Consider me a skeptic of the "Sorry Works" idea, which suggests that if doctors immediately disclose all potential errors and apologize, total malpractice costs will go down. To the extent a disclosure regime has merit, reduction of litigation expense always seemed to me to be an implausible reason.
The one comparative empirical study, of a VA hospital (which, because of federal law, has a malpractice litigation environment somewhat different than that of private practitioners), had ambiguous results at best for its small sample. The Kentucky hospital that adopted it ended up having more claims than 30 out of 35 other VA hospitals, and four of those other five had significantly higher workloads. It did have the eighth-least amount of claims costs out of the 36 hospitals (not including legal fees, and we don't know whether those were higher or lower). But as I've discussed before, claims experience is not a smooth function, and just one more slightly above-average settlement or judgment against the hospital would've dropped it below the median. Moreover, the main proponent, Dr. Steve Kraman, acknowledged that 5 to 10 of the 14 claims a year brought against his hospital are still "illegitimate or nuisance" claims. (I'm not counting the bizarre tactics of some Sorry Works proponents against the merits of the idea, though the fact that the plaintiffs' bar is pushing the idea is a revealed preference for litigation regimes.)
That said, I don't think that this Studdard et al. paper in Health Affairs (via Childs), "Disclosure Of Medical Injury To Patients: An Improbable Risk Management Strategy," which purports to come to the same conclusion by modelling the effects of implementing "Sorry Works," adds anything substantive to the debate. The results are certainly striking:
The model computed a 5% chance that total claim volume would decrease or remain unchanged and a 95% chance that it would increase. The distribution also indicated a 60% chance that comprehensive disclosure of severe injuries would at least double the annual number of claims nationwide, and a 33% chance that volume would increase by threefold or more.
Under the assumption that average payments would not change, the model predicted a 6% chance that total direct costs of compensation would decrease or remain unchanged under routine disclosure and a 94% chance that they would increase. Under the assumption that disclosure reduced average payments by 40%, a net increase in costs remained more likely than a decrease or no change, and there was a 34% chance that costs would at least double.
But the methodology is far from persuasive:
We modeled the litigation consequences of disclosure by combining existing data on the epidemiology of medical injuries and malpractice claims with expert opinion about likely patient reactions to disclosure. We used Monte Carlo simulations to incorporate uncertainty around the experts' judgments of the proportion of patients who would be prompted to sue and the proportion who would be deterred from suing.
In other words, the vast majority of experts think that litigation will go up; those assumptions were built into the model, which promptly simulated that litigation would go up. That doesn't tell us anything new.