The Spine Journal 15 (2015) 1602–1608
Clinical Study
Spine surgery and malpractice liability in the United States Symeon Missios, MDa, Kimon Bekelis, MDb,* a Department of Neurosurgery, Louisiana State University Health Sciences Center, 1541 Kings Hwy, Shreveport, LA 71103, USA Department of Surgery, Section of Neurosurgery, Dartmouth-Hitchcock Medical Center, One Medical Center Dr., Lebanon, NH 03756, USA
b
Received 17 July 2014; revised 9 February 2015; accepted 20 March 2015
Abstract
BACKGROUND CONTEXT: The correlation of negative outcomes with aggressiveness of malpractice liability has been questioned in the literature. PURPOSE: The aim of this study was to investigate the association of malpractice liability with unfavorable outcomes and hospitalization charges in spine surgery. STUDY DESIGN/SETTING: This was a retrospective cohort study. PATIENT SAMPLE: The sample included a total of 709,951 patients undergoing spine surgery who were registered in the Nationwide Inpatient Sample (NIS) database from 2005 to 2010. OUTCOME MEASURES: The outcome measures were state-level mortality, length of stay (LOS), and hospitalization charges after spinal surgery. METHODS: We performed a retrospective cohort study involving patients who underwent spine surgery from 2005 to 2010 and were registered in NIS. We used data from the National Practitioner Data Bank from 2005 to 2010 to create measures of volume and size of malpractice claim payments. Their association of the latter with the outcome measures was investigated. RESULTS: During the study period, there were 707,951 patients (mean age, 54.4 years, with 49.7% females) who underwent spine surgery and were registered in NIS. In a multivariable regression model, higher number of claims per 100 physicians in a state was associated with increased hospitalization charges (b50.14; 95% confidence interval [CI], 0.13–0.14) and LOS (b50.041; 95% CI, 0.036– 0.047). On the contrary, there was no association with mortality (odds ratio [OR], 0.99; 95% CI, 0.87–1.12). Larger magnitude of awarded claims was associated with increased hospitalization charges (b50.08; 95% CI, 0.075–0.09) and LOS (b50.02; 95% CI, 0.016–0.031). On the contrary, there was no association with mortality (OR, 0.95; 95% CI, 0.82–1.11). CONCLUSIONS: In the present national study, aggressive malpractice environment was not correlated with mortality but was associated with higher hospitalization charges after spine surgery. Further research is needed to identify ways to regulate the malpractice system to address these disparities. Ó 2015 Elsevier Inc. All rights reserved.
Keywords:
Spine surgery; Malpractice; Liability; Claims; NPDB; NIS
Introduction Physicians are at constant risk of litigation throughout their professional careers [1]. Neurosurgeons and orthopedic surgeons are faced with the heaviest burden, both in size FDA device/drug status: Not applicable. Author disclosures: SM: Nothing to disclose. KB: Nothing to disclose. SM and KB contributed equally to this work and are co-primary authors. * Corresponding author. Section of Neurosurgery, DartmouthHitchcock Medical Center, One Medical Center Dr., Lebanon, NH 03756, USA. Tel.: (603) 650-5110; fax: (603) 650-4547. E-mail address: [email protected] (K. Bekelis) http://dx.doi.org/10.1016/j.spinee.2015.03.041 1529-9430/Ó 2015 Elsevier Inc. All rights reserved.
and numbers, of malpractice claims. Most of these cases involve spine surgery [2]. Despite the rising cost of malpractice for physicians and patients, there is growing concern about the effectiveness and the impact of this system [3,4]. Some argue that the current system plays a role in maintaining the quality of care [4,5]. Others point out that it fails to compensate most patients who suffer avoidable injuries and punishes many physicians for adverse events that were not caused by negligence [4,5]. Although there is evidence for the latter in some medical specialties [4], this paradoxical imbalance between liability environment and outcomes has not been investigated before in spinal surgery, which is particularly prone to litigation [2].
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
Concerns have also been raised [6] with regard to the effect of increasing liability on the practice of defensive medicine, resulting in increased hospitalization charges. Several studies have attempted to characterize malpractice claims in spine surgery [6–17]. These reports are in the context of the broader specialty (neurosurgery or orthopedics) and do not focus on spine surgery. In addition, most of these studies are based on physician surveys about defensive medicine [6,11,13,18], whereas others focus on the context and the size of malpractice claims [7,9,13,16,17]. Most of the literature involves retrospective analyses of single-institution experiences [7,9], demonstrating results with limited generalization, given their inherent selection bias. The interpretation of other multicenter studies is equally limited given their focus on specific subgroup data or a specific region of the United States [10,12,13,17]. There has been no investigation of the association of the local malpractice liability environment with unfavorable outcomes and hospitalization charges for spine surgery. The Nationwide Inpatient Sample (NIS) is a hospital discharge database that represents approximately 20% of all inpatient admissions to nonfederal hospitals in the United States [19]. It allows for the unrestricted study of the patient population in question. By combining data from the NIS, National Practitioner Data Bank (NPDB), and Area Resource File (ARF), we investigated the association of the volume and size of claims payments at the state level with mortality, length of stay (LOS), and hospitalization charges after spine surgery.
