Medical Malpractice Reform: Noneconomic Damages Caps Reduced Payments 15 Percent, With Varied Effects By Specialty

Abstract

The impact of medical malpractice reforms on the average size of malpractice payments in specific physician specialties is unknown and subject to debate. We analyzed a national sample of 220,653 malpractice claims from 1985–2010 merged with information on state liability reforms. We estimated the impact of state noneconomic damage caps on average malpractice payment size for physicians overall and for 10 different specialties, and compared how the effects differed according to the restrictiveness of the cap ($250,000 vs. $500,000 cap). We found noneconomic damage caps reduced payments by $42,980 (15%; p<0.001), with a $250,000 cap reducuing average payments by $59,331 (20%; p<0.001), while a $500,000 cap had no significant effect. Effects varied according to specialty and were largest in specialties with high average payments, such as pediatrics. This suggests that the effect of noneconomic damage caps differs by specialty, and only more restrictive caps result in lower average payments.

INTRODUCTION

Medical malpractice liability remains one of the most hotly debated issues in health policy and an area of intense concern among physicians.^1,2 While proponents of the system argue that it is an important tool to protect patients against negligent care, opponents argue that it is costly and inefficient and does a poor job of delivering compensation to patients.^3,4 In response to these concerns, many states have enacted legislative reforms that either directly or indirectly limit the damages that patients can recover in medical malpractice cases. However, some have questioned whether malpractice liability reforms have successfully lowered the cost of malpractice liability or assuaged physician fears of malpractice.^5,6

Previous studies that have examined the impact of malpractice reforms on expected liability have produced mixed results. Several studies have found that limitations on allowable damage recovery, particularly for noneconomic damages (defined as payments for pain and suffering, loss of companionship, etc.), reduce the average size of malpractice payments,^7–11 while others have found no effect.^12,13 There is a similar lack of consensus on the relationship between malpractice reform and the frequency of claims, with several studies finding that the adoption of liability reform lowers the probability that physicians experience a malpractice claim^7,9,14 and others finding no effect.^8,12,13

Several limitations of existing studies may explain the lack of definitive evidence on the impact of malpractice tort reforms on the size and frequency of malpractice payments. First, prior studies focus on the effect of malpractice reforms on all physicians rather than on distinct specialties. Malpractice risk varies considerably across specialties,^15–19 driven by differences in patient populations, rates and risks of procedural use, and physician characteristics, indicating that there is no reason to expect malpractice reform to have a uniform impact across all physicians. Second, prior studies usually rely on highly aggregate measures of malpractice reforms and, with several exceptions^20,21, do not distinguish between the different dimensions across which states have adopted reforms. A cap on damages with a lower value will be more restrictive, both because it will affect more cases and lead to lower average payments in cases to which it applies. Just as we expect reforms to impact liability risk differently across specialties, we expect reforms that impose different restrictions to have different effects as well.

Additionally, previous studies suffer from important data limitations, ranging from small samples from individual insurers^8,22 or a single state,²¹ to data from the National Practitioner Data Bank (NPDB). Although the NPDB theoretically includes all payments made to claimants on behalf of health care providers,²³ it has been criticized for under-reporting payments, in part because of the so-called “corporate shield” loophole that ignores payments made by hospitals or corporate entities, but also due to underreporting of lawsuits that jointly involve physicians and hospitals, in which hospitals assume the full liability payments so that their physicians are not recorded in the NPDB.²⁴ Moreover, although the NPDB includes information on whether an adverse event was broadly related to surgery, obstetric procedure, or anesthesia, it does not include information on specific physician specialty; the NPDB also specifically states that surgery, obstetric, and anesthesia codes may not refer to physicians who were actually surgeons, obstetricians, or anesthesiologists.

Using data from the Physician Insurer’s Association of America (PIAA) Data Sharing Project (DSP), the largest available collection of paid and unpaid malpractice claims from private insurers in the U.S., we estimated the effect of state malpractice reforms from 1985 to 2010 on the payment size of malpractice claims according to different physician specialties. In addition, we compared the effects of restrictive noneconomic damage caps of $250,000 to less restrictive caps. We estimated these effects overall and for 10 distinct categories of physician specialty.

