It’s time for us all to face a hard truth: you can, in fact, compare apples and oranges. Really, they’re not all that dissimilar. The fundamental traits of each are pretty much the same: they’re both spherical fruits and they both can be made into delicious candy. By almost any measure, they’re much more similar to each other than they are to just about anything else. If you’re trying to learn about fruit, you’d be remiss not to include apples and oranges as data points.
Statisticians and research scientists would say that information about apples and oranges is poolable. The term describes data sets that are similar enough to be included in the same analysis or meta-analysis. Sometimes moral and social considerations can complicate the question,1 but mostly the issue is whether, statistically and clinically, the data sets are homogeneous enough to be thrown into a pot and analyzed together. Basically, homogeneous data are poolable, but heterogeneous data are not.
I’m only bringing all this up because I’m not especially convinced that all the data we have about the effect of tort reform on physician malpractice is poolable. The first problem you encounter when trying to wade through the metric ton of studies on medical malpractice is that there’s an incredible amount of disagreement. Pick any 2 articles on the topic and they’re as likely as not to have come to opposite conclusions.2,3 Maybe some of the discrepancies can be explained away by confirmation biases, ideological differences, or varying methodologies — but maybe they can’t, and, in any event, the lack of consensus renders intelligent policy debate and design next to impossible.
Nearly all of those studies try to quantify the effect of the various malpractice claim regulations on physician practice by linking it to the prevalence of defensive medicine, which is when a physician alters his or her clinical approach for the purpose of reducing malpractice exposure. That’s a reasonable basis for going about the analysis, since most investigations agree that malpractice limits are correlated with defensive practices. The idea works intuitively, too. It’s easy to imagine a physician — perhaps even the one in the mirror — becoming concerned about excessive malpractice premiums and either ordering needless tests or avoiding certain patients altogether in order to limit his or her exposure.
But, as it happens, most of the investigations only consider one of those two responses. There’s plenty of analysis discussing the impact of malpractice policy on how many extra tests physicians will order when trying to avoid a lawsuit; that’s what researchers call “positive defensive medicine.” On the other hand, there’s quite a bit less information out there about whether and to what extent malpractice pressure causes physicians to engage in “negative defensive medicine” — that is, to avoid particular high-risk patients entirely.
There are even fewer studies that consider both genres of defensive medicine simultaneously, which is a shame because the little data we do have are pretty intriguing. The best example I could find was a paper from a couple of years ago that took a shot at modeling the effects of malpractice pressure on negative and positive defensive medicine in the most realistic way that it could.2 The study came to a couple of key conclusions. The first was that healthcare spending and healthcare quality always move in the same direction in response to changes in the malpractice environment. This suggests that — much to the chagrin of think tanks across the political spectrum — for any group that’s subject to the same malpractice threat, tort reform is unlikely to both decrease costs and increase quality. The second finding was that health spending and quality will increase with rising malpractice pressure up to a certain threshold, at which point everything flips and further increases in malpractice liability result in healthcare spending and quality decreases; basically, if you plot malpractice burden on the x-axis and healthcare spending/quality on the y-axis, the graph would be an upside-down V. Since we know that relationship, it follows that if you can quantify the current malpractice burden on a particular population, then you can accurately predict how (if not precisely how much) tort reform will affect that group.
These results haven’t been widely replicated, but they’re at least a step toward reconciling the inconsistencies seen in malpractice reform studies; many of the studies conclude that increasing malpractice pressure (by doing things like lifting liability caps or making it easier to file a lawsuit) would increase healthcare spending, based on data derived from environments with relatively low malpractice pressures; those are data points on the upslope of the upside-down graph that I mentioned earlier. Conversely, studies with the opposite findings were usually carried out in high-pressure malpractice environments; data points on the downslope of the graph.
To some extent, the particulars of any given malpractice environment are dictated by state law; places like California and Texas, for instance, have hard caps on non-economic damages that systematically limit payments to victims of physician negligence. But, even more than geography, malpractice liability is related to physician specialty.4 Some specialties, like obstetrics, get sued (and are forced to make payments to wronged patients) all the time; others, like cardiology, don’t. The difference is made obvious by the wide variation in litigation frequency, which is directly related to the risk of a negative judgment or costly out-of-court settlement. These data are public, consistent, and conveniently can be used to place each specialty on our quality and spending graph — cardiology near the left, obstetrics on the right — which in turn tells us how tort reform is likely to affect each specialty’s outcomes. Increasing the malpractice pressure on cardiologists will increase quality and spending in that sector; the same change will decrease those measures for obstetrics.
All of which is to say that malpractice liability reform may or may not be a worthy policy, but that issue shouldn’t be analyzed for the profession as a whole, but instead on a specialty-by-specialty basis. All we need is to figure out the optimal quality and spending levels for each specialty — no mean feat, to be sure — and we’re off to the races. Sometimes you need to put the apples with the apples, and the oranges with the oranges.
- Chan L, Macdonald ME, Carnevale FA, Steele RJ, Shrier I. Reconciling disparate data to determine the right answer: a grounded theory of meta analysts’ reasoning in meta‐analysis [published online July 25, 2017]. Res Synth Methods. doi: 10.1002/jrsm.1258
- Montanera D. The importance of negative defensive medicine in the effects of malpractice reform. Eur J Health Econ. 2016;17(3):355-369.
- Lakdawalla DN, Seabury SA. The welfare effects of medical malpractice liability. Int Rev Law Econ. 2012;32(4):356-369.
- Jena AB, Seabury S, Lakdawalla D, Chandra A. Malpractice risk according to physician specialty. N Engl J Med. 2011;365(7):629-636.