The Patient Is Beside the Point
Two stories caught my eye this week, and together they captured something that has been bothering me about American healthcare.
The first was a newsletter by former insurance executive Wendell Potter about the enormous revenues of companies like UnitedHealth, CVS, Cigna, and McKesson. These are businesses I have written about extensively, with major roles in insurance, pharmacy benefits, and drug distribution. He shares the below diagram along with this quote, “If you rank the ten largest health care companies in America by revenue, you will not find a single hospital system or drugmaker. Not Pfizer, not Eli Lilly, not the Mayo Clinic or HCA. Every company in the top ten is a middleman, and together they take in about $2.6 trillion a year.” That is exactly what struck me — how much of our healthcare economy is organized around managing the money and transactions surrounding care, without actually providing any.
The second story, in the New York Times, described hospitals and insurers using AI to fight over medical bills. It cited a Blue Cross Blue Shield Association analysis estimating that increasingly complex medical coding by AI contributed to nearly $1 billion in additional healthcare costs over two years, without corresponding changes in the treatment measures it examined.
Hospitals were using AI to collect more money from claims. Insurers were using sophisticated AI tools to scrutinize claims and fight back. Software companies were selling their services into the dispute and keeping a piece of the extra revenue.
Apparently, what American healthcare needed was another company taking a cut.
The connection felt obvious to me. We have built enormous businesses around the process of paying for healthcare. Now, presented with technology that could change medicine for the better, we are instead investing in making that process even more elaborate. The care itself can apparently wait.
From behind the pharmacy counter at C.O. Bigelow, it can sometimes feel like everyone has figured out how to get rich off healthcare except the people actually giving and receiving it.
There always seems to be money for another intermediary. Another platform. Another service to help navigate the requirements imposed by the last one.
Meanwhile, I can help a patient understand their medication, work through a side effect, or figure out how to take their prescriptions safely, and that conversation usually comes with no separate payment to my pharmacy.
I want patients to have that access. I also want the people providing it to be paid for their expertise.
So my blood boils when I read about a hospital’s AI software finding additional billing opportunities in a medical record while so much of the work happening in front of me remains financially invisible.
The system is remarkably inventive about monetizing the process. Recognizing the value of the care seems to require considerably more discussion.
Medical coding translates your diagnoses and care into the language used for payment. In certain hospital payment arrangements, documenting an additional qualifying condition can move a stay into a higher-paying category. AI helps search the record for conditions that might otherwise go undocumented.
There are legitimate reasons to do this. I do believe a hospital should get paid appropriately for treating a complicated patient. And to be fair, BCBS’s findings do not establish that every additional diagnosis is improper or that AI caused the entire spending increase. Additionally, STAT’s reporting on an earlier analysis raised questions about the methods and whether some conditions had previously been underdocumented.
On the other hand, an insurance company complaining that it paid out more than it wanted to obviously does not have my sympathy.
But more complete documentation can increase payment for care already delivered without changing what happened to the patient. That may fix an underpayment. It does not, by itself, make anyone healthier.
The Times described McLaren Health Care, whose chief financial officer said an AI coding tool increased revenue by $1 million a month. Its vendor, SmarterDx, receives a share of the additional revenue it generates. The software company gets a cut of what it finds.
Admittedly, I would buy a tool like that for my pharmacy.
Of course I would. If it helped us collect money we were legitimately owed for care we legitimately provided, I would happily pay for it.
My guilt and willingness to buy it is part of what makes this so frustrating.
Getting paid has become so sufficiently difficult that hiring another company to help is a sensible business decision. Insurers buy tools to scrutinize claims. Providers buy tools to submit and appeal them. Each side has a reasonable explanation for its investment.
None of those explanations requires the patient to be better off.
Caroline Pearson of the Peterson Health Technology Institute told the Times that automation allows the parties to go additional rounds over payment because doing so is cheap. Making each argument cheaper can encourage more arguing. A process can become faster at every individual step while becoming more complicated overall.
This feels like a frustratingly small ambition for technology with so much potential.
Don’t get it twisted, I am excited about AI in medicine. I especially love the possibility of revisiting old clinical trial data and discovering benefits we may have missed: patients who responded particularly well, or clues about another use for an existing drug. Those findings would need to be tested, but that is a future worth pursuing.
I also want AI that reduces paperwork, helps overturn an inappropriate denial, and gives clinicians more time with patients.
What I really want is a system that reliably pays people for providing good care, including the work that never produces another billing opportunity.
A software’s intelligence cannot compensate for a system that rewards the wrong results.
A hospital can point to the money it recovered. An insurer can point to the payments it prevented. A software company can point to another contract. Each can declare victory without showing that a patient got an appointment sooner, filled a prescription more easily, or spent less time fighting for care.
The final focus should be the person seeking care and whether any of this has made their experience better.
A bigger payment is evidence that the billing software worked. It is not evidence that healthcare got better. The companies selling this as progress should have to show the difference.
That is what connects these two stories. We have built a healthcare economy where improving the business of getting paid can be enormously valuable even when it does nothing to improve the patient’s health.
Let’s make sure we don’t forget that the patient is the point.
Giddy up!
Alec Wade Ginsberg, PharmD, RPh
4th-Gen Pharmacist | Owner & COO, C.O. Bigelow
Founder, Drugstore Cowboy