METHODS

Malpractice Claims Data

We used data on malpractice claims by medical specialty and state from the PIAA Data Sharing Project from 1985 to 2010. The PIAA is an organization of insurers that currently includes 60 domestic and 12 international insurers plus 46 affiliated members. The Data Sharing Project includes medical professional liability claims against physicians insured by approximately half of its member companies, with claims from all 50 US states (medical professional liability is the name of the insurance line that medical malpractice claims fall under). PIAA’s member companies insure over 325,000 medical practitioners, and the Data Sharing Project contains information on all closed claims and claims open for one year or longer. The data represent approximately 25% of medical malpractice claims in the US and have been used to study malpractice risk in past studies.^14,25–29

The PIAA provided data for this study through an agreement with the RAND Corporation. The PIAA was not involved in the design or conduct of our study or in the interpretation and reporting of findings. We have no relevant conflicts of interest to report.

Our data consisted of the total number of malpractice claims, the number of claims that involved some indemnity payment to the plaintiff, and total indemnity payments (if any), at the state-specialty-year level. Consistent with past work,^17–19,30 we only included claims that involved some defense costs in order to eliminate cases in which there was a suspicion but no formal allegation of negligence was ever made. We used these data to compute average malpractice payment per specialty, state and year for claims in which an indemnity payment to a patient was made. It is important to note that the relationship between tort reform and the frequency of medical malpractice claims has been a common target of study in past work. However, despite its strengths, the Data Sharing Project has limited utility for studying the frequency of claims because it does not collect any information on the exposure of its member organizations (i.e., number of physicians insured). Thus, there is no denominator against which to normalize the number of claims.

We aggregated our data into the following specialties: internal medicine (including family practice), general surgery, surgical sub-specialties, obstetrics and gynecology, pediatrics, cardiology, neurology, anesthesiology, radiology, and ophthalmology. While PIAA data were available for more detailed specialties, other specialties were relatively small with too few claims to examine separately and so were combined into an “other” category. All dollar values were adjusted to 2010 values using the Consumer Price Index. All analyses were done using STATA™ Version 12. The data were de-identified and exempted from human subjects review by the institutional review board at RAND.

Malpractice Reform Data

We used published historical data on state level tort reforms from the Database of State Tort Law Reforms (DSTLR 4th).³¹ This database represents a systematic effort to provide a comprehensive list of malpractice reforms in place by state and year, and have become widely used in empirical work on malpractice. The database tracks which reforms were rejected by state courts, allowing us to identify whether a reform was in effect for a particular year. The database is currently available through 2010. We merged the DSTLR with the PIAA claims data at the state-year level.

We created indicator variables for whether a state had in effect in a year any of the following reforms: statutory caps on noneconomic damages (our main variable of interest), caps on punitive damage awards, restrictions on the contingency fees that attorneys charge their clients, reforms to the collateral source rule and reforms to the rules on joint and several liability. We focused on caps on noneconomic damages because these caps have been the most controversial reform in the last several years and among all types of tort reforms have been argued to be the most impactful, reducing average award sizes by 20% – 30%.²⁰ While some caps are indexed to the rate of inflation, so that they rise over time, we record the nominal value of a cap in place in each year.

Statistical Analysis

We estimated the impact of malpractice liability reforms on average payments using a difference-in-differences approach. By estimating changes in malpractice risk within states before and after the adoption of reforms, this approach accounts for fixed differences between states in malpractice risk and national trends that affect malpractice risk in all states.

We estimated multivariate linear regressions with average payment amounts as the dependent variables. The main independent variables were state-year indicators for specific malpractice tort reforms, of which noneconomic damage caps were the primary reform of interest. Other covariates included state fixed effects to account for time-invariant differences across states in their health care systems and liability regimes and year fixed effects to control for time trends. To control for demographic trends across states, we also included average income per capita, average age of the population, and the average percent of the population that was white using data from the Area Health Resource File (AHRF) published by the Health Resources and Services Administration (HRSA). To capture the overall effect of reforms across specialties, we estimated a regression which pooled specialties together. To capture specialty-specific effects of reforms, we used an interacted model that allowed for differential effects of the cap across specialties.