1603
Context The authors sought to correlate outcomes with malpractice environment in an analysis that considered patients treated using spinal surgery in the Nationwide Inpatient Sample (NIS). Contribution This study considered more than 700,000 patients treated with spine surgery between 2005 and 2010. The authors maintain that more litigious malpractice environments were associated with increased hospitalization charges and longer hospital lengths of stay. Implications The authors findings regarding hospital charges are not adjusted in hierarchical fashion and may be over-inflated as a result. In addition, the study itself is subject to the potential for ecological fallacy in that the authors cannot ensure the malpractice environment is responsible for the findings related in the analysis. Moreover, it is not certain that the findings related to hospital length of stay, while statistically significant given the numbers involved, are clinically meaningful. These facts should be appreciated by readers when considering the authors contentions and interpretation. —The Editors
Methods
Area Resource File
NIS database
We used the ARF, 2005 to 2010, a national county-level health information database, maintained by the US Department of Health and Human Services, to create measures of resource availability. By combining the county data and the 2010 census data, the state density of all physicians, neurosurgeons, and orthopedic surgeons was calculated.
All patients undergoing spine surgery, who were registered in the NIS database [19] (Healthcare Cost and Utilization Project, Agency for Healthcare Research and Quality, Rockville, MD, USA), between 2005 and 2010 were included in the analysis. For these years, the NIS contains discharge data regarding 100% of discharges from a stratified random sample of nonfederal hospitals in several states to approximate a representative 20% subsample of all nonfederal US hospital discharges. More information about the NIS is available at http://www. ahcpr.gov/data/hcup/nisintro.htm. National Practitioner Data Bank We used data from the NPDB from 2005 to 2010 to create measures of volume and magnitude of claims payments [20]. This database is maintained by the Health Resources and Services Administration and contains approximately 200,000 medical malpractice payments made on behalf of physicians since 1990. Despite limitations (such as the ‘‘corporate shield’’ loophole and potential underreporting), NPDB is the most representative national database on medical malpractice payments, and the size of these potential biases is limited [4].
Cohort definition To establish the cohort of patients, we used International Classification of Disease-9-Current Modification codes to identify patients in the NIS who underwent any spine surgery (03.2–03.29, 03.0, 03.01, 03.02, 03.09, 03.1, 03.4, 03.51, 03.53, 03.59, 03.6, 80.5, 80.50, 80.51, 80.52, 80.59, 81.00, 81.01, 81.02, 81.03, 81.04, 81.05, 81.06, 81.07, 81.08, 81.09, 81.3, 81.30, 81.31, 81.32, 81.33, 81.34, 81.35, 81.36, 81.37, 81.38, 81.39, 81.62, 81.63, 81.64, 84.51) between 2005 and 2010. Outcome variables The primary outcome variables were mortality, average LOS of hospitalization, and the average hospitalization
1604
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
charges after spine surgery. All outcome values were calculated at the state level. We selected charges, instead of cost, because the former is a better surrogate of defensive medicine because it is directly tied to the physician practice and not the insurance status of the patient (which plays a major role in cost). Exposure variables The association of the outcomes with the pertinent exposure variables was examined using regression analysis. Age, modified Charlson Comorbidity Index (CCI) [21,22], state density of neurosurgeons (average number of neurosurgeons per 100,000 population in a state), and state density of orthopedic surgeons (average number of orthopedic surgeons per 100,000 population in a state) were continuous variables. Gender, race (African American, Hispanic, Asian, or other, with Caucasian being the reference value), insurance (private insurance, self-pay, Medicaid, with Medicare being the reference value), and income were categorical variables. Income was defined as the median income based on zip code and was divided into quartiles, with the lowest quartile being the reference value. We used two continuous variables reflecting state-level measures of the malpractice liability environment. The one was the mean dollar value of malpractice payments (arising from both judgments and settlements) per physician in each state. The second was the mean number of malpractice claims per 100 physicians in each state. The choice of these measures was based on findings [4] that physicians respond to the number of awarded claims and to the average size of malpractice awards. The number of malpractice claims per 100 physicians in each state was used as opposed to the number of claims per physician per state as a means of rescaling. This provided odds ratios (ORs) and confidence intervals (CIs) that were easier to interpret. As part of a sensitivity analysis, the regressions were repeated using the mean number of malpractice claims per physician in each state as a dependent variable, and the relationship between the independent variables with the exposure variable did not change. The hospital characteristics, used in the analysis as categorical variables, included hospital region (Northeast, Midwest, South, West, with Northeast being the reference value), hospital location (rural, urban teaching, and urban nonteaching, with rural being the reference value), and hospital bed size (small, medium, and large, with small being the reference value). More information of the definitions of the various categories of hospital characteristics can be found at http://www.hcup-us.ahrq.gov/db/vars/nis_stratum/ nisnote.jsp. Statistical analysis Categorical variables are expressed in percentages. Normally distributed continuous variables are expressed using
