Association Between State Medical Malpractice Environment and Postoperative Outcomes in the United States

Accepted Manuscript Association between State Medical Malpractice Environment and Postoperative Outcomes in the United States Christina A. Minami, MD, MS, Catherine R. Sheils, BA, Emily Pavey, MA, Jeanette W. Chung, PhD, Jonah J. Stulberg, MD, PhD, MPH, David D. Odell, MD, MMSc, Anthony D. Yang, MD, MS, FACS, David J. Bentrem, MD, MS, FACS, Karl Y. Bilimoria, MD, MS, FACS PII:

S1072-7515(16)31703-3

DOI:

10.1016/j.jamcollsurg.2016.12.012

Reference:

ACS 8565

To appear in:

Journal of the American College of Surgeons

Received Date: 23 October 2016 Revised Date:

1 December 2016

Accepted Date: 2 December 2016

Please cite this article as: Minami CA, Sheils CR, Pavey E, Chung JW, Stulberg JJ, Odell DD, Yang AD, Bentrem DJ, Bilimoria KY, Association between State Medical Malpractice Environment and Postoperative Outcomes in the United States, Journal of the American College of Surgeons (2017), doi: 10.1016/j.jamcollsurg.2016.12.012. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Association between State Medical Malpractice Environment and Postoperative Outcomes in the United States

RI PT

Christina A Minami, MD, MS1,2; Catherine R Sheils, BA1,3, Emily Pavey, MA1; Jeanette W Chung, PhD1; Jonah J Stulberg, MD, PhD, MPH1,2; David D Odell, MD, MMSc1,2; Anthony D Yang, MD, MS, FACS1,2; David J Bentrem, MD, MS, FACS1; Karl Y Bilimoria, MD, MS,

1

SC

FACS1,2

Surgical Outcomes and Quality Improvement Center (SOQIC), Department of Surgery,

2

M AN U

Feinberg School of Medicine, Northwestern University, Chicago, IL

Center for Healthcare Studies in the Institute for Public Health and Medicine, Feinberg School

of Medicine, Northwestern University, Chicago, IL

University of Rochester School of Medicine, University of Rochester, Rochester, NY

TE D

3

Disclosure Information: Nothing to disclose.

Support: Dr Bilimoria has received support from the National Institutes of Health, Agency for

EP

Healthcare Research and Quality, American Board of Surgery, American College of Surgeons, Accreditation Council for Graduate Medical Education, National Comprehensive Cancer

AC C

Network, American Cancer Society, Health Care Services Corporation, Northwestern University, the Robert H Lurie Comprehensive Cancer Center and Northwestern Memorial Hospital and honoraria from hospitals and professional societies for clinical care and quality improvement research presentations. Research for this project was done while Dr Minami was a National Research Service Award postdoctoral fellow at the Center for Education in Health Sciences under an institutional award from the Agency for Healthcare Research and Quality, T-32 HS 1

ACCEPTED MANUSCRIPT

000078 (PI: Jane L Holl, MD MPH). Dr Yang has received support from the American College of Surgeons. Disclaimer: None of these funding sources had any role in the design and conduct of the study;

RI PT

collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

M AN U

SC

Correspondence address: Karl Y Bilimoria MD, MS, FACS Surgical Outcomes and Quality Improvement Center (SOQIC) Department of Surgery and Center for Healthcare Studies Feinberg School of Medicine, Northwestern University 633 N St. Clair Street, 20th floor Chicago, IL 60611 [email protected]

AC C

EP

TE D

Running Head: Malpractice Environment and Surgical Outcomes

2

ACCEPTED MANUSCRIPT

ABSTRACT Background: The U.S. medical malpractice system assumes that the threat of liability should deter negligence but it is unclear whether malpractice environment impacts healthcare quality.

RI PT

We thus sought to explore the association between state malpractice environment and postoperative complication rates.

Study Design: This observational study included Medicare fee-for-service beneficiaries

SC

undergoing one of the following surgeries in 2010: colorectal, lung, esophageal, or pancreatic resection, total knee arthroplasty, craniotomy, gastric bypass, abdominal aortic aneurysm repair,

M AN U

coronary artery bypass grafting, or cystectomy. State-specific malpractice environment was measured by 2010 medical malpractice insurance premiums, state average award size, paid malpractice claims/100 physicians, and a composite malpractice measure. Outcomes of interest included 30-day readmission, mortality, and post-operative complications (e.g. sepsis,

TE D

myocardial infarction (MI), pneumonia). Using Medicare administrative claims data, associations between malpractice environment and postoperative outcomes were estimated using hierarchical logistic regression models with hospital random-intercepts.

EP

Results: Measures of malpractice environment did not have significant, consistent associations with post-operative outcomes. No individual tort reform law was consistently associated with

AC C

improved post-operative outcomes. Higher-risk state malpractice environment, based on the composite measure, was associated with higher likelihood of sepsis (OR 1.22, 95% CI 1.071.39), MI (OR 1.14, 95% CI 1.06-1.23), pneumonia (OR 1.09, 95% CI 1.03-1.16), acute renal failure (OR 1.15, 95% CI 1.08-1.22), DVT/PE (OR 1.22 95% CI 1.13-1.32), and gastrointestinal bleed (OR 1.18, 95% CI 1.08-1.30).

3

ACCEPTED MANUSCRIPT

Conclusions: Higher risk malpractice environments were not consistently associated with a lower likelihood of surgical post-operative complications, bringing into question the ability of

RI PT

malpractice lawsuits to promote healthcare quality.

