Sorry is Never Enough: The Effect of State Apology Laws on Medical Malpractice Liability Risk Benjamin J. McMichael* R. Lawrence Van Horn† W. Kip Viscusi‡ Last Modified: March 25, 2016 * JD/PhD Graduate, Vanderbilt University Program in Law and Economics. † Associate Professor of Management and Law, Executive Director of Health Affairs, Owen Graduate School of Management, Vanderbilt University. ‡ University Distinguished Professor of Law, Economics, and Management, Vanderbilt University Law School, 131 21st Avenue South, Nashville, TN 37203, [email protected]. Phone: (615) 343 – 6835. Fax: (615) 322 – 5953. Contact Author. 1 Abstract State apology laws offer a separate avenue from traditional damages-centric tort reforms to address medical malpractice liability. These laws facilitate apologies from physicians by excluding statements of apology from malpractice trials. Prior research suggests that apologies may assuage the anger of patients, decrease the number of claims filed, and lower settlement amounts. Using a unique dataset that covers all malpractice claims for 90% of physicians practicing in a single specialty across the country, this study examines whether apology laws limit malpractice liability risk. In general, apology laws increase the probability of malpractice lawsuits. Apology laws can reduce the average payment per malpractice case, but this decrease is primarily the result of an increase in the number of cases involving no payment to a claimant, not a decrease in the amounts that claimants actually receive. Overall, the evidence suggests that apology laws do not effectively limit medical malpractice liability risk. JEL Codes: I18, K13, K41 Keywords: Medical malpractice, apology law, tort reform, 2 I. Introduction The traditional response to an increase in the frequency or size of medical malpractice lawsuits includes reforms such as caps on noneconomic or punitive damages and changes to the traditional collateral source and joint and several liability rules. The common feature these reforms, which were enacted beginning in the 1970s and 1980s, share is their focus on post-liability damages awards. They either limit those awards directly (as with damages caps) or indirectly (as with collateral source and joint and several liability reform). However, reforms to damages awards are not state legislatures’ only options when attempting to attenuate medical malpractice liability risk. Another reform option, which has become popular among states over the last 15 years, is enacting laws that facilitate greater communication between patients and their providers (see Ho and Liu 2011a; 2011b). State apology laws are reforms to state rules of evidence and exclude from trials statements of apology, condolence, or sympathy made by healthcare providers (sometimes only physicians) to patients.4 Traditionally, if a physician expressed sympathy to a patient for a complication arising from care provided by the physician, the patient could later use that statement as evidence of malpractice in a lawsuit. State apology laws bar these statements made by physicians to patients from being admitted into evidence. Partial apology laws, which are the most common type of apology law, bar the admission of apologies and expressions of sympathy and condolence into evidence. Full apology laws bar the admission of these types of statements as well as outright admissions of error or liability. 4 In many states, apology laws apply to all types of individual healthcare providers such as nurse practitioners or physician assistants. Throughout this article, we will focus on physicians since our dataset includes only physicians. 3 Typically, states enact apology laws under the assumption that these laws facilitate communication between patients and physicians, which presumably will induce patients to file fewer lawsuits or accept lower settlement amounts (Ho and Liu 2011a). Patients report that, in addition to seeking compensation, they sue their physicians out of anger (Vincent, Phillips, and Young 1994). Apologies from their physicians could potentially assuage this anger and prevent some marginal suits from being filed. However, physicians are often advised to avoid apologizing to patients in order to avoid providing those patients with evidence that could be used in a lawsuit (Lamb et al. 2003). In other words, prior research suggests that physicians may not apologize to or communicate with patients in an effort to avoid lawsuits even when doing so could avoid a lawsuit. Apology laws address this paradox by allowing physicians to apologize to and communicate with patients with less risk of liability. This study evaluates the effect of state apology laws on medical malpractice litigation risk using a unique dataset from a large national malpractice insurer. We focus on one type of specialty physician.5 For this particular specialty, our dataset includes approximately 90% of all physicians,6 so we are able to analyze nearly the universe of malpractice claims filed against an entire specialty over an eight-year period. We observe all of the claims filed against those physicians whether or not a claim resulted in a positive payment to the claimant and whether or not it resulted in a formal lawsuit being filed. Thus, unlike most studies of medical malpractice litigation, which often include only observations of positive payments made at trial or as part of a settlement, we analyze malpractice cases regardless of their final disposition or the nature of the payment to the claimant. 5 For confidentiality reasons, we cannot reveal the specialty. 6 The insurer estimates that it insures more than 90% of this type of specialist physician, and we are able to verify this estimate using independent information provided in the Area Health Resource File, which includes data on the number of physicians across a wide variety of specialties. 