mean and standard deviation, and nonnormally distributed continuous variables are expressed using median and interquartile range (IQR). When comparing demographics, continuous variables were compared using the Student t test or Mann-Whitney test, as appropriate, and categorical variables were compared using the chi-square test. Patients with missing variables were excluded from the analysis using listwise deletion. The number of excluded patients can be found in the Supplemental Tables. Examination of the distribution of values for hospitalization charges and average claims payment per physician per state revealed significant positive skewness (5.1 and 1.4, respectively). To normalize the distribution of these values and provide a better fit for the data, log transformation was performed using the natural logarithm (ln) of the values. After ln transformation, skewness improved to 0.2 for hospitalization charges and 0.6 for claims payments per physician per state. The association of mortality with the variable of interest [number of awarded claims per 100 physicians per state and ln (average claims payment per physician per state)] was examined using a multivariable logistic regression model. The assumptions for using this model were met because we had a binary outcome, and independent observations, and error terms. The relationship of mortality with the exposure variable of interest remained nonsignificant. The association of LOS with the exposure variables of interest was examined using a generalized linear regression model using gamma distribution, which provided better fit of the data. The assumptions for using this model were met with normal distribution of residuals, homoscedasticity, and linear fit of the data. The association of hospitalization charges with the variables of interest was examined using a generalized linear model after ln transformation of the charges, which significantly improved fit of the data. The covariates included in all the models were age, sex, race, Charlson Comorbidity Index, payment, income, state density of neurosurgeons, state density of orthopedic surgeons, hospital region, hospital location, and hospital bed size. We included all the potential confounder variables in the models. Our very large sample size minimizes the possibility of misspecification by including these variables. Statistical analyses were performed using SPSS, version 20 (IBM, Armonk, NY, USA), and XLSTAT, version 2013.6.02 (Addinsoft, New York, NY, USA). All probability values are the results of two-sided tests, and the level of significance was set at p!.05.
Results Patient characteristics In the selected study period, there were 709,951 patients (Figure) undergoing spine surgery (mean age, 54.4 years, with 49.7% females), who were registered in NIS. Of these, 186,629 were operated in states that belong to the
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
1605
claims per physician was associated with a 0.08% increase in average hospitalization charges after spine surgery.
Discussion
Figure. Cohort selection for the study.
highest quartile of claims’ volume and 182,870 in states that belong to the lowest quartile of claims’ volume. The relative distribution of the exposure variables is demonstrated in Table 1.
Clinical outcomes Nationwide mortality for spinal procedures in our cohort was 0.3%. The median LOS was 2.0 days (IQR, 3.0 days). The median hospitalization charges were $36,633 (IQR, $50,706). Patients in the highest quartile of claims’ volume produced higher hospitalization charges (Table 2).
Multivariable analysis In a multivariable regression analysis (Table 3, Table SDC1, SDC2, SDC3), higher number of claims per 100 physicians per state was associated with increased lntransformed hospitalization charges (b50.14; 95% CI, 0.13–0.14) and LOS (b50.041; 95% CI, 0.036–0.047). On the contrary, number of claims was not associated with mortality (OR, 0.99; 95% CI, 0.87–1.12) after spine surgery. After back transformation, one more paid claim in a state per 100 physicians was associated with a 15.03% increase in average hospitalization charges after spine surgery. Similarly (Table 3, Table SDC4, SDC5, SDC6), increased size of claims awarded per physician per state (ln transformed for better fit of data) was associated with increased ln-transformed hospitalization charges (b50.08; 95% CI, 0.075–0.09) and LOS (b50.02; 95% CI, 0.016– 0.031) but demonstrated no association with mortality (OR, 0.95; 95% CI, 0.82–1.11), after spine surgery. After back transformation, a 1% increase in the size of paid
An environment of increased malpractice liability was not associated with higher mortality but was related to rising hospitalization charges among patients undergoing spine surgery, in this large population-level study. We used the number and size of paid claims as surrogates of the state-level malpractice environment. Prior research [4] has identified these as the central determinants of the perceived threat of malpractice among physicians. This burden can be particularly heavy for spine surgeons, not only in cost of malpractice insurance but also in time, stress, and additional work [4]. Neurosurgeons and orthopedic surgeons face the most hostile liability environment among the surgical specialties [2]. In this setting, the study of the association of litigation with negative outcomes and its impact on the practice of spine surgery is of major significance. Mortality after spinal procedures, the most common reason for litigation overall, was not associated with an aggressive malpractice environment. This paradoxical imbalance between negative outcomes and litigation has been detected before in other medical specialties [4] and brings the orientation of this system into question. Several other studies have demonstrated a disconnect between negligence and litigation, without specific analysis of spine surgery [5,23,24]. The Harvard Malpractice Study [5] and the California Medical Association’s Medical Insurance Feasibility Study [25] found negligence rates ranging from 0.8% to 1%. In addition, Studdert et al. [26] have shown that nearly 40% of claims were not associated with medical errors. It appears that local factors (likely financial), rather than medical outcomes, are driving the aggressiveness of litigation. Despite the lack of association of the malpractice environment with negative outcomes in spine surgery, the former was associated with increased hospitalization charges and increased LOS. Specifically, every additional paid claim was associated with a 15% increase in average hospitalization charges after spine surgery. Several studies have identified the practice of defensive medicine, resulting in more procedures and tests ordered, which might be associated with lengthier hospitalizations, as major contributors to this phenomenon [6,11,18]. Specifically, increased charges for imaging have been associated with hostile litigation environments in other areas of medicine [4]. The impact of this system on hospitalization charges for spine surgery has not been highlighted before in the literature. In the context of the national efforts for cost containment through accountable care [27], malpractice claims can become a target for cost optimization. In view of the impact of malpractice on the economics of health care, the need to regulate the medical litigation system becomes very significant.