Keywords: medical malpractice, postoperative complications, medical liability, surgical

AC C

EP

TE D

M AN U

SC

outcomes, quality of care

4

ACCEPTED MANUSCRIPT

INTRODUCTION Classic tort deterrence theory comprises one of the main theoretical struts of the U.S. medical malpractice system. According to this rationale, the threat of medical litigation should

RI PT

heighten physician vigilance and lower negligent care. It follows that quality of care and patient outcomes should be better in higher liability environments.1,2 Medical malpractice litigation is especially relevant to surgical specialties, which garner a disproportionate number of malpractice

SC

claims.3 In addition, surgeons appear to be well-aware of their malpractice risk, as their

perceived threat of malpractice suits has been shown to be the highest across medical and

M AN U

surgical specialties.4

Critics of this system often argue that the threat of malpractice results in “defensive medicine,” or the ordering of tests, procedures, or visits, and avoidance of high-risk patients or procedures primarily out of concern for malpractice liability,5 leading to a preponderance of

TE D

studies on this “unintended” consequence of medical liability.6 How medical malpractice environments affect the quality of care however, is less clear.7 Previous studies of the association between malpractice environment and surgical outcomes have been limited by scope, including a

EP

focus on individual surgical subspecialties, unique procedures or diseases, and disparate measures to characterize malpractice environments.6 Furthermore, studies of the association

AC C

between malpractice environment and patient outcomes have not demonstrated consistent results. A number of studies found evidence of deterrence,8-11 while others documented no evidence of an association between malpractice environment and healthcare quality.6,7,12-19 There is thus the need for further assessment of the effect of malpractice environment on surgical quality. Our objective was to determine whether state malpractice environment is associated with post-operative complications in patients undergoing a range of surgical procedures. We 5

ACCEPTED MANUSCRIPT

hypothesized that there would be no significant, consistent associations between malpractice environment and the likelihood of postoperative complications.

RI PT

METHODS Data Sources and Study Sample

The study cohort was constructed using CMS Medicare 2009-2010 fee-for-service data

SC

for patients undergoing one of eleven surgeries: colon, rectal, lung, esophageal, or pancreatic resection, total knee arthroplasty (TKA), craniotomy, gastric bypass, abdominal aortic aneurysm

M AN U

(AAA) repair, coronary artery bypass grafting (CABG), or cystectomy. The surgeries were determined by searching the MEDPAR (Medicare Provider Analysis and Review) principal surgery field for International Classification of Diseases, Ninth Revision, Clinical Modification (ICD9) procedure codes. All patients must have had continuous coverage at the time of the

TE D

procedure and for at least 30 days post procedure. All patients were 65 or older at the time of the procedure. Because malpractice environment metrics from Washington D.C. and U.S. territories were unavailable, beneficiaries treated in these regions were excluded (N=3,488). Hospital

EP

characteristics were obtained from the 2010 American Hospital Association (AHA) Annual Survey data. This observational study did not require review by the Northwestern University

AC C

Institutional Review Board as it involved pre-existing, de-identified data.

Outcome Measures

The outcome measures included 30-day readmission, prolonged length of stay (LOS)

(defined as being in the top quartile of length of stay), as well as several 30-day postoperative complications: sepsis, myocardial infarction (MI), pneumonia (PNA), surgical site infection 6

ACCEPTED MANUSCRIPT

(SSI), acute renal failure (ARF), respiratory failure, deep vein thrombosis (DVT) or pulmonary embolism (PE), gastrointestinal (GI) bleed, hematoma/hemorrhage, and mortality.20 The Agency for Healthcare Research and Quality (AHRQ) patient safety indicators were used as guidelines in

RI PT

determining the 30-day complications for sepsis, DVT or PE, hemorrhage or hematoma, and respiratory failure, applying exclusion criteria as necessary.21 The International Classification of Diseases, Ninth Revision (ICD9) codes used in previous studies were used to determine MI,

SC

pneumonia, ARF, SSI, and GI bleed as well as incorporating some of the standard patient safety indicator guidelines.20 For each complication the MEDPAR, inpatient claims, and outpatient

M AN U

claims were searched. A patient was considered to have the post-operative complication if it occurred either during the inpatient stay or post-discharge, within 30 days of the procedure.

Measures of State-Level Malpractice Environment

TE D

Similar to our previous work in this area,19 data on general surgery malpractice premiums were drawn from the 2010 Medical Liability Monitor Annual Rate Survey, an overview of physicians’ liability insurance rates quoted by large U.S. malpractice insurance underwriters.22

EP

Calculated mean state premiums for general surgery were used as an indicators of state-level practice costs due to liability threat in a given location.12,23,24 As a proxy for claims frequency, a

AC C

measure of paid claims per surgeons was constructed from 2010 data from the National Practitioner Data Bank (NPDB) Public Use File on paid surgery-related malpractice claims in each state.8,25

Overall malpractice environment was characterized using a composite measure.26 A

principal component analysis (PCA) of seven measures (number of paid claims per physician, average malpractice award size, lawyers per capita, general medicine/general 7

ACCEPTED MANUSCRIPT

surgery/obstetrics/gynecology premiums, and an index of malpractice laws) was used to construct this composite measure. The index of malpractice laws was drawn from the National Conference of State Legislature’s (NCSL) 2008 summary of Medical Liability/Malpractice Law.

RI PT

This summary detailed the following reforms for each state: caps on noneconomic damages, joint-and-several liability reform, limits on attorneys’ fees, mandatory periodic payment of

malpractice damages awards, patient compensation funds, laws protecting physician apologies,

SC

pretrial screening panels, alternative dispute resolution mechanisms, requirements for certificates of merit, expert witness standards, and pretrial screening panels.27 Each reform was coded as a

M AN U

dichotomous variable indicating the absence of that reform in a given state. Examples of laws that create high liability environments include: lack of caps on noneconomic damages, no limit on attorneys’ fees, and lack of pretrial screening panels, which encourage parties to settle meritorious claims and work to deter frivolous claims. A more detailed description of the

TE D

development of this composite state-level malpractice measure have been previously published.26 When characterizing state malpractice environment, we defined it from the perspective of physicians; that is, “low-risk” environments are those that are favorable to physicians. Quintiles

EP

for each continuous measure of malpractice risk (average premiums, average number of paid claims per surgeons, and the composite) were constructed. States were deemed to have the