All regression models and sample means were weighted by the counts of total claims in each specialty-state-year combination. State-year observations in which there were no closed claims with an indemnity payment were not included. Standard errors were clustered at the state and specialty level to allow for correlation in the error terms over time.³² F-tests were used to test whether the interactions between specialty and noneconomic damage caps were jointly different from zero.

Our regressions included all reform types in each model, although we focused our discussion on the effects of noneconomic damage caps. (In the Appendix, we report estimates for the impact of other reforms on claim outcomes).³³ In order to estimate the impact of the restrictiveness of damage caps on indemnity payments, our models allowed the effect of noneconomic damage caps to differ according to cap size. Specifically, we estimated the effect of adopting less restrictive and more restrictive damage caps, defined as a cap value of $500,000 or $250,000, respectively, compared to no cap on noneconomic damages at all. For context, $250,000 is the size of the cap adopted by the California Medical Injury Compensation Act, which is often used as a model by proponents of reform.^34,35

Limitations

Our study had several limitations. First, we did not address the implications of adopting noneconomic damage caps for other important outcomes, such as the cost and quality of medical care or on patient safety outcomes. While evidence about the impact of malpractice reform on patient outcomes is mixed,^14,36–38 advocates for the civil justice system argue that shielding physicians from the full cost of malpractice liability could give incentives to take less care and reduce patient safety. Second, while the PIAA DSP is the largest available database of medical professional liability claims and covers all 50 states, it was not designed to be nationally representative. Third, as with any retrospective analysis of state policies using a difference-in-differences methodology, we assumed that the policy adoption was uncorrelated with other, unobserved factors that also drive claim outcomes. Fourth, our estimated reductions in average indemnity payments associated with stringent caps may underestimate the true effect if stringent caps reduce malpractice claims and those claims that are discouraged have marginally weaker merits and lower average indemnity payments.

RESULTS

Our data included 220,653 total claims, of which 74,366 (33.7%) involved indemnity. This is a higher proportion of paid claims than in some past work,¹⁷ but consistent with other studies.³⁹ The proportion of claims which involved indemnity varied across specialties (Exhibit 1), with the highest proportions in obstetrics and gynecology (39.4%), general surgery (38.8%), and anesthesiology (38.3%).

Exhibit 1.

Medical malpractice claims by physician specialty, 1985–2010

	Number of Claims	Number (%) of Claims with Indemnity
All Physicians	220,653	74,366 (33.7%)
Anesthesiology	8,151	3,123 (38.3)
Cardiology	4,225	911 (21.6)
Internal Medicine	55,390	17,870 (32.3)
General Surgery	23,245	9,030 (38.8)
Obstetrics and Gynecology	32,666	12,860 (39.4)
Ophthalmology	6,339	2,085 (32.9)
Pediatrics	6,340	2,045 (32.3)
Radiology	12,733	4,267 (33.5)
Surgical Specialties	46,451	14,685 (31.6)
Other Specialties	25,113	7,490 (29.8)

Source/Notes. Source: Data come from the Physician Insurers Association of America (PIAA) Data Sharing Plan database (see text for details).

Exhibit 2 reports trends in the average payment for paid claims by year from 1985 to 2010. Average payments rose steadily from 1985 through 1994. After a brief decline, there was rapid growth until 2002. After 2003, average payments leveled off, ranging from $90,000 to $110,000 per year. Across all years, the average payment among claims with indemnity was $291,433.

EXHIBIT 2.

Trends in average malpractice payments, 1985–2010

Source/Notes: SOURCE Authors’ analysis of PIAA data. NOTES Figure reports the average payment in the PIAA year (Panel B). PIAA data are at the claim level, and only include claims that involved some indemnity payment to the plaintiff.

Exhibit 3 reports the number of states with noneconomic damage caps in place by year and cap size. From 1985 to 2010, the majority of caps ranged from $250,000 to $500,000, with a small number of states adopting caps at $250,000 or greater than $500,000. The majority of caps were adopted in the late 1980s or the early 2000s, corresponding to periods of instability in the malpractice insurance markets.⁴⁰ The national average payment stabilization appears to coincide with increases in the number of states with noneconomic damage caps in 2003, though this simple correlation does not control for other potentially confounding factors addressed by regression analysis.

