As the Medical Director of the Kansas Business Group on Health I’m sometimes asked to weigh in on topics that might affect employers or employees. This is a reprint of a blog post from KBGH:
I’ve heard a few times over the past couple weeks that hospitals are padding their case counts of COVID-19 patients in order to increase revenue. This is transparently, obviously false, as we’ll get into later. But before we wade into that, let’s take this chance for a quick review of how hospitals and doctors get paid for the care of patients.
The history and process for how physicians are paid
Once upon a time, billing for medical care was very informal. Hospitals and doctors largely set individual, almost artisanal, rates for each patient according to a “sliding scale” of what the patient was expected to be able to pay. Poor patients paid less, and wealthier patients paid more.
Once medical insurance became common, insurance companies, including Medicare, attempted to hold physicians and hospitals to the standard of “customary, prevailing, and reasonable charges.” Unsurprisingly, this loose standard led to steadily inflated billing, so much so that the passage of Medicare is arguably what vaulted physicians from middle-class professionals into the upper reaches of national income. As early as 1970, congressional testimony referred to federal insurance as the “Goose that laid the golden egg” for physicians and hospitals.
Through a series of reforms in the 1970s, ‘80s, and ‘90s, billing for medical care became much more standardized (and in part led to the administrative bloat that is now the number one source of waste in American health care). Nearly every diagnostic or therapeutic procedure performed by a medical professional is now captured by a “Current Procedural Terminology” (CPT) code. For example, your dermatologist codes a “2029F” for a skin exam. A cardiothoracic surgeon codes a “33945” for a heart transplant. A routine, but fairly comprehensive new visit to a primary care doctor is coded a “99204.” All these codes are reimbursed according to the complexity of the task, taking into account the amount of time a procedure is expected to take, the amount of resources like syringes and protective equipment expected to be consumed, and the skill or level of training required to provide the service.
Hospitals themselves bill not according to CPT codes, but rather according to Diagnosis related groups (DRGs), which were introduced in the 1980s. DRGs are meant to make sure that reimbursement account for the severity and mix of the type of patients the hospital treats, and thus the resources that the hospital needs to treat those patients. For example, someone who presented with fever, cough, and a density on their chest x-ray, and who tested positive for COVID-19, would be coded a discharge diagnosis of “J12.81” for “pneumonia due to SARS-related coronavirus.” If that same patient needed ventilator support during her hospitalization, though, she would be coded “J96.01” for “Acute respiratory failure with hypoxia,” which pays in the ballpark of $54,000 (about three times as much as a COVID-related diagnosis). The additional payment is meant to pay for the increased duration of the visit and the increased intensity of treatment, since patients on ventilators are typically cared for by a single, specialized nurse, a respiratory therapist, a pulmonary physician, and others.
Our healthcare system has some inherent issues
The purpose of this post is not to defend current medical coding and billing. Our system is bizarre by almost any developed country’s standard. Take the way payment is determined for those CPT codes. The American Medical Association owns the Relative Value Scale Update Committee, or “RUC,” which is tasked with updating physician payment for those roughly 4,000 CPT codes. The RUC is powerful. It ultimately guides about 70% of all physician payment in the United States. Most of its 31 members are assigned by professional societies like the American College of Radiology and the American Society of Plastic Surgeons. Therefore, primary care doctors, the most cost-effective and crucial part of the health care workforce, make up only a tiny fraction of the committee. So the natural momentum of the committee is to steadily increase the payment for specialty care, while keeping reimbursement for routine care relatively flat. And the committee arguably works with faulty data. RUC recommendations are based on survey results of only about 2% of physicians, updated only every 5-20 years. Perhaps because of this, estimates of the time it takes to complete a given procedure—a vital component in calculating the complexity of care—are notoriously inaccurate.
In spite of these limitations, RUC recommendations are accepted without change by CMS more than 90% of the time, and commercial insurers largely base their payments on a multiple of the CMS charge as a baseline for negotiations with individual health systems.
Even though doctors can largely set their own rates without competition or pushback, they don’t get off scot-free. Because coding of routine visits is tied directly to the “complexity” of the patient, documentation requirements dictate that the average physician note in the U.S. is four times the length (paywall) of notes in peer countries. This is why you may have found notes from your doctor so long, repetitive, and bewildering. To make this worse, the advent of electronic health records has led to “chart bloat,” a phenomenon in which notes, thanks to cut-and-paste and other features, lead to an illusion of complexity and thus increased charges.
DRG rates, at least, are set by a slightly more predictable, scientific method. This isn’t to say that some gamesmanship doesn’t go into hospital billing; every physician in America has been coached on billing for the exact level of sickness of her patients at some point in her career. The words, “Don’t bill a uroseptic patient for a simple UTI” still ring in my ears from residency.
So does this mean the number of COVID-19 cases are being inflated?
In spite of these faults, there is no evidence that we’re over-attributing illness to COVID-19. We are still under-testing compared to most of our peer countries, and this is reflected in the mortality data we’re seeing. The “background” mortality rate in America is about 2.8 million deaths per year, with a little more than half of those deaths from cardiovascular disease and cancer. Deaths are seasonal and pretty steady year-over-year. But right now we’re seeing an excess mortality rate that is roughly double what COVID-19 accounts for. That is, only about half of observed excess premature deaths are in people diagnosed with COVID-19. So if anything, we are under-attributing deaths to COVID-19. After all, a patient who dies of a heart attack brought on by low oxygen levels and sticky blood due to an undiagnosed case of COVID-19 was still killed by COVID-19.
What about those increased payments for COVID-19 patients? It is true that hospitals make about 20% more for a patient infected with SARS-CoV-2. This is the result of the $100 billion slice of the federal stimulus passed in March that is allocated to hospitals. Why did hospitals get their own cut? Because volumes in hospitals are down by more than half as elective procedures like hip replacements and cardiac catheterizations—the lifeblood of hospital systems, for better or for worse—have been delayed or cancelled. Here is Harvard data on ambulatory visit volume through mid-April:
As a result, health care jobs—long considered “recession-proof,” are going away. Almost 43,000 health care jobs were lost in March alone. Health care is such a giant part of the American economy—a stunning $3.5 trillion per year, good for almost a fifth of gross domestic product, again, for better or for worse—that this reduction in health service delivery is thought to account for about half of our current loss of GDP. That’s why you hear our current financial predicament being referred to as a “health care-led recession.”
So if COVID-19 is a huge conspiracy to allow doctors, nurses, and hospitals to make extra money, it isn’t a very good one.