AC C

“highest” malpractice risk environment if they were in the top quintile while those that were in the bottom quintile were designated as having the “lowest” malpractice risk environments.19

Data Analyses

To examine the association between 30-day complications and malpractice variables, hierarchical logistic regression models with hospital random-intercept were used to account for 8

ACCEPTED MANUSCRIPT

the clustered nature of the data. The variables for limits on attorney fees, caps on damages, and pretrial panels malpractice were adjusted for simultaneously due to their combined contribution to the overall malpractice environment. The odds ratios (ORs) and corresponding 95%

RI PT

confidence interval (CI) indicates the likelihood of occurrence of a given complication in states with laws that are “unfavorable” to physicians (i.e., an OR less than one indicates that having a malpractice law that is “unfavorable” to physicians is negatively associated with the

SC

complication). The continuous variables were split into quintiles and each adjusted for in a

separate models. The models adjusted for patient-level (age, race, sex), procedure (emergent),

M AN U

and hospital-level variables (number of beds, teaching status, and hospital type). The odds ratios (ORs) and corresponding 95% confidence interval (CI) indicate the odds of having the complication in states contained in the highest quintile relative to states in the lowest quintile. The Elixhauser Comorbidity scores were calculated at the index procedure and were entered into

TE D

the model as covariates, excluding the comorbidities AIDS, drug use, and ulcer due to extremely low frequency. As a sensitivity analysis, we ran a hierarchical logistic regression model simultaneously including all of the state medical malpractice tort laws summarized by the NCSL

EP

(Appendix 1). In addition, analyses were also run excluding TKAs, which comprised more than 50% of cases, to determine if findings were driven solely by complications of this one procedure.

AC C

The alpha level for significance level was set at 0.05. Significance is noted with and

without the Bonferroni correction for multiple comparisons, as correction for multiple comparisons would actually favor our hypothesis (i.e., more stringent significance threshold would make it less likely to identify any significant associations between malpractice environment and outcomes). The Bonferroni corrected significance value was p<0.0006. All

9

ACCEPTED MANUSCRIPT

reported p-values are two-tailed. All analysis was done using SAS, version 9.4 (Cary, North Carolina).

RI PT

RESULTS Patient Characteristics

Our study had 890,237 beneficiaries from 3,168 hospitals with a median age of 74 years.

SC

The majority of patients were Non-Hispanic whites (90.6%) (Table 1). More than half of the patients in the sample underwent TKA (50.4%), while 16.9% had colon surgery, 13.5%

M AN U

underwent CABG, and 5.7% underwent craniotomy. No other procedure comprised more than 5% of the population. There was a small percentage of emergent procedures (3.7%). Patients were more frequently hospitalized in larger hospitals (74.3%) and teaching hospitals (54.1%).

TE D

State Malpractice Environments

The mean annual malpractice premium for General Surgeons was $46,947 (SD=$21,421), and there was a mean of 1.2 surgery-related paid claims per 100 surgeons

EP

(SD=0.68) (Table 2). Table 2 also shows the number of states with malpractice laws “unfavorable” to physicians per the NCSL, (e.g., caps on damages (20), limits on attorneys’ fees

AC C

(33), and pretrial screening panels (34)). The composite measure identified Pennsylvania, New York, New Jersey, and Illinois among the highest-risk states, and Arkansas, Minnesota, Idaho, and North Dakota are among the lowest-risk states.19,26

State Malpractice Environment and Postoperative Outcomes

10

ACCEPTED MANUSCRIPT

Regression results for the models predicting postoperative complications as a function of state malpractice environment can be found in Table 3. Greater general surgery malpractice insurance premiums were significantly associated with a greater likelihood of postoperative

RI PT

sepsis (OR 1.26, 95% CI 1.11-1.43), pneumonia (OR 1.13, 95% CI 1.07-1.20), ARF (OR1.09, 95% CI 1.03-1.15), and GI bleed (OR 1.03-1.23). In states with higher numbers of paid

claims/100 physicians, a higher likelihood of MI (OR 1.09, 95% CI 1.01-1.17), SSI (OR 1.13,

SC

95% CI 1.07-1.20), ARF (OR 1.09, 95% CI 1.03-1.15), hemorrhage (OR 1.17, 95% CI 1.041.31), prolonged LOS (OR 1.12, 95% CI 1.07-1.17), readmission (OR 1.16, 95% CI 1.10-1.22),

M AN U

and mortality (OR 1.08, 95% CI 1.02-1.15) were noted. State malpractice award size offered a more mixed picture, with states with higher award sizes found to be associated with a higher likelihood of MI (OR 1.12, 95% CI 1.04-1.19) and DVT/PE (OR 1.08, 95% CI 1.01-1.16) but a lower likelihood of SSI (OR 0.94, 95% CI 0.89-0.99) and mortality (OR 0.94, 95% CI 0.89-

TE D

0.99). Higher-risk state malpractice environment based on the composite measure was associated with higher likelihood of sepsis (OR 1.22, 95% CI 1.07-1.39), MI (OR 1.14, 95% CI 1.06-1.23), pneumonia (OR 1.09, 95% CI 1.03-1.16), ARF (OR 1.15 95% CI 1.08-1.22), DVT/PE (OR 1.22

EP

95% CI 1.13-1.32), and gastrointestinal bleed (OR 1.18, 95% CI 1.08-1.30). The composite measure was not associated with a significantly lower likelihood of any of the postoperative

AC C

complications studied.