1606
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
Table 1 Patient and hospital characteristics
Variables
All patients
Sample size (N) 707,951 Age (mean6SD) 54.41616.18 Sex, N (%) F 349,863 (49.66) M 354,607 (50.34) Unreported data 3,481 Quartiles of median income based on zip code, N (%) 1st Quartile 151,716 (21.96) 2nd Quartile 179,974 (26.05) 3rd Quartile 184,513 (26.71) 4th Quartile 174,679 (25.28) Unreported data 17,069 Payer, N (%) Medicare 224,803 (31.83) Medicaid 40,696 (5.76) Private payer 354,251 (50.17) Self-payer 13,086 (1.85) Other 73,332 (10.38) Unreported data 1,783 Charlson Comorbidity Index (CCI), N (%) Low (0–3) 657,046 (92.81) Moderate/high ($4) 50,905 (7.19) Race, N (%) Caucasian 443,783 (82.62) African American 38,552 (7.18) Hispanic 31,098 (5.79) Asian 7,144 (1.33) Other 16,557 (3.08) Unreported cases 170,817 Region, N (%) Northeast 111,640 (15.77) Midwest 169,448 (23.93) South 276,845 (39.11) West 150,018 (21.19) Location, N (%) Rural 32,645 (4.61) Urban nonteaching 295,284 (41.71) Urban teaching 380,022 (53.68) Bed size, N (%) Small 94,755 (13.38) Medium 150,273 (21.23) Large 462,923 (65.39) Neurosurgeons per 100,000 1.7260.32 population (mean6SD) Orthopedic surgeons per 100,000 7.7561.32 population (mean6SD)
States at the 4th quartile of paid claims’ volume
States at the 3rd quartile of paid claims’ volume
States at the 2nd quartile of paid claims’ volume
States at the 1st quartile of paid claims’ volume
p Value
165,681 53.36615.87
146,474 55.02616.62
156,429 54.71616.17
239,367 54.57616.10
!.0001
82,265 (49.65) 83,409 (50.35) 7
72,779 (49.69) 73,692 (50.31) 3
77,989 (49.86) 78,434 (50.14) 6
116,830 (49.52) 119,072 (50.48) 3,465
35,346 38,008 39,026 47,720 5,581
(22.08) (23.74) (24.38) (29.81)
35,023 42,139 39,838 25,677 3,797
(24.55) (29.53) (27.92) (18.00)
29,362 41,811 46,977 35,713 2,566
(19.08) (27.17) (30.53) (23.21)
51,985 58,016 58,672 65,569 5,125
(22.19) (24.77) (25.05) (27.99)
!.0001 !.0001 !.0001 !.0001
46,066 9,809 86,273 2,817 19,983 733
(27.93) (5.95) (52.30) (1.71) (12.11)
52,210 9,071 67,153 3,050 14,823 167
(35.69) (6.20) (45.90) (2.08) (10.13)
50,554 8,562 79,851 2,826 14,479 157
(32.35) (5.48) (51.10) (1.81) (9.27)
75,973 13,254 120,974 4,393 24,047 726
(31.84) (5.55) (50.69) (1.84) (10.08)
!.0001 !.0001 !.0001 !.0001 !.0001
.239
154,712 (93.38) 10,969 (6.62)
134,801 (92.03) 11,673 (7.97)
145,286 (92.88) 11,143 (7.12)
222,247 (92.85) 17,120 (7.15)
!.0001
114,350 11,319 5,821 1,471 6,049 26,671
(82.26) (8.14) (4.19) (1.06) (4.35)
100,601 7,461 6,224 911 3,355 27,922
(84.86) (6.29) (5.25) (0.77) (2.83)
79,381 4,954 1,861 650 1,914 67,669
(89.43) (5.58) (2.10) (0.73) (2.16)
149,451 14,818 17,192 4,112 5,239 48,555
(78.32) (7.77) (9.01) (2.16) (2.75)
!.0001 !.0001 !.0001 !.0001 !.0001
89,524 28,944 37,044 10,169
(54.03) (17.47) (22.36) (6.14)
18,426 19,034 98,161 10,853
(12.58) (12.99) (67.02) (7.41)
0 75,269 23,498 57,662
(0) (48.12) (15.02) (36.86)
3,690 46,201 118,142 71,334
(1.54) (19.30) (49.36) (29.80)
!.0001 !.0001 !.0001 !.0001
6,195 (3.74) 53,930 (32.55) 105,556 (63.71)
8,723 (5.96) 69,364 (47.36) 68,387 (46.69)
7,758 (4.96) 67,193 (42.95) 81,478 (52.09)
9,969 (4.16) 104,797 (43.78) 124,601 (52.05)
!.0001 !.0001 !.0001
25,570 (15.43) 43,916 (26.51) 96,195 (58.06) 1.7560.35
13,062 (8.92) 30,083 (20.54) 103,329 (70.54) 1.6460.29
22,636 (14.47) 35,793 (22.88) 98,000 (62.65) 1.9660.32
33,487 (13.99) 40,481 (16.91) 165,399 (69.10) 1.6060.21
!.0001 !.0001 !.0001 !.0001
8.1161.50
7.2861.57
8.1660.84
7.5461.16
!.0001
SD, standard deviation; CCI, Charlson Comorbidity Index.