EXHIBIT 3.

Trends in average malpractice payments and noneconomic damage cap adoption, 1985–2010

Source/Notes: SOURCE Authors’ analysis of Ronen Avraham’s Database of State Tort Law Reforms (4^th Edition). NOTES Figure reports the number of states with noneconomic damage caps in place by size of cap and year.

Exhibit 4 reports estimates of the effect of a state’s adoption of a noneconomic damage cap on average malpractice payments, both overall and by physician specialty. These estimates do not discern between caps of varying restrictiveness. Overall, noneconomic damage caps reduced average malpractice awards by $42,980 (p<0.001), about a 15% reduction. Noneconomic damage caps had variable effects across specialties. Noneconomic damage caps had their largest effects on indemnity payments in pediatrics ($116,662 reduction; p<0.001), a specialty known to have a low frequency of claims but among the highest average payments.⁴¹ Noneconomic damage caps also lead to large and statistically significant reductions in average payment size in obstetrics and gynecology ($104,809 reduction; p<0.001) and cardiology ($57,480 reduction; p=0.05). Noneconomic damage caps had a lower than average impact on payments in surgical specialties, even though these specialties tend to be at highest risk in terms of frequency of claims.¹⁷

EXHIBIT 4.

Effect of noneconomic damage caps on average malpractice payments according to physician specialty

Source/Notes: SOURCE Authors’ analysis of PIAA data. NOTES Figure reports the estimated impact of noneconomic damage caps on average payments by physician specialty. The estimates are based on difference-in-differences regression of average payments against the presence of noneconomic damage caps as well as fixed effects for physician specialty, year, state, the presence of other forms of tort reform in the state (punitive damage caps, joint and several liability reform, collateral source rule reform and caps on attorney contingency fees) and state demographics (race and per capita income). Specialty-specific estimates come from a separate regression with cap-specialty interactions. The dot represents the coefficient estimate and the line the 95% confidence interval. Confidence intervals reflect variance estimates calculated to allow for clustering at the state and specialty level.

Importantly, the estimates in Exhibit 4 do not differentiate between more and less restrictive caps. More restrictive noneconomic damage caps would be expected to have a greater impact on malpractice award size especially for specialties with high average payments. Exhibit 5 reports estimates of the effect of noneconomic damage caps on malpractice payments, differentiating between a restrictive cap of $250,000 and a less restrictive cap of $500,000 for noneconomic damages. Average payment amounts are reported to provide context. Consistent with past findings,¹⁷ there was wide variation in the size of average payments by specialty.

Exhibit 5.

Estimated Impact of Noneconomic Damage Caps on Average Payments by Cap Size and Physician Specialty

		Effect of $500,000 cap on average indemnity payment			Effect of $250,000 cap on average indemnity payment
	Average payment ($s)	Dollar effect	% effect	p- value	Dollar effect	% effect	p- value
All Physicians	293,645	−17,866	−6.1	0.18	−59,331	−20.2	<0.01
Anesthesiology	344,356	29,349	8.5	0.68	−42,552	−12.4	0.09
Cardiology	325,337	−27,735	−8.5	0.48	−79,235	−24.4	<0.01
Internal Medicine	268,900	−40,232	−15.0	0.03	−51,225	−19.0	<0.01
General Surgery	267,007	−27,052	−10.1	0.09	−38,135	−14.3	0.01
Obstetrics and Gynecology	376,845	−65,307	−17.3	0.06	−124,005	−32.9	<0.01
Ophthalmology	256,380	−37,067	−14.5	0.30	−17,728	−6.9	0.26
Pediatrics	390,141	−31,471	−8.1	0.57	−146,481	−37.5	<0.01
Radiology	271,760	−4,705	−1.7	0.86	−37,810	−13.9	0.09
Surgical Specialties	259,228	2,542	1.0	0.90	−35,717	−13.8	<0.01
Other Specialties	304,052	40,604	13.4	0.19	−47,599	−15.7	<0.01

Notes: The table reports the estimated impact of noneconomic damage caps on average payments according to physician specialty and the size of the damage cap. Estimates are based on difference-in-differences regression of average payments against the presence of noneconomic damage caps and the size of the cap, as well as fixed effects for physician specialty, year, state, the presence of other forms of tort reform in the state (punitive damage caps, joint and several liability reform, collateral source rule reform and caps on attorney contingency fees) and state demographics (race and per capita income). We report estimates for the estimated effect of two caps: one at $250,000 (more restrictive) and one at $500,000 (less restrictive). Specialty-specific estimates come from a separate regression with cap-specialty interactions. P-values reflect clustering at the state and specialty level.