None of the tort reform variables included in the regression model (limits on attorneys’

fees, caps on damages, and pretrial panels) were consistently associated with an increased or decreased likelihood of 30-day complications (Table 4). Similarly, in our sensitivity analysis, which included all tort reforms listed in the NCSL, the presence of reforms were not consistently significantly associated with a change in the likelihood of the 30-day postoperative 11

ACCEPTED MANUSCRIPT

complications studied (Appendix 1). When TKAs were excluded from the analysis, associations between malpractice environment and post-operative complications were largely unchanged. DISCUSSION

RI PT

The U.S. medical malpractice system assumes that the threat of liability should deter negligence and discourage poor quality of care, but whether it leads to measurable changes in patient outcomes is unclear.7,9,10 In previous work, our group was unable to detect an association

SC

between malpractice environment and postoperative outcomes in colorectal surgery.19 In this study, with a much wider range of surgical procedures, we found that higher liability state

M AN U

malpractice environments were not consistently associated with a lower likelihood of surgical post-operative complications. In some instances, higher liability environments were actually associated with an increased likelihood of certain 30-day post-operative complications. If medical malpractice risk serves as an effective deterrent to negligent care, then

TE D

improved quality should be reflected by improved surgical outcomes in higher liability environments. Many of the studies that have examined the relationship between malpractice environment and patient outcomes have found conflicting results.12,13,15,17,18 While some authors

EP

have found evidence of deterrence,8-11,28,29 many of these studies had inconsistent overall findings10,11,29 and suffered from certain limitations, such as a lack of generalizability to other

AC C

procedures or fields8 and short observation periods for longitudinal studies.9 Our study used several measures of state medical malpractice environment to test for

associations with a number of post-operative complications after a wide range of surgical procedures in order to broadly capture the possible effects. The lack of consistent association between select measures of malpractice environment and surgical complications could reflect the clinical complexity inherent in preventing surgical complications. According to tort deterrence 12

ACCEPTED MANUSCRIPT

theory, adopting a “defensive” stance in higher-risk malpractice environments may, in some cases, result in a lower incidence of a given complication, (e.g., fear of suit may prompt higher vigilance regarding appropriate beta-blockade through surgery, thus lowering post-operative MI

RI PT

rates). In other instances, defensive practices may actually lead to the increased incidence of postoperative complications via surveillance bias, such as DVT (i.e., “the more you look, the more you find” phenomenon),30 where the concern related to being sued for a missed diagnosis

SC

of a DVT can lead to a low threshold for obtaining deep venous imaging, with subsequent

increased identification of clinically inconsequential DVTs. The bleeding risk associated with the

M AN U

subsequent treatment of these DVTs is not negligible and more harm than good may result. Similarly, complications may arise from unnecessary cardiac catheterizations performed to assuage fears of missing a treatable MI or Clostridium difficile infections may result from a defensive overuse of antibiotics to prevent SSI. Other “assurance” practices, or supplying

TE D

additional services with little or no medical value primarily out of concern of medical liability,31 may also have the unintended consequence of higher readmission rates and lengths of stay. While the precise drivers of this association with worse outcomes cannot be delineated using

EP

these administrative data, the finding does raise questions regarding the ability of the tort system to improve patient care. Prior publications have illustrated of how defensive practices can

AC C

ultimately lead to patient harm32,33 but an overarching conceptual mechanism by which litigation anxiety directly leads to worse outcomes is not readily available. The overall pattern of our results suggests that the severity of state medical malpractice

system may fail to effectively promote higher quality surgical care. Other research has speculated that the “poor fit” between those who are negligently injured and those who sue1 (i.e. most negligently injured patients do not file claims, and many patients who do file claims have 13

ACCEPTED MANUSCRIPT

not actually suffered a negligent injury),34,35 may partially explain why malpractice environments don’t result in improved quality of care. As a consequence, providers perceive that malpractice claiming is haphazard and does not help derive risk-reduction information from the outcomes of

RI PT

tort litigation.36 In fact, because malpractice claims are relatively rare events with many false positives, improvements in rates of negligent injury are not likely to lead to reductions in claims rates for the individual physician.37

SC

Limitations

Our results should be interpreted in light of several limitations. First, given the cross-

M AN U

sectional nature of this study, the directionality of the association between malpractice environment and outcomes cannot be determined; that is, we cannot determine whether patient outcomes led to the development of a state’s given liability environment or vice versa. However, these laws have generally been in place and unchanged in most states for many years prior to our

TE D

analyses, so if the deterrence theory holds true, we should still see a reduction in complications. Certain methodological approaches, such as building in a time lag between the measures of malpractice environment and the outcome measures, or using instrumental variable analysis,

EP

have been suggested as means of better defining this relationship.1 Both of these strategies, however, carry their own drawbacks.1 Second, the frequency of complications may not be

AC C

accurately coded using administrative data; however, we used complications that have been demonstrated to be reliably abstracted from claims data.20,38,39 Furthermore, under- or overcoding should not disproportionately affect states with either high or low malpractice risk, therefore limiting the effect of under-or over-coding in our analysis of difference. Third, while the use of Medicare data allowed us to perform a nationwide study on the relationship between malpractice environment and surgical complications, use of this dataset limited our population to 14

ACCEPTED MANUSCRIPT

Medicare beneficiaries, thus limiting the generalizability of results to patients with other insurance types. Fourth, as there is no gold standard for measuring malpractice environment, we chose to include a number of different measures (selected tort reforms, malpractice insurance

RI PT

premiums, number of paid claims per 100 surgeons, average malpractice award size, and a composite measure) in the same regression model. Clearer associations may have been

demonstrated in models using a single indicator of malpractice environment. This approach,

SC

however, is probably better suited to studies evaluating hypotheses relating to specific facets of malpractice, rather than evaluations of malpractice environment as a whole.26 The malpractice

M AN U

composite measure is likely the best assessment of overall state malpractice environment. Finally, our current malpractice system is designed to deter negligent care, and we cannot draw any conclusions regarding negligence from the occurrences of these postoperative complications. However, we reason that physicians, in responding to high-liability environments, may take

TE D

greater precautions with the care of their patients and adopt a higher state of vigilance against postoperative complications, which should then result in better patient outcomes, but this did not turn out to be the case.