4 Including all malpractice cases whether or not they resulted in a positive payment to a claimant is important because over half of all claims filed against physicians result in no payment to the claimant. However, even if a claimant receives nothing, the physician or insurance company must still expend (often substantial) resources defending the claim. Analyzing the effect of apology laws on all malpractice cases allows us to obtain a complete and an unbiased picture of the effect of these laws. Decomposing medical malpractice liability risk into the frequency of claims and the magnitude of the loss associated with those claims, we find that apology laws do not work as intended by state legislatures. In general, apology laws have only a small effect on the probability a physician faces any claim in a given year, but these laws increase the probability that a physician faces a lawsuit. For example, full and partial apology laws increase the probability that a physician not rated for surgery is a party to a lawsuit by 75% and 50%, respectively, relative to the national average. While full apology laws reduce the average payment received by claimants from physicians by $65,000, this reduction is due primarily to an increase in the number of claims resulting in no payment to claimants, not a reduction in the amount of money that is actually paid out to resolve a claim. These results are not consistent with either the intended effect of apology laws or prior research on apology laws and physician disclosure programs designed to encourage physician-patient communication. Overall, the results suggest that apology laws do not effectively limit malpractice risk and that these laws may not be able to accomplish the original goals of tort reform. The next section provides more detail on apology laws and their role in medical malpractice litigation. 5 II. Reforming Medical Malpractice and Apology Laws A. Medical Malpractice and Tort Reforms The debate over the need for and efficacy of tort reforms stretches back over 30 years. Groups, such as the American Medical Association (AMA), assert that physicians respond to rising liability costs by performing expensive and unnecessary procedures to avoid liability (defensive medicine), providing less care, retiring early, or relocating to states that have enacted tort reforms (AMA 2015). Since the first modern tort reforms began to emerge in the 1970s and 1980s, states have taken different tacks when attempting to attenuate medical malpractice risk. Damages-centric tort reforms, such as damages caps, joint and several liability reform, and collateral source rule reform, directly or indirectly reduce the amount of damages physicians must pay. Damages caps restrict different types of damages to less than the cap amount. Joint and several liability reform makes it more difficult for plaintiffs to recover an entire damages award from a single defendant when multiple defendants are held liable. Collateral source rule reform reduces plaintiffs’ damages awards by the amount of compensation they received from other sources. These tort reforms have been the most popular among states and have received the most attention in the law and economics literatures. A number of studies have considered the effect of damages-centric tort reforms on medical malpractice payments and premiums. Viscusi and Born (2005) and Born, Viscusi, and Baker (2009) show that tort reforms affect malpractice insurance losses, premiums, and loss ratios. A number of early studies focusing on the tort reforms of the 1970s and 1980s, which are reviewed by Zuckerman, Koller, and Bovbjerg (1986), find mixed evidence of the effect of damages-centric reforms on medical malpractice payments. Similarly, the Office of Technology Assessment (1993) and the Congressional Budget Office (2004) reviewed later waves of research examining the effect of reforms on malpractice payments in the late 1980s and 1990s and in the 1990s and early 2000s, 6 respectively. Both reviews conclude that while earlier studies suffered from various methodological flaws, damages caps were the only types of reforms to have a consistent effect on medical malpractice payments. Examining a large sample of positive medical malpractice settlements and judgments contained in the National Practitioner Data Bank (NPDB), Avraham (2007) finds similar results, with noneconomic damages caps reducing both the number of annual payments and the average size of those payments. Unlike traditional, damages-centric reforms, apology laws directly affect how plaintiffs establish the liability of defendant physicians but have no direct effect on the damages award. In states without an apology law, if a physician expresses sympathy for an injury or apologizes to the plaintiff, the plaintiff may generally introduce the physician’s statement as evidence of liability. In states with an apology law, these statements by the physician are excluded from evidence, making it more difficult for a plaintiff to establish the physician’s liability. The next section provides further background on apology laws. B. Background on State Apology Laws State apology laws come in two different varieties. Partial apology laws bar plaintiffs from admitting into evidence statements by physicians that express sympathy or condolence. Full apology laws immunize physician statements that express sympathy as well as outright admissions of fault, error, or liability. Both types of apology laws function as rules of evidence to exclude otherwise admissible evidence against physicians. We provide more background on apology laws in an online appendix. Table 1 provides a comprehensive overview of states’ adoption of the two different types of apology laws,7 and Figure 1 provides an overview of changes in state apology 7 In practice, the distinction between full and partial laws is not always clear. McDonnell and Guenther (2008) classify eight states as having a full apology law, the National Council of State Legislatures classifies six states as full apology law states, and Morse (2009) and Ho and Liu (2011a) classify five states having a full apology law. Our independent 7 laws between 2004 and 2011—the beginning and end of our data period, respectively. By 2004, 12 states had enacted partial apology laws and one state had enacted a full apology law. Between 2004 and 2011, 20 additional states enacted partial apology laws and four states enacted full apology laws.8 Thus, the time period over which our analysis occurs includes substantial variation in state apology laws. Figure 1 demonstrates that this variation was not