Table 2 Primary outcomes for patients undergoing spinal surgery in the United States
Variables
All patients
States at the 4th quartile of paid claims’ volume
Mortality, N (%) Length of stay, median (IQR) Hospitalization charges, median (IQR)
2,277 (0.32) 2.0 (3.0) 36,633 (50,706)
553 (0.33) 2.0 (3.0) 35,015 (49,106)
IQR, interquartile range.
States at the 3rd quartile of paid claims’ volume
States at the 2nd quartile of paid claims’ volume
States at the 1st quartile of paid claims’ volume
p Value
500 (0.34) 2.0 (3.0) 40,364 (53,233)
441 (0.28) 2.0 (3.0) 29,781 (42,448)
783 (0.33) 2.0 (3.0) 40,486 (55,016)
.015 !.0001 !.0001
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
1607
Table 3 Regression models for primary outcomes and variables of interest Number of paid claims per 100 physicians per state*
ln-Transformed size of paid claims per physician per statey
Variables
OR
95% CI
p Value
OR
95% CI
p Value
Mortalityz Length of stayx Ln-transformed hospitalization chargesk
0.99 0.041 0.138
0.87–1.12 0.036–0.047 0.13–0.14
.852 !.0001 !.0001
0.95 0.02 0.08
0.82–1.11 0.016–0.031 0.075–0.09
.541 !.0001 !.0001
OR, odds ratio; 95% CI, 95% confidence interval. * The odds ratio and confidence interval for this regression were large, and rescaling was applied for the exposure variable so that it reflects the number of paid claims per 100 physicians per state. y The payment amount was ln transformed to provide a better fit for the data. z Based on a logistic regression model including age, sex, income, payer, Charlson Comorbidity Index (CCI), race, hospital region, hospital location, hospital bedsize, density of neurosurgeons, and orthopedic surgeons as covariates. x Based on a linear regression model including age, sex, income, payer, CCI, race, hospital region, hospital location, hospital bedsize, density of neurosurgeons, and orthopedic surgeons as covariates. k Based on a linear regression including age, sex, income, payer, CCI, race, hospital region, hospital location, hospital bedsize, density of neurosurgeons, and orthopedic surgeons as covariates. Hospitalization charges underwent a ln transformation because this provided the best fit for our data.
On average, physicians spend nearly 11% of their 40year careers with an open, unresolved malpractice claim [28]. The burden is heavier for litigation-prone specialists, such as neurosurgeons and orthopedic surgeons. In addition to this cost for providers at a personal level, the economic burden of the current malpractice system to physicians and society is extremely heavy, especially considering the liability outcomes dissociation. The impact of this system on the quality of care is questionable in the context of practicing defensive medicine. Unnecessary tests and procedures, in response to high litigation risk, expose patients to increased radiation risk and the possibility of adverse events, while increasing hospitalization charges. Finally, there is some evidence [12] for avoidance behavior on the part of the physicians, in high malpractice settings, resulting in increased travel times for patients needing to undergo high-risk procedures. Therefore, the need for reform of the malpractice system is imperative. The present study has several limitations common to administrative databases. Residual confounding could account for some of the observed associations. In addition, some coding inaccuracies will undoubtedly occur and can affect our estimates. This is no different than other studies involving the NIS. The NIS during the years studied did not include hospitals from all states. However, the creation of the 20% random sample to be included in NIS is done in such a way by the Healthcare Cost and Utilization Project that the hospitals included are still representative of the nation. In addition, some data categories were not available for all patients. To avoid the introduction of further bias, we excluded those patients from any analysis. Additionally, we were lacking disease severity, as well as posthospitalization and long-term data, including functional outcomes on our cohort. However, because two-thirds of paid claims involve death [20], we believe that the use of mortality as a metric is representative. The NIS does not provide a breakdown of the charges, and we cannot identity if the higher charges are the result of more coding for surgical
procedures or other hospitalization aspects, such as postoperative imaging studies. Finally, we are analyzing the observed associations, but causality is very hard to establish based on ecologic data. An important limitation of the NPDB is the inability to analyze claims that do not result in payments. However, the literature supports that paid claims are the most important determinants of perceived aggressiveness of the litigation system [20]. In addition, the NPDB underestimates the number of malpractice payments because settlements paid on behalf of corporate entities instead of physicians are exempt from reporting [29]. However, this phenomenon is not expected to be unevenly distributed throughout the United States [30]. Conclusions Spine surgeons are faced with a large number of malpractice claims during their careers. The potential imbalance between malpractice liability cost and quality of care has been an issue of debate in several areas of medicine. In the present national study merging three large databases (NIS, NPDB, and ARF), aggressive malpractice environment was not correlated with mortality but was associated with higher hospitalization charges after spine surgery. In view of the impact of malpractice on the economics of health care, the need to regulate the medical litigation system becomes a priority. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.spinee.2015.03.041. References [1] Kane C. Policy research perspectives—medical liability claim frequency: a 2007-2008 snapshot of physicians. Chicago, IL: American Medical Association, 2010.