EP

Conclusion

The lack of a consistent association between malpractice environment and improved

AC C

surgical outcomes not only questions the efficacy of our current medical malpractice system as an effective vehicle for deterrence, but also suggests a possible unintentional potential to increase harm to patients. Future studies may help to elucidate the precise mechanisms by which high-risk medical malpractice environments may increase the likelihood of surgical complications. The effect of tort reforms, particularly non-traditional reforms,40 on patient outcomes should be further studied and used to inform discussions on malpractice reform. 15

ACCEPTED MANUSCRIPT

REFERENCES 1.

Mello M, Brennan T. Deterrence of medical errors: theory and evidence for malpractice reform. Tex Law Rev 2002;80:1595-1637. Posner RA. A theory of negligence. J Legal Stud 1972;1:29-96.

3.

Jena AB, Seabury S, Lakdawalla D, Chandra A. Malpractice risk according to physician

RI PT

2.

specialty. N Engl J Med 2011;365:629-636.

Deshpande SP, Deshpande SS. Factors impacting perceived threat of malpractice lawsuits

SC

4.

by various medical specialists. Health Care Manag (Frederick) 2011;30:55-65. U.S. Congress OTA. Defensive Medicine and Medical Malpractice, OTA-H--6O2.

M AN U

5.

Washington DC: U.S. Government Printing Office; 1994. 6.

Minami CA, Chung JW, Holl JL, Bilimoria KY. Impact of medical malpractice environment on surgical quality and outcomes. J Am Coll Surg 2014;218:271-278. e279. Frakes M, Jena AB. Does Medical Malpractice Law Improve Health Care Quality? :

TE D

7.

National Bureau of Economic Research;2014. 8.

Dhankar P, Khan M, Bagga S. Effect of medical malpractice on resource use and

9.

EP

mortality of AMI patients. J Empirical Legal Stud 2007;4:163-183. Zabinski ZaB, Bernard S. The Deterrent Effect of Tort Law: Evidence from Medical

10.

AC C

Malpractice Reform. Northwestern Law & Econ Research Paper 2015 No. 13-09.

Currie J, MacLeod W. First do no harm? Tort reform and birth outcomes. Q J Econ

2008;123:795-830.

11.

Iizuka T. Does Higher Malpractice Pressure Deter Medical Errors? J Law Econ 2013:161-188.

17

ACCEPTED MANUSCRIPT

12.

Dubay L, Kaestner R, Waidmann T. The impact of malpractice fears on cesarean section rates. J Health Econ 1999;18:491-522. Kessler D, McClellan M. Do doctors practice defensive medicine? Q J Econ 1996;111:353-390.

14.

Kessler D, McClellan M. Malpractice law and health care reform: optimal liability policy in an era of managed care. J Pub Econ 2002;84:175-197.

Klick J, Stratmann T. Medical malpractice reform and physicians in high-risk specialties.

SC

15.

J Legal Stud 2007;36:S121-S142.

Sloan F, Shandle J. Is there empirical evidence for "Defensive Medicine"? A

M AN U

16.

RI PT

13.

reassessment. J Health Econ 2009;28:481-491. 17.

Sloan F, Whetten-Goldsteon K, Githens P, Entman S. Effects of the threat of medical malpractice litigation and other factors on birth outcomes. Med Care 1995;33:700-714. Yang YT, Studdert DM, Subramanian S, Mello MM. Does tort law improve the health of

TE D

18.

newborns, or miscarry? A longitudinal analysis of the effect of liability pressure on birth outcomes. J Empirical Legal Stud 2012;9:217-245. Bilimoria KY, Sohn MW, Chung JW, et al. Association between state medical

EP

19.

malpractice environment and surgical quality and cost in the United States. Ann Surg

20.

AC C

2016;263:1126-1132.

Ghaferi AA, Birkmeyer JD, Dimick JB. Complications, failure to rescue, and mortality with major inpatient surgery in medicare patients. Ann Surg 2009;250:1029-1034.

21.

AHRQ. Patient Safety Indicators Technical Specifications. http://www.qualityindicators.ahrq.gov/Modules/PSI_TechSpec.aspx. Accessed 9/1/2014,

2014. 18

ACCEPTED MANUSCRIPT

22.

Medical Liability Monitor. Medical Liability Monitor Annual Rate Survey Issue 2011;36.

23.

Rock S. Malpractice premiums and primary cesarean section rates in New York and Illinois. Public Health Rep 1988;103:459-463. Localio A, Lawthers A, Bengtson J, et al. Relationship between malpractice claims and

RI PT

24.

cesarean delivery. JAMA 1993;269:366-373. 25.

Chandra A. The growth of physician medical malpractice payments: evidence from the

26.

SC

national practitioner data bank. Health Aff (Millwood) 2005;W5:240-249.

Chung JW, Sohn MW, Merkow RP, et al. Development of a composite measure of state-

M AN U

level malpractice environment. Health Serv Res 2014;49:751-766. 27.

National conference of State Legistures. Available at www.ncsl.org. Accessed 7/7/2013.

28.

Lakdawalla DN, Seabury SA. The Welfare effects of medical malpractice liability. Int Rev Law Econ 2012;32:356-369.

Baicker K, Chandra A. The effect of malpractice liability on the delivery of health care.

TE D

29.

Health Econ Pol 2005:8(4). 30.

Bilimoria KY, Chung J, Ju MH, et al. Evaluation of surveillance bias and the validity of

31.

EP

the venous thromboembolism quality measure. JAMA 2013;310:1482-1489. Studdert D. Defensive medicine among high-risk specialist physicians in a volatile

32.

AC C

malpractice environment. JAMA 2005;293:2609-2617.

Becker MC, Galla JM, Nissen SE. Left main trunk coronary artery dissection as a

consequence of inaccurate coronary computed tomographic angiography. Arch Intern Med 2011;171:698-701.

33.

Srinivas SV, Deyo RA, Berger ZD. Application of "less is more" to low back pain. Arch Intern Med 2012;172:1016-1020. 19

ACCEPTED MANUSCRIPT

34.