concentrated in one area of the country or among rural or urban states, as states across the country enacted apology laws. In theory, apology laws can facilitate communication between patients and physicians by allowing physicians to express sympathy to assuage patient anger without fear that their communications could later be the basis of liability. Increased communication can theoretically reduce malpractice litigation by decreasing the motivation of patients to sue. Patients often cite the desire for remuneration as a reason for filing a malpractice suit, but many patients report suing their physicians for reasons unrelated to financial compensation (Hickson et al. 1992; May & Stengel 1990; Vincent & Young 1994). Hickson et al. (1992) find that a patient’s anger at her physician and breakdowns in communication between the two are often the primary reason for litigation, with a third of patients surveyed reporting that their physicians did not speak openly with them and half reporting that their physicians attempted to mislead them (see also Vincent et al. 1994). Other work has found that an apology from a physician can reduce patient anger, improve communication, and thereby reduce a patient’s desire to sue (Liebman & Hyman 2004, 2005; Sloan & Hsieh 1995). One survey found that about two thirds of physicians and patients agreed that improved communication could classification of apology laws is consistent with Ho and Liu (2011a). An online appendix provides additional background on how different state laws were classified as having full or partial apology laws. 8 Illinois enacted a partial apology law in 2005 and repealed it in 2010. It is the only state that has repealed an apology law. 8 reduce malpractice litigation (Shapiro et al. 1989). Similarly, when patients are presented with hypothetical situations, the likelihood that they litigate following a medical error generally decreases when they receive an apology (Mazor et al. 2004; Wu 1999; Wu et al. 2009). Apology laws can facilitate communication between individual patients and physicians, but they may also encourage hospitals and other institutions to adopt system-wide approaches to the disclosure of medical errors. Liebman and Hyman (2004, 2005) and Kraman and Hamm (1999) find that hospital-level programs designed to promote disclosure of medical errors and encourage apologies reduced the malpractice payments of hospitals in Pennsylvania and Tennessee, respectively. Boothman et al. (2009) and Kachalia et al. (2010) study a disclosure and apology program at the University of Michigan Health Service and find that claims decreased by 36%, lawsuits decreased by 65%, patient compensation costs dropped by 59%, and legal costs dropped by 61%. Studying the same program but focusing only on gastroenterology claims, Adams et al. (2014) find a statistically significant reduction in the number of patient encounters resulting in a claim, the average payment per claim, and the time to resolution of a claim. Apology laws are not necessary for hospitals to introduce early disclosure and apology programs (e.g., Michigan did not enact an apology law until 2011), but they can facilitate these programs by allowing hospitals to encourage apologies with less fear of subsequent malpractice litigation. Prior work focusing on specific disclosure programs has found results generally consistent with the justification for enacting apology laws. These programs reduce the number of claims filed, decrease the average payment per claim, and reduce the time between the initiation of a claim and claim resolution. However, these evaluations of apologies have either been limited to specific programs at specific medical facilities (e.g., Boothman 2009) or have not been able to observe all malpractice claims filed regardless of whether a payment was made to the claimant (e.g. Ho and 9 Liu 2011a; 2011b). In the next section, we describe our dataset which is both national in scope and inclusive of all claims regardless of whether the claimant received a payment. III. Data Our data come from a national malpractice insurer, and they include information on 90% of all physicians practicing in a single specialty from 2004 through 2014.9 The data include all physicians with a malpractice insurance policy with this insurer in each year and all malpractice claims filed against those physicians whether or not they resulted in a lawsuit or a payment to the claimant. For each physician in each year, we observe the state of practice and the beginning and ending date of the physician’s malpractice policy. The data also include whether the physician is rated for surgery. For this particular specialty, some physicians focus primarily on seeing patients in an office setting while others both see patients in an office and perform surgery. In total, the data include nearly 110,000 physician-years. Three different types of malpractice events appear in the data. “Incidents” are physician- reported instances of malpractice that do not involve a claim for payment by a patient. Because they are voluntarily self-reported, we ignore these events in our analysis.10 “Claims” involve a patient demand for compensation but are resolved prior to the filing of a lawsuit either with or without the patient receiving compensation. “Suits” involve actual lawsuits whether or not the patient received any compensation. We do not observe whether a payment or non-payment as part of a lawsuit is the result of a settlement or pursuant to a judgment. 9 For confidentiality reasons, we cannot reveal either the malpractice insurer or the specialty of the physicians in our dataset. To determine the percentage of all practicing specialty physicians included in our dataset, we obtained information on the number of physicians practicing in this specialty each year from the Area Health Resource File. In any given year, the number of physicians included in our dataset is between 89% and 93% of the number of physicians listed in the Area Health Resource File. 10 Some medical malpractice insurers require incident reporting for coverage of malpractice events. The insurer from which we obtain our data did not require incident reporting as a condition of coverage over the relevant study period. 10