1608
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
[2] Jena AB, Seabury S, Lakdawalla D, Chandra A. Malpractice risk according to physician specialty. N Engl J Med 2011;365:629–36. [3] Mello MM, Studdert DM, Brennan TA. The new medical malpractice crisis. N Engl J Med 2003;348:2281–4. [4] Baicker K, Fisher ES, Chandra A. Malpractice liability costs and the practice of medicine in the Medicare program. Health Aff (Millwood) 2007;26:841–52. [5] Brennan TA, Leape LL, Laird NM, Hebert L, Localio AR, Lawthers AG, et al. Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. N Engl J Med 1991;324:370–6. [6] Studdert DM, Mello MM, Sage WM, DesRoches CM, Peugh J, Zapert K, et al. Defensive medicine among high-risk specialist physicians in a volatile malpractice environment. JAMA 2005;293: 2609–17. [7] Mukherjee S, Pringle C, Crocker M. A nine-year review of medicolegal claims in neurosurgery. Ann R Coll Surg Engl 2014;96: 266–70. [8] Eloy JA, Svider PF, Folbe AJ, Couldwell WT, Liu JK. Comparison of plaintiff and defendant expert witness qualification in malpractice litigation in neurological surgery. J Neurosurg 2014;120:185–90. [9] Emery E, Balossier A, Mertens P. Is the medicolegal issue avoidable in neurosurgery? A retrospective survey of a series of 115 medicolegal cases from public hospitals. World Neurosurg 2014;81:218–22. [10] Rovit RL, Simon AS, Drew J, Murali R, Robb J. Neurosurgical experience with malpractice litigation: an analysis of closed claims against neurosurgeons in New York State, 1999 through 2003. J Neurosurg 2007;106:1108–14. [11] Mello MM, Studdert DM, DesRoches CM, Peugh J, Zapert K, Brennan TA, et al. Effects of a malpractice crisis on specialist supply and patient access to care. Ann Surg 2005;242:621–8. [12] Dranove D, Gron A. Effects of the malpractice crisis on access to and incidence of high-risk procedures: evidence from Florida. Health Aff (Millwood) 2005;24:802–10. [13] Matsen FAr, Stephens L, Jette JL, Warme WJ, Posner KL. Lessons regarding the safety of orthopaedic patient care: an analysis of four hundred and sixty-four closed malpractice claims. J Bone Joint Surg Am 2013;95:e201–8. [14] Groff MW, Heller JE, Potts EA, Mummaneni PV, Shaffrey CI, Smith JS. A survey-based study of wrong-level lumbar spine surgery: the scope of the problem and current practices in place to help avoid these errors. World Neurosurg 2013;79:585–92. [15] Daniels EW, Gordon Z, French K, Ahn UM, Ahn NU. Review of medicolegal cases for cauda equina syndrome: what factors lead to an adverse outcome for the provider? Orthopedics 2012;35:e414–9.
[16] Quraishi NA, Hammett TC, Todd DB, Bhutta MA, Kapoor V. Malpractice litigation and the spine: the NHS perspective on 235 successful claims in England. Eur Spine J 2012;21:S196–9. [17] Epstein NE. A medico-legal review of cases involving quadriplegia following cervical spine surgery: is there an argument for a nofault compensation system? Surg Neurol Int 2010;1:3. [18] Nahed BV, Babu MA, Smith TR, Heary RF. Malpractice liability and defensive medicine: a national survey of neurosurgeons. PLoS One 2012;7:e39237. [19] Steiner C, Elixhauser A, Schnaier J. The healthcare cost and utilization project: an overview. Eff Clin Pract 2002;5:143–51. [20] Chandra A, Nundy S, Seabury SA. The growth of physician medical malpractice payments: evidence from the National Practitioner Data Bank. Health Aff (Millwood) 2005. Suppl Web Exclusives:W5-240-W5-9. [21] Charlson ME, Ales KL, Simon R, MacKenzie CR. Why predictive indexes perform less well in validation studies. Is it magic or methods? Arch Intern Med 1987;147:2155–61. [22] Romano PS, Roos LL, Jollis JG. Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: differing perspectives. J Clin Epidemiol 1993;46:1075–90. [23] Brennan TA, Sox CM, Burstin HR. Relation between negligent adverse events and the outcomes of medical-malpractice litigation. N Engl J Med 1996;335:1963–7. [24] Localio AR, Lawthers AG, Brennan TA, Laird NM, Hebert LE, Peterson LM, 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;325:245–51. [25] California Medical Association. Report on the medical insurance feasibility study. San Francisco, CA: California Medical Association, 1977. [26] Studdert DM, Mello MM, Gawande AA, Gandhi TK, Kachalia A, Yoon C, et al. Claims, errors, and compensation payments in medical malpractice litigation. N Engl J Med 2006;354:2024–33. [27] Fisher ES, McClellan MB, Safran DG. Building the path to accountable care. N Engl J Med 2011;365:2445–7. [28] Seabury SA, Chandra A, Lakdawalla DN, Jena AB. On average, physicians spend nearly 11 percent of their 40-year careers with an open, unresolved malpractice claim. Health Aff (Millwood) 2013;32:111–9. [29] General Accounting Office. Major improvements are needed to enhance data bank’s reliability. Washington, DC: General Accounting Office, 2000. [30] Bishop TF, Ryan AM, Casalino LP. Paid malpractice claims for adverse events in inpatient and outpatient settings. JAMA 2011;305: 2427–31.
Clinical Study
Spine surgery and malpractice liability in the United States Symeon Missios, MDa, Kimon Bekelis, MDb,* a Department of Neurosurgery, Louisiana State University Health Sciences Center, 1541 Kings Hwy, Shreveport, LA 71103, USA Department of Surgery, Section of Neurosurgery, Dartmouth-Hitchcock Medical Center, One Medical Center Dr., Lebanon, NH 03756, USA
b
Received 17 July 2014; revised 9 February 2015; accepted 20 March 2015
Abstract
BACKGROUND CONTEXT: The correlation of negative outcomes with aggressiveness of malpractice liability has been questioned in the literature. PURPOSE: The aim of this study was to investigate the association of malpractice liability with unfavorable outcomes and hospitalization charges in spine surgery. STUDY DESIGN/SETTING: This was a retrospective cohort study. PATIENT SAMPLE: The sample included a total of 709,951 patients undergoing spine surgery who were registered in the Nationwide Inpatient Sample (NIS) database from 2005 to 2010. OUTCOME MEASURES: The outcome measures were state-level mortality, length of stay (LOS), and hospitalization charges after spinal surgery. METHODS: We performed a retrospective cohort study involving patients who underwent spine surgery from 2005 to 2010 and were registered in NIS. We used data from the National Practitioner Data Bank from 2005 to 2010 to create measures of volume and size of malpractice claim payments. Their association of the latter with the outcome measures was investigated. RESULTS: During the study period, there were 707,951 patients (mean age, 54.4 years, with 49.7% females) who underwent spine surgery and were registered in NIS. In a multivariable regression model, higher number of claims per 100 physicians in a state was associated with increased hospitalization charges (b50.14; 95% confidence interval [CI], 0.13–0.14) and LOS (b50.041; 95% CI, 0.036– 0.047). On the contrary, there was no association with mortality (odds ratio [OR], 0.99; 95% CI, 0.87–1.12). Larger magnitude of awarded claims was associated with increased hospitalization charges (b50.08; 95% CI, 0.075–0.09) and LOS (b50.02; 95% CI, 0.016–0.031). On the contrary, there was no association with mortality (OR, 0.95; 95% CI, 0.82–1.11). CONCLUSIONS: In the present national study, aggressive malpractice environment was not correlated with mortality but was associated with higher hospitalization charges after spine surgery. Further research is needed to identify ways to regulate the malpractice system to address these disparities. Ó 2015 Elsevier Inc. All rights reserved.
Keywords:
Spine surgery; Malpractice; Liability; Claims; NPDB; NIS
Introduction Physicians are at constant risk of litigation throughout their professional careers [1]. Neurosurgeons and orthopedic surgeons are faced with the heaviest burden, both in size FDA device/drug status: Not applicable. Author disclosures: SM: Nothing to disclose. KB: Nothing to disclose. SM and KB contributed equally to this work and are co-primary authors. * Corresponding author. Section of Neurosurgery, DartmouthHitchcock Medical Center, One Medical Center Dr., Lebanon, NH 03756, USA. Tel.: (603) 650-5110; fax: (603) 650-4547. E-mail address: [email protected] (K. Bekelis) http://dx.doi.org/10.1016/j.spinee.2015.03.041 1529-9430/Ó 2015 Elsevier Inc. All rights reserved.
and numbers, of malpractice claims. Most of these cases involve spine surgery [2]. Despite the rising cost of malpractice for physicians and patients, there is growing concern about the effectiveness and the impact of this system [3,4]. Some argue that the current system plays a role in maintaining the quality of care [4,5]. Others point out that it fails to compensate most patients who suffer avoidable injuries and punishes many physicians for adverse events that were not caused by negligence [4,5]. Although there is evidence for the latter in some medical specialties [4], this paradoxical imbalance between liability environment and outcomes has not been investigated before in spinal surgery, which is particularly prone to litigation [2].
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
Concerns have also been raised [6] with regard to the effect of increasing liability on the practice of defensive medicine, resulting in increased hospitalization charges. Several studies have attempted to characterize malpractice claims in spine surgery [6–17]. These reports are in the context of the broader specialty (neurosurgery or orthopedics) and do not focus on spine surgery. In addition, most of these studies are based on physician surveys about defensive medicine [6,11,13,18], whereas others focus on the context and the size of malpractice claims [7,9,13,16,17]. Most of the literature involves retrospective analyses of single-institution experiences [7,9], demonstrating results with limited generalization, given their inherent selection bias. The interpretation of other multicenter studies is equally limited given their focus on specific subgroup data or a specific region of the United States [10,12,13,17]. There has been no investigation of the association of the local malpractice liability environment with unfavorable outcomes and hospitalization charges for spine surgery. The Nationwide Inpatient Sample (NIS) is a hospital discharge database that represents approximately 20% of all inpatient admissions to nonfederal hospitals in the United States [19]. It allows for the unrestricted study of the patient population in question. By combining data from the NIS, National Practitioner Data Bank (NPDB), and Area Resource File (ARF), we investigated the association of the volume and size of claims payments at the state level with mortality, length of stay (LOS), and hospitalization charges after spine surgery.
1603
Context The authors sought to correlate outcomes with malpractice environment in an analysis that considered patients treated using spinal surgery in the Nationwide Inpatient Sample (NIS). Contribution This study considered more than 700,000 patients treated with spine surgery between 2005 and 2010. The authors maintain that more litigious malpractice environments were associated with increased hospitalization charges and longer hospital lengths of stay. Implications The authors findings regarding hospital charges are not adjusted in hierarchical fashion and may be over-inflated as a result. In addition, the study itself is subject to the potential for ecological fallacy in that the authors cannot ensure the malpractice environment is responsible for the findings related in the analysis. Moreover, it is not certain that the findings related to hospital length of stay, while statistically significant given the numbers involved, are clinically meaningful. These facts should be appreciated by readers when considering the authors contentions and interpretation. —The Editors
Methods
Area Resource File
NIS database
We used the ARF, 2005 to 2010, a national county-level health information database, maintained by the US Department of Health and Human Services, to create measures of resource availability. By combining the county data and the 2010 census data, the state density of all physicians, neurosurgeons, and orthopedic surgeons was calculated.
All patients undergoing spine surgery, who were registered in the NIS database [19] (Healthcare Cost and Utilization Project, Agency for Healthcare Research and Quality, Rockville, MD, USA), between 2005 and 2010 were included in the analysis. For these years, the NIS contains discharge data regarding 100% of discharges from a stratified random sample of nonfederal hospitals in several states to approximate a representative 20% subsample of all nonfederal US hospital discharges. More information about the NIS is available at http://www. ahcpr.gov/data/hcup/nisintro.htm. National Practitioner Data Bank We used data from the NPDB from 2005 to 2010 to create measures of volume and magnitude of claims payments [20]. This database is maintained by the Health Resources and Services Administration and contains approximately 200,000 medical malpractice payments made on behalf of physicians since 1990. Despite limitations (such as the ‘‘corporate shield’’ loophole and potential underreporting), NPDB is the most representative national database on medical malpractice payments, and the size of these potential biases is limited [4].
Cohort definition To establish the cohort of patients, we used International Classification of Disease-9-Current Modification codes to identify patients in the NIS who underwent any spine surgery (03.2–03.29, 03.0, 03.01, 03.02, 03.09, 03.1, 03.4, 03.51, 03.53, 03.59, 03.6, 80.5, 80.50, 80.51, 80.52, 80.59, 81.00, 81.01, 81.02, 81.03, 81.04, 81.05, 81.06, 81.07, 81.08, 81.09, 81.3, 81.30, 81.31, 81.32, 81.33, 81.34, 81.35, 81.36, 81.37, 81.38, 81.39, 81.62, 81.63, 81.64, 84.51) between 2005 and 2010. Outcome variables The primary outcome variables were mortality, average LOS of hospitalization, and the average hospitalization
1604
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
charges after spine surgery. All outcome values were calculated at the state level. We selected charges, instead of cost, because the former is a better surrogate of defensive medicine because it is directly tied to the physician practice and not the insurance status of the patient (which plays a major role in cost). Exposure variables The association of the outcomes with the pertinent exposure variables was examined using regression analysis. Age, modified Charlson Comorbidity Index (CCI) [21,22], state density of neurosurgeons (average number of neurosurgeons per 100,000 population in a state), and state density of orthopedic surgeons (average number of orthopedic surgeons per 100,000 population in a state) were continuous variables. Gender, race (African American, Hispanic, Asian, or other, with Caucasian being the reference value), insurance (private insurance, self-pay, Medicaid, with Medicare being the reference value), and income were categorical variables. Income was defined as the median income based on zip code and was divided into quartiles, with the lowest quartile being the reference value. We used two continuous variables reflecting state-level measures of the malpractice liability environment. The one was the mean dollar value of malpractice payments (arising from both judgments and settlements) per physician in each state. The second was the mean number of malpractice claims per 100 physicians in each state. The choice of these measures was based on findings [4] that physicians respond to the number of awarded claims and to the average size of malpractice awards. The number of malpractice claims per 100 physicians in each state was used as opposed to the number of claims per physician per state as a means of rescaling. This provided odds ratios (ORs) and confidence intervals (CIs) that were easier to interpret. As part of a sensitivity analysis, the regressions were repeated using the mean number of malpractice claims per physician in each state as a dependent variable, and the relationship between the independent variables with the exposure variable did not change. The hospital characteristics, used in the analysis as categorical variables, included hospital region (Northeast, Midwest, South, West, with Northeast being the reference value), hospital location (rural, urban teaching, and urban nonteaching, with rural being the reference value), and hospital bed size (small, medium, and large, with small being the reference value). More information of the definitions of the various categories of hospital characteristics can be found at http://www.hcup-us.ahrq.gov/db/vars/nis_stratum/ nisnote.jsp. Statistical analysis Categorical variables are expressed in percentages. Normally distributed