Localio A, Lawthers A, Brennan T, et al. Relation between malpractice claims and adverse events due to negligence. Results of the Harvard Medical Practice Study III. N Engl J Med 1991;325:245-251. Studdert D, Mello M, Gawande A, et al. Claims, errors, and compensation payments in

RI PT

35.

medical malpractice litigation. N Engl J Med 2006;354:2024-2033.

Hupert N, Lawthers A, Brennan T, Peterson L. Processing the tort deterrent signal: a qualitative study. Soc Sci Med 1996;43(1).

Mello MM, Hemenway D. Medical malpractice as an epidemiological problem. Soc Sci Med 2004;59:39-46.

38.

M AN U

37.

SC

36.

Iezzoni LI, Daley J, Heeren T, et al. Identifying complications of care using administrative data. Med Care 1994;32:700-715.

39.

Weingart SN, Iezzoni LI, Davis RB, et al. Use of administrative data to find substandard

Mello MM, Studdert DM, Kachalia A. The medical liability climate and prospects for

EP

reform. JAMA 2014;312:2146-2155.

AC C

40.

TE D

care: validation of the complications screening program. Med Care 2000;38:796-806.

20

ACCEPTED MANUSCRIPT

Data 890,237 74 (65-108) 485,029 (54.5)

SC

806,843 (90.6) 51,250 (5.8) 10,328 (1.2) 21,816 (2.5) 150,467 (16.9) 17,682 (2.0) 448,958 (50.4) 50,720 (5.7) 8,522 (1.0) 4,119 (0.5) 36,791 (4.1) 3,759 (0.4) 39,388 (4.4) 120,447 (13.5) 9,384 (1.1)

AC C

EP

TE D

M AN U

Characteristic Patients Age, y, median (range) Women, n (%) Race (ethnicity), n (%) Non-Hispanic white Non-Hispanic black Hispanic Other Principal procedure, n (%) Colonic surgery Rectal surgery Total knee arthroplasty Craniotomy Pancreatic surgery Esophageal surgery Lung resection Gastric bypass surgery Abdominal aortic aneurysm surgery Coronary artery bypass Cystectomy Emergent procedure, n (%) No Yes Hospitals (n=3,168) No. of beds, n (%) < 100 100 to 199 200 to 499 ≥500 Hospital ownership/control, n (%) Not-for-profit, non-government Government Investor (for profit) Teaching status, n (%) Non-teaching Teaching

RI PT

Table 1. Patient and Hospital Characteristics

857,029 (96.3) 33,208 (3.7)

80,487 (9.0) 147,465 (16.6) 428,240 (48.1) 234,045 (26.3)

675,237 (75.9) 127,590 (14.3) 87,410 (9.8) 408,447 (45.9) 481,790 (54.1)

21

ACCEPTED MANUSCRIPT

Table 2. State Malpractice Environment Measures (N=50 States) Characteristic

Data 46,947 (21,421) 1.2 (0.68) 20 (40) 33 (66) 34 (68) 38 (76) 26 (52) 2 (4) 12 (24) 14 (28) 20 (40) 10 (20) 14 (28)

AC C

EP

TE D

M AN U

SC

RI PT

Annual malpractice insurance premiums for general surgeons, $, mean (SD) No. of paid, surgery-related claims per 100 surgeons, mean (SD) States with the following tort reforms “unfavorable” to physicians, n (%) No caps on damages No limits on attorneys’ fees No pre-trial screening panels No general compensation fund No affidavit/certificate of merit No peer review panels No alternative dispute resolution mechanisms No physician apology/sympathetic gestures No expert witness standards No joint and several liability No periodic payment

22

ACCEPTED MANUSCRIPT

Table 3. Associations between Measures of State Malpractice Environment and 30-Day Complications

p Value

<0.001†

1.13 1.09 1.05 (0.99,1.28) (1.01,1.17) (0.99,1.12)

0.06

0.02‡

0.08

0.003‡

0.87

1.13 1.09 (1.07,1.20) (1.03,1.15)

<0.001†

0.003‡

RI PT

SC

1.05 1.12 (0.98,1.12) (1.03,1.23)

M AN U

0.06

1.05 (0.95,1.17)

0.35

1.05 (0.94,1.16)

TE D

2010 number of paid claims/100 physicians, odds ratio (95% CI)*

<0.001†

1.00 1.09 (0.95,1.06) (1.03,1.15)

0.40

0.20

0.009‡

1.05 1.07 (0.98,1.13) (0.97,1.17)

0.17

0.16

1.01 (0.90,1.13)

1.03 (0.99,1.07)

0.98 (0.93,1.03)

1.00 (0.94,1.06)

0.85

0.20

0.41

0.99

1.17 (1.04,1.31)

1.12 (1.07,1.17)

1.16 (1.10,1.22)

1.08 (1.02,1.15)

0.008‡

<0.001†

<0.001†

0.008‡

EP

p Value

1.26 1.07 1.13 (1.11,1.43) (1.00,1.15) (1.07,1.20)

Respiratory DVT or failure, PE, n=485,441 n=793,636

AC C

State law* General surgery malpractice insurance premiums, odds ratio (95% CI)*

Sepsis, MI, Pneumonia SSI, ARF, n=448,500 n=768,698 , n=790,888 n=795,869 n=784,427

Hematoma or Prolonged GI bleed, hemorrhage, LOS, Readmission Mortality, n=793,061 n=752,957 n=890,232 , n=853,752 n=890,232

23

ACCEPTED MANUSCRIPT

0.001‡

0.02‡

0.98

Malpractice composite 1.22 1.14 1.09 variable, odds (1.07,1.39) (1.06,1.23) (1.03,1.16) ratio (95% CI)* p Value

0.004‡

<0.001†

0.11

0.98 1.15 (0.92,1.04) (1.08,1.22)

0.004†

0.52

<0.001†

0.95 (0.86,1.04)

0.27

0.99 (0.88,1.10)

1.08 1.06 (1.01,1.16) (0.98,1.15)

0.02‡

0.80

0.98 (0.94,1.02)

1.04 (1.00,1.09)

0.94 (0.89,0.99)

0.24

0.31

0.07

0.02‡

1.05 (0.93,1.18)

1.05 (1.00,1.1)

1.05 (0.99,1.11)

0.99 (0.93,1.05)

0.42

0.05

0.08

0.76

0.13

1.22 1.18 (1.13,1.32) (1.08,1.30)

<0.001†

1.06 (0.96,1.17)

RI PT

0.19

0.94 0.96 (0.89,0.99) (0.91,1.01)

SC

p Value

0.93 1.12 1.00 (0.83,1.04) (1.04,1.19) (0.95,1.05)

M AN U

2010 state average malpractice award size, odds ratio (95% CI)*

<0.001†

*“Worst” malpractice environment quintile compared to “best” malpractice environment quintile. p<0.0006 (Bonferroni-corrected level).

