I am getting caught up on the news after a couple of weeks away and two stories caught my attention. The first is the ongoing debate about the tax exempt status of Illinois nonprofit hospitals, which has received extensive coverage in the Chicago Tribune. Nonprofits avoid paying most state taxes, notably property taxes. For some nonprofit, the tax exemption could be worth tens of millions of dollars annually.
The question before the state is what they should expect of nonprofits in exchange for tax exemption. The current law requires nonprofits to provide “community benefits” commensurate with their tax savings. The state and the Illinois Hospital Association have been unable to find mutually acceptable language to replace this vague standard. The most draconian approach limits community benefits to charity care. At the other extreme, the IHA (and the Chicago Tribune) largely back a proposal by the Civic Federation that defines community benefits broadly to include losses incurred on Medicare, Medicare, bad debt, and community outreach programs.
A few years ago, I advised the state Attorney General’s office on this issue. I argued for the following conceptual approach: In exchange for tax exemption, nonprofit hospitals should be required to perform a commensurate level of “charitable acts,” which I defined as services and programs for which the hospital expects to lose money. Alternatively, charitable acts are those that investor-owned hospitals would not undertake.
This approach gives us some guidance as we consider what to include as community benefits. Consider bad debt, which is often the largest component of a hospital’s community benefit. If a hospital makes every effort to collect its medical bills from the uninsured but ends up writing off some of them, the resulting bad debt is not necessarily a charitable act. The fact that billed charges vastly exceed costs allows hospitals to simultaneously reap profits while incurring bad debt. For example, suppose that one in three uninsured patients pays their bills. Now consider a hospital that treats three patients at a cost of $3,000 each and bills each patient $10,000. The hospital collects $10,000 from one patient and nothing from the other two. Overall, the hospital has made a profit, and expected to turn a profit at the time it admitted these uninsured patients. There nothing charitable about the decision to admit them. If the hospital truly wishes to commit a charitable act, it should not bill the patients in the first place. (Clearly the accounting is a bit more complicated than this, as the hospital might not collect enough to fully cover its costs. But it is misleading to say that all unpaid bills represent community benefits.)
Now consider “losses” on Medicare patients. Medicare might be unprofitable in the accounting sense, but only if we compare inpatient revenues against fully allocated inpatient costs. Hospitals make money off of complementary outpatient services. Moreover, hospitals would still bear a substantial fraction of overhead costs even if they stopped taking Medicare inpatients. The bottom line is that Medicare is economically profitable, regardless of what the accountants tells us. The proof: investor-owned hospitals continue to seek out Medicare business. One can use similar logic to determine whether community outreach, teaching and education are truly charitable acts. (Free nutrition assessments in Chicago public schools; probably yes. Free mammograms at Old Orchard Shopping Center? Probably not.)
Overall, I would take a rather hard line against nonprofit hospitals. There is lots of research suggesting that there are few significant differences in the performance of nonprofits and investor-owned hospitals. Nonprofits may argue that without the tax exemptions they will have to cut back on investments, especially in modernization and new technology. They are undoubtedly correct, but let’s not get caught up in the rhetoric. Hospitals can and will continue to grow to meet demand. But there is no need to artificially subsidize growth. If taxpayers are going to give something to nonprofits above and beyond normal payments in exchange for care, then the nonprofits should give something back, something that is truly charitable.
The other item that caught my attention was a study in Health Affairs showing that physicians who have electronic access to test results (via electronic medical records) order more tests than physicians who do not. The New York Times ran both an article and an editorial about the study. The finding is based on a single year of cross-section data. All first year graduate students in the social sciences should immediately recognize that this study suffers from endogeneity bias. The “treatment” (i.e., access to test results through EMR) is potentially correlated with unobservable physician characteristics that end up in the error term of the regression. As a result, causality is in doubt.
Let me offer several candidate explanations for the findings. First, it could be that access to test results encourages doctors to find ways to increase testing (the negative connotation offered by the author and the New York Times.) Second, it could be that some doctors have a technological bent. They tend to order lots of tests and they are first to have access to EMR. Thus, there is a correlation between testing and EMR but the latter does not cause the former. Third, some doctors may order an above average number of tests. Forced by managed care and Medicare to reduce costs, they seek a way to get testing under control. They install EMR which gives them access to prior test results. This allows them to reduce testing , though not by enough to put them below the average. Again, there is a correlation between testing and EMR, but in this case EMR leads to less testing. There may be other possibilities. My point is not that I know what causes what. My point is that no one knows what causes what.
Fortunately, there are research methods that can help sort out causality. I happen to be working on a similar study using these methods. In an upcoming blog, I will share some preliminary findings. My coauthors and I get a similar correlation, but reveal an interesting pattern of causality that may provide hope to those who believe that EMR can rescue the healthcare system.