Across all physicians, a less restrictive $500,000 cap on noneconomic damages was associated with a small reduction in average award size ($17,866 or 6.1% reduction), however this reduction was not statistically significant (p=0.18). In contrast, a more restrictive $250,000 cap was associated with a substantial reduction in award size overall ($59,311 or 20.2% reduction; p<0.001). Generally, a less restrictive $500,000 cap had no effect on malpractice awards for various specialties (with one exception, internal medicine), whereas a more restrictive $250,000 cap was associated with statistically significant reductions in malpractice award sizes overall and for 7 of 10 specialties. The effect of a $250,000 cap was strongest for pediatrics, reducing average payments by $146,481 (37.5%, p<0.001), followed by obstetrics and gynecology ($124,005 or 32.9%, p<0.001) and cardiology ($79,235 or 24.4%, p=0.007).

Across specialties, the effect of a $250,000 cap was larger for specialties with higher average indemnity payments. For example, pediatrics, obstetrics and gynecology, anesthesiology, and cardiology were the specialties with the four highest average indemnity payments and were also those for which the estimated effect on malpractice awards of a $250,000 noneconomic damage cap was greatest. More generally, the correlation between average specialty malpractice award size and estimated specialty-specific restrictive cap effects was 0.82, suggesting a greater impact of restrictive caps for specialties with higher malpractice awards. A joint F-test of the specialty interaction terms with the noneconomic damage cap suggested that the effects were statistically different from one another for a damage cap of $250,000 (p<0.001).

Additional results

The Appendix reports the effects of other reform types (caps on punitive damage awards, restrictions on the contingency fees that attorneys charge their clients, reforms to the collateral source rule and reforms to the rules on joint and several liability) on average payments, estimated in the same differences-in-differences model as the noneconomic damage cap analysis.³³ Other reforms had little consistent effect on outcomes, except caps on attorney contingency fees, which were associated with higher payments. This is consistent with other evidence that these reforms may discourage attorneys from accepting lower value cases.⁴² We also found joint and several liability reform was associated with higher payments, also consistent with past findings.³⁷ The Appendix also reports results of the effects of noneconomic damage caps on all claims (paid and unpaid).³³ Noneconomic damage caps are associated with a reduction in average payment overall ($16,948 reduction; p<0.001), and had no effect on the share of claims resulting in payment. Finally, the Appendix verifies that our findings and conclusions are robust to several sensitivity analyses, including: (1) a discrete, nonparametric approach to incorporating the size of the cap on average payments, (2) inclusion of lead effects of tort reforms in order to assess for pre-existing trends in malpractice awards in states implementing reforms, (3) inclusion of lag effects to account for the possibility of delayed onset of the impact of the cap, (4) clustering of standard errors at the state rather than state-specialty level, (5) focusing on caps that apply to all tort cases as opposed to caps that apply only to malpractice cases, and (6) adjustment for state-year health maintenance organization penetration rates.

DISCUSSION

We analyzed the impact of noneconomic damage caps and other medical malpractice reforms on the size of malpractice indemnity payments according to physician specialty. We found that restrictive noneconomic damage caps were associated with lower average payments across all specialties with particularly large reductions among specialties with high average payment sizes such as pediatrics and obstetrics. In addition, we found that the size of the noneconomic damage cap was important to determining whether the cap had any impact. For example, whereas a $250,000 cap reduces average indemnity payments by more than a third in several specialties, a less restrictive $500,000 cap generally has no effect.