‡

p<0.05.

TE D

†

ARF, acute renal failure; DVT, deep vein thrombosis; GI, gastrointestinal; LOS, length of stay; MI, myocardial infarction; PE,

AC C

EP

pulmonary embolism; SSI, surgical site infection

24

ACCEPTED MANUSCRIPT

Table 4. Associations between Selected State Tort Reform Variables and 30-Day Complications

0.78

0.004†

0.94 1.08 (0.87,1.02) (1.03,1.13) 0.16

<0.001†

1.17 0.97 (1.06,1.31) (0.91,1.02) 0.003†

0.24

0.99 (0.96,1.03)

1.03 (0.99,1.06)

1.04 (1.01,1.08)

1.12 (1.05,1.20)

0.71

0.16

0.01†

<0.001‡

0.99 (0.95,1.02)

0.99 (0.95,1.02)

0.99 (0.96,1.03)

0.44

0.46

0.77

1.01 (0.97,1.06)

0.96 (0.91,1.00)

1.06 (1.02,1.11)

0.55

0.08

0.008†

RI PT

0.99 0.94 (0.92,1.07) (0.90,0.98)

ARF, n=784,427

1.01 (0.94,1.08) 0.83

1.06 (0.97,1.16) 0.17

Hematoma or Prolonged GI bleed, hemorrhage, LOS, Readmission, Mortality, n=793,061 n=752,957 n=890,232 n=853,752 n=890,232

0.99 0.99 (0.95,1.03) (0.93,1.04)

SC

MI, Pneumonia, SSI, n=768,698 n=790,888 n=795,869

DVT or PE n=793,636

0.61

M AN U

Sepsis, n=448,500

Respiratory failure, n=485,441

TE D

State malpractice variable Limits on attorney fees, odds ratio (95% CI)* p Value Caps on damages, odds ratio (95% CI)* p Value Pretrial panels, odds ratio (95% CI)* p Value

0.58

1.03 1.00 (0.98,1.08) (0.95,1.06) 0.20

0.89

0.98 1.06 (0.93,1.04) (0.98,1.14) 0.57

0.15

0.95 (0.89,1.02)

0.99 (0.96,1.02)

1.03 (0.99,1.06)

1.06 (1.02,1.1)

0.16

0.48

0.12

0.001†

1.03 (0.96,1.11)

1.03 (1.01,1.06)

1.06 (1.03,1.09)

1.03 (0.99,1.07)

0.43

0.01†

<0.001‡

0.13

0.93 (0.85,1.02)

1.01 (0.97,1.04)

1.02 (0.99,1.06)

1.09 (1.04,1.14)

0.13

0.77

0.23

<0.001†

AC C

EP

*State law coded as being “unfavorable” for physicians. † p<0.05. ‡ p<0.0006 (Bonferroni-corrected level). ARF, acute renal failure; DVT, deep vein thrombosis; GI, gastrointestinal; LOS, length of stay; MI, myocardial infarction; PE, pulmonary embolism; SSI, surgical site infection

25

ACCEPTED MANUSCRIPT

Precis In this observational study of Medicare beneficiaries undergoing a variety of surgeries, we found that higher risk malpractice environments were not consistently associated with a lower

RI PT

likelihood of surgical postoperative complications and were, in some instances, associated with

AC C

EP

TE D

M AN U

SC

worse postoperative outcomes.

26

ACCEPTED MANUSCRIPT

eTable 1. Association of Malpractice Variables and 30-Day Complications Using All Dichotomous Malpractice Variables

0.95 1.08 (0.87,1.04) (1.03,1.14) 0.24

0.002†

Pretrial panels, odds 1.10 0.93 ratio (95% (0.97,1.25) (0.87,0.99) CI) p Value Periodic payments, odds ratio (95% CI) p Value Physicians apologies/ sympathetic gestures, odds ratio (95% CI) p Value

0.14

0.04†

0.95 1.01 (0.87,1.03) (0.96,1.06)

0.28 0.98 (0.94,1.02)

0.13

0.27

1.00 1.04 (0.96,1.04) (1.00,1.08)

0.34

0.84

0.07

1.02 (0.97,1.08)

0.91 (0.86,0.97)

1.04 (0.99,1.1)

0.43

0.002†

1.05 (1.01,1.09)

0.23

0.73

0.01†

1.09 (0.98,1.2)

1.09 (1.03,1.15)

1.01 (0.97,1.06)

0.1

0.003†

0.63

0.12

0.92 1.01 (0.89,0.96) (0.98,1.05) <0.001†

0.48

1.02 1.02 (0.98,1.07) (0.97,1.06)

0.36

1.08† (1,1.16)

0.04† 1.02 (0.95,1.10)

0.49

RI PT

0.004†

1.03 1.02 (0.99,1.07) (0.98,1.06)

0.56

1.03 (0.92,1.14) 0.62

1.06 (0.99,1.14)

Hematoma or Prolonged GI bleed, hemorrhage, LOS, Readmission, Mortality, n=793,061 n=752,957 n=890,232 n=853,752 n=890,232

0.99 0.97 (0.95,1.04) (0.92,1.03)

