On these pages, Michelle Mello recently posted a discussion of her new article with Lindsey Murtagh, Thomas Gallagher, and Penny Andrew, called “Disclosure-And-Resolution Programs That Include Generous Compensation Offers May Prompt A Complex Patient Response.”
In this vignette-based online study, the authors put respondents in clinical scenarios with medical errors, and then added experimental conditions where the error was simply confessed, or confessed with an offer of waiver of the medical bills, with an offer to reimburse a limited amount of out-of-pocket expenses related (specifying $25,000 out of pocket plus $5,000 lost time), or with an offer of “full compensation.” As the headline suggests, the authors conclude that offers of full compensation may be sometimes be bad ideas for self-interested hospitals. I’m a big fan of this sort of vignette-based research, because it allows randomized manipulation that is impossible in observational field research. Still, allow me to offer some of my own questions and interpretations below the fold.
As I understand the study from looking at the authors’ appendix on the Health Affairs site, it appears that the “full compensation” condition did not specify a dollar amount. Methodologically, this could be a very important confound compared to the prior “limited” condition, which specifies $30,000. The lack of specificity may be doing the work in the “full” compensation condition, rather than whether the proposal was for “full” or “limited” compensation. Strangely, even without a specified amount, “nearly 60 percent of those offered full compensation said that they felt they should have been offered more.” More than what? According to the appendix, the vignette provided that, “They will pay you a sum for your pain and suffering that they say is comparable to what you could expect to recover in a lawsuit.” If patients thought the unspecified offer was insufficient, then perhaps that is because they presumed that they couldn’t trust the hospital (their adversary, which just injured them) to fill in the blank later, unilaterally deciding what amount is fair. That skepticism sounds prudent to me, and its hard to imagine a hospital structuring a liability waiver as an open-price contract, as was suggested here. Thus, I’m not sure the experiment tells us anything about what would happen in the real world if patients were actually presented with a check representing full compensation.
The “full compensation” condition is also confounded because it is the only one that includes a waiver of the right to sue. For the other conditions — those where less than full compensation was provided — the vignette provided that, “They [the doctor and the hospital manager] explain that if you accept this offer, you will still have the right to sue them.” This choice of research design is perplexing. Wouldn’t hospitals prefer to get waivers across the board? Isn’t the management of liability the whole point of these efforts by hospitals? I would be interested in seeing the impact of the demand for waiver itself, perhaps in a future study with a factorial design that can isolate that particular variable.
More generally, in these other conditions with no waiver requirement, the authors’ describe the dependent variable as “willingness to accept compensation.” (Unfortunately, the actual survey instrument does not appear in the authors’ appendix, so I’m not sure what was actually asked.) Normally, a willing to accept (WTA) or willing to pay (WTP) analysis presumes some sort of quid pro quo exchange. But here it appears to be just free money — say $30,000, and across conditions 20-30% are refusing it. Why? Does the question prompt really make it clear that they are leaving money on the table? Are the authors finding utterly irrational behavior here? The article says that “a small proportion (about 6 percent) of respondents showed problems comprehending the vignettes.” In my view, utterly irrational responses are evidence of a lack of comprehension, unless we have some theory of behavioral rationality to explain them. (Is this a status quo bias?)
Assuming that the subjects do understand the vignette and questions, then this may reflect a real-world distrust of the physician and hospital in this sort of scenario. More precisely, it may reflect a hesitancy or lack of confidence about being pressed to make such a decision without legal counsel. (Note that the “you’ll still have the right to sue” is itself legal advice, coming from the adversary.) Such a hesitance to make uncounseled decisions is probably the most charitable way to interpret the patient’s behavior, and it is a theme I would like to see developed much more clearly in this vein of scholarship.
When reporting results, the full article uses the section headline “interest in pursuing litigation” but then describes the survey question as whether the subject was “likely to seek legal advice about suing the physician.” In the discussion section, the authors tease apart the injured patient’s desire to have legal counsel from the desire to litigate. Future research could do so empirically. After all, from the other side of the coin, we do not suppose that hospitals are eager to litigate merely because they have hired attorneys. We should expect no less of rational patients.
On this point, the authors conclude the “seeking legal advice” discussion by noting that, “some institutions currently operating disclosure-and-resolution programs refer patients—particularly those who have been seriously injured or appear distrustful—to reputable plaintiff’s attorneys who provide good-quality representation and have demonstrated openness toward early settlement.” This trend is very worrisome to me. Early settlement is only in the interests of the patient if the hospital is making a very generous offer. Attorneys already have incentives to settle cases too quickly and too cheaply (adopting a volume strategy rather than a quality strategy). An attorney that is trying to cultivate referrals from the adversary has absolutely no ability to press the adversary. Indeed, a mere request for discovery of the material facts may get the lawyer bumped from the sycophant list to the dispreferred list. I think that these sorts of referrals should be recognized as a per se violation of the Rules of Professional Conduct, due to the inherent conflict of interest. (If the hospital wants to provide a link to a website maintained by the plaintiff’s bar, that is a different proposal of course.)
Overall, Murtagh and colleagues have thrown open the doors for more research in this vein — experimental, observational, and analytical. .