In a settlement announced this week with the New York State Attorney General, UnitedHealth Group has agreed to shut down two databases maintained by its Ingenix subsidiary. In addition, UnitedHealth will pay $50 million to re-create the databases under the aegis of a non-profit organization to be established for the purpose.
The two Ingenix databases involved are Prevailing Health Charges System and the Medical Data Research database. The two products track "usual, customary and reasonable" physicians' fees. Most insurers use Ingenix's data, which is not available to the public or doctors, to calculate patients' out-of-pocket costs when they seek out-of-network care. Physician groups have complained bitterly about under-reimbursement driven by these databases, and consumers may have been under-reimbursed as well. This settlement does not address the claim of under-reimbursement.
In addition to the under-reimbursement claim, the New York State Attorney General also attacked the Ingenix data products for "lacking transparency." It does seem remarkable that a commercial, subscription-based data product would be obligated at all to be transparent. The New York Attorney General also charged a conflict of interest in that Ingenix is a subsidiary of a health insurer that also makes use of the Ingenix data for calculating reimbursements. Again, what's implicit in this charge is that these databases are so important that they've become bigger than the organization collecting them and become something of a public trust.
This settlement brings to mind the furor that erupted in 2006 when the First DataBank unit of Hearst Corporation was engaged in what appeared to be some very sloppy updating of its wholesale average price database for pharmaceuticals. Instead of surveying the industry to develop a true price average, it was instead gathering all its data from a single drug wholesaler. As a private, subscription data product, this would normally be just a huge embarrassment. In the case of First DataBank, however, this database was driving pharmacy reimbursement rates nationwide and reportedly led to inflated reimbursements to the tune of many billions of dollars. Here again, we have an example of a private industry database with outsized influence. While Hearst can't be claimed to have had any conflict of interest, this clearly seems to be a case were increased transparency regarding data collection practices might have prevented the problem in the first place.
Should private sector databases that are used to drive payment systems, particularly where taxpayer dollars are involved, seek to meet a higher standard of transparency? On reflection, I think the answer is "yes." Data publishers whose products fuel mission-critical applications shouldn't need to hide their work. If your work is sloppy, you have a perpetual litigation threat hanging over your head, as these two cases well illustrate. If your work is first-rate, you have a selling advantage. As to the proprietary aspects of building a database, I'm not (necessarily) suggesting that you open your algorithms to the world, since their quality can be measured based on results. Rather, I am suggesting that your data inputs and your process for creating gold from dross might benefit from some sunshine. After all, any data publishing veteran knows well that most of our value is wrapped up not in secret methodologies but rather that we sat down and did the messy work nobody else wanted to do. We aggregate, we scrub, we normalize, we purify. It's not rocket science, and there are few secrets, just skilled practitioners. In our business, transparency builds trust, and trust builds increased utilization, meaning greater revenues. Suddenly, the case for transparency seems perfectly clear.