actual pre-game is the following:
“Our first priority today is player safety and our second is fairness; when we make errors, I want our errors to lean in those two directions. I want this game to be safe, and I want this game to be fair”
An experienced but new to me assistant referee said that he was chewing on that line for the rest of the game as he had not thought about the game in that way before. An assessor commented that he thought the thought process was very good for where I am now but if I was ten years younger and 15% faster, it would have limited my advancement.
I had made a decision several years ago that I would limit some of the risk in the game in order to favor safety and fairness. This limits my ability to be in a game where the entire game is played at the cliff edge of beauty and chaos, but it also limits the long term injuries and the probability of a game going to shit. This was a deliberate decision as to where I want my errors to lie.
Provider data is messy data. In my experience, the only reliable provider data is data that is directly tied to money. That means the tax id number is reliable, that means the NPI, Medicare and Medicaid numbers are reliable, and that means the address to send the payment to is reliable. And if it is not reliable, the providers are very good about changing that data without prompting as a missing check or EFT fund transfer will get their attention quickly. Everything else is far fuzzier. Addresses are often not in US Postal Service standards, locations that have not been open for years will not be reported as closed, providers who moved to Arizona to work on their golf game or have joined the choir invisible are not reported as terminations. Panel status (the ability to take new patients) and office hours are also often fuzzy. My PCP has three sets of hours for Thursday reported in the three largest commercial network directories (Mayhew Insurance is accurate because I took a photo of the new hours and sent it to our provider folks who I used to work with) .
Messy data means significant assumptions need to be made in how it is presented.
I think about this in the context of directories as most insurers are making a decision (given my experience, an unconscious decision) to have errors that cast the illusion of a much broader network. Right now the cost of errors going in that direction are mostly borne by members and not the insurance company unless the state has aggressive regulators with strong tools to do their job. California has a hold harmless provision that makes the cost of a bad directory be borne by the insurer.
The Maryland study does show an interesting possibility of a counter-example. There a decision was made to show an accurate network and potentially under-report the network. Errors would be errors of omitting participating providers instead of errors of including non-participating providers. Members would be very confident that if they picked up the phone, they would be talking to a provider that could help them. In the other cases, it was a crap shoot if the provider that they reached was actually able to help them. There may have been providers in the network but out ofthe directory because those providers elected to sign up for the network to maintain a relationship with one or two patients but they did not want any more calls for treatment slots that did not exist or they could have an idiosyncratic group contract that brought them into the network but for practical reasons, they were effectively unable to treat. One example is a provider who is technically in network as he is a guest lecturer and very odd case surgeon who flies into Shelbyville two or three times a year for weird surgeries but is otherwise practicing 97% of the time in Los Angeles.
The health insurance industry seldom makes a conscious decision as to what type of network directory errors they want and it hurts.