0.78

0.34

SC

0.61

0.98 (0.94,1.02)

1.08 1.01 (1.02,1.13) (0.95,1.08)

M AN U

p Value

0.98 0.93 (0.90,1.06) (0.89,0.98)

TE D

Caps on damages, odds ratio (95% CI)

MI, Pneumonia, SSI, n=768,698 n=790,888 n=795,869

EP

p Value

Sepsis, n=448,500

AC C

State law* Limits on attorney fees, odds ratio (95% CI)

Respiratory DVT or ARF, failure, PE, n=784,427 n=485,441 n=793,636

0.004†

0.71

0.95 1.05 (0.89,1.02) (0.96,1.15) 0.16

0.27

0.95 0.97 (0.91,1.00) (0.91,1.03)

0.96 (0.89,1.03)

1.00 (0.97,1.02)

1.02 (0.98,1.05)

1.04 (1.00,1.08)

0.26

0.74

0.30

0.04†

1.04 (0.96,1.12)

1.00 (0.97,1.03)

1.05 (1.02,1.09)

1.01 (0.97,1.06)

0.38

0.78

0.002†

0.51

0.87 (0.78,0.97)

0.99 (0.95,1.04)

1.01 (0.96,1.05)

1.10 (1.03,1.16)

0.01†

0.79

0.79

0.002†

0.94 (0.87,1.02)

1.03 (1.00,1.06)

0.99 (0.96,1.02)

1.07 (1.03,1.11)

0.10

0.06

0.27

0.11

0.05

0.51

0.002†

1.01 (0.93,1.10)

1.06 (1,1.12)

1.05 (0.98,1.13)

1.10 (1.01,1.20)

1.01 (0.98,1.04)

1.03 (0.99,1.06)

1.07 (1.03,1.12)

0.76

0.04†

0.14

0.03†

0.59

0.19

0.002†

ACCEPTED MANUSCRIPT

0.94

0.93 0.92 (0.85,1.01) (0.87,0.97)

0.08

0.002†

1.09 0.98 (0.88,1.35) (0.88,1.10) 0.42

0.76

1.02 1.03 (0.93,1.13) (0.98,1.09) 0.62

0.23

1.01 0.97 (0.90,1.13) (0.91,1.04) 0.89

0.39

0.55

0.97 (0.93,1.01)

0.10 0.98 (0.90,1.08) 0.69 1.00 (0.95,1.04) 0.82

0.50

0.99 0.97 (0.95,1.03) (0.94,1.01) 0.60

0.19

1.03 0.88 (0.98,1.07) (0.85,0.92)

<0.001†

0.23

0.99 0.88 (0.90,1.08) (0.80,0.96) 0.76

1.09 (1.03,1.15) 0.001†

0.004†

1.07 0.95 (1.02,1.12) (0.91,0.99) 0.003†

0.01†

0.91 1.07 (0.86,0.96) (1.02,1.13) <0.001†

0.13 0.96 (0.89,1.04) 0.30

0.94 (0.87,1.01)

0.007†

<0.001†

0.09

0.93 (0.78,1.12) 0.46

0.99 (0.91,1.07) 0.79 1.04 (0.95,1.15) 0.37

0.88

0.88 0.91 (0.84,0.93) (0.86,0.97)

<0.001†

0.004†

0.84 0.89 (0.75,0.95) (0.76,1.03) 0.004†

0.12

1.07 1.04 (1.02,1.13) (0.97,1.11) 0.01†

0.27

1.00 1.13 (0.94,1.07) (1.04,1.22) 0.90

1.04† (1.00,1.08)

1.06 (1.01,1.11)

1.05 (1.00,1.11)

0.11

0.04†

0.01†

0.04†

1.00 (0.92,1.08)

0.98 (0.95,1.01)

0.99 (0.95,1.02)

0.97 (0.93,1.01)

0.96

0.24

0.47

0.16

1.01 (0.94,1.10)

1.00 (0.97,1.03)

0.95 (0.91,0.98)

0.99 (0.95,1.03)

0.76

0.98

0.004†

0.49

0.94 (0.78,1.13)

1.21 (1.13,1.30)

1.04 (0.96,1.12)

0.99 (0.90,1.09)

0.49

<0.001†

0.30

0.85

1.09 (1.0,1.18)

0.99 (0.96,1.02)

0.98 (0.95,1.02)

0.92 (0.88,0.96)

0.06

0.60

0.26

<0.001†

0.86 (0.77,0.95)

0.98 (0.94,1.02)

1.05 (1.00,1.1)

1.03 (0.98,1.09)

0.003†

0.29

0.04†

0.22

0.89

0.97 1.01 (0.92,1.03) (0.94,1.08) 0.34

0.93 (0.85,1.02)

RI PT

0.12

0.99 (0.95,1.03)

0.29

SC

0.93 1.00 (0.84,1.02) (0.95,1.05)

0.16

0.83 1.00 (0.78,0.88) (0.93,1.07)

M AN U

0.04†

1.07 (0.98,1.16)

TE D

Expert witness, odds ratio (95% CI) p Value Affidavit/ certificate of merit expert witness, odds ratio (95% CI) p Value Peer review panels, odds ratio (95% CI) p Value Compensation funds, odds ratio (95% CI) p Value Joint and several liability, odds ratio (95% CI) p Value

0.64

0.97 0.98 (0.93,1.02) (0.94,1.03)

EP

p Value

1.03 (0.99,1.09)

AC C

Alternative dispute resolution 1.02 0.94 mechanisms (0.93,1.14) (0.88,1.00) , odds ratio (95% CI)

0.003†

*State law coded as being “unfavorable” for physicians. † p<0.05; comparisons not included in calculation of the Bonferroni correction, so no Bonferroni correction was applied. ARF, acute renal failure; DVT, deep vein thrombosis; GI, gastrointestinal; LOS, length of stay; MI, myocardial infarction; PE, pulmonary embolism; SSI, surgical site infection