Our analysis contributes to prior work in two important ways. First, we analyzed how malpractice reforms affect various physician specialties differently. Malpractice risk is known to vary considerably across specialties,^15–19 due to differences in patient populations, rates and risks of procedural use, and physician characteristics, all of which may lead malpractice reforms to have distinct impacts across specialties. Second, we analyzed how the effects of noneconomic damage caps on payment size vary according to the restrictiveness of the cap. In particular, we estimated the potential impact of two relatively common cap sizes – a $250,000 damage cap, often considered the model for tort reform advocates, or a less restrictive $500,000 cap – and compared their impact on claims in different specialties. Our findings complement prior work which used data from Texas to simulate the impact in other states of noneconomic damage caps of varying stringency²¹; analyses of the NPDB which demonstrated that states implementing more stringent caps on noneconomic damages during 2003–2006 had larger declines in malpractice claim rates and average indemnity sizes²⁰; and prior studies exploring the impact of cap stringency on outcomes such caesarean section rates (proxy for defensive medicine) and physician labor supply.^43–45

The impact of a more restrictive $250,000 cap was greatest for specialties with higher average indemnity payments (such as pediatrics, obstetrics and gynecology, and cardiology), rather than for specialties which have traditionally had more frequent claims. For example, surgical specialties are typically considered the highest risk due to the size of malpractice premiums and frequency of payments, but may actually benefit less from restrictive caps compared to other specialties (in terms of lower payment sizes) since indemnity payments in surgical subspecialties are generally similar in magnitude to other lower-risk specialties. Put differently, restrictive noneconomic damage caps would not be expected to have large effects in cardiovascular-thoracic surgery, a commonly considered ‘high-risk’ specialty which in a prior study ranked second-highest in the rate of malpractice claims but was average compared to other specialties in the size of indemnity payments.¹⁷

Our findings may also help clarify why prior literature has produced mixed results on the impact of tort reforms on malpractice risk, with most evidence suggesting that noneconomic damage caps reduce the size of malpractice payments by as much as 20% to 30% and also reduce the frequency of claims^7–11,43; however, some other studies find no effect.^12,13 Our findings not only unpack the results of these prior studies into differential effects of malpractice reforms across physician specialties and differential effects of more versus less restrictive caps, but also suggest that the indeterminacy of prior studies may be partially due to their limited analyses of these two important dimensions.

The implications of our findings for policy depend in part on the goal of noneconomic damage caps. If the intent is to lower the expected cost of malpractice liability for the average physician, our findings suggest that only more restrictive caps, such as the $250,000 noneconomic damage limit utilized by the California MICRA and followed by 7 other states, are likely to have a significant effect on malpractice liability. If however, the goal of caps is to trim extreme verdicts in outlier cases, without necessarily impacting the average or “typical” case, then a less restrictive cap might suffice (though past work indicates that malpractice policy limits might already accomplish this⁴⁶).

Our study also relates to ongoing policy debates over the size of noneconomic damage caps. For example, a proposed ballot initiative in California would raise the noneconomic damage cap from $250,000 to $1.1 million. As this would exceed the typical policy limit, which is usually binding in terms of payment size, this would effectively eliminate the cap for the majority of cases. Although the specific effect of this policy would depend, however, on the distribution of policy limits across physicians in California, our findings suggest this would lead to about a 20% increase in average indemnity payments, with larger increase in obstetrics and gynecology and pediatrics (from Exhibit 5).

Our findings do not necessarily mean that damage caps are socially beneficial. Others have questioned the fairness of noneconomic damage caps by pointing out that they have the largest effect on patients with grave and disfiguring injuries.⁴⁷ Our study was also limited to assessing the impact of noneconomic damage caps on claim outcomes. A comprehensive assessment of the implications of noneconomic damage caps for social welfare would require a complete accounting of the costs and benefits of reform, including the potential impact on patients.

Conclusion

Our study provides new insights into the relationship between the adoption of noneconomic damage caps and medical malpractice liability risk for physicians in different specialties. The restrictiveness of specific malpractice policies also determines whether those policies have any effect on malpractice awards. Future evaluations of malpractice reforms and the design of malpractice policy should consider both the types of physicians who are most likely to be affected by reforms as well as the differential impact that reforms may have depending on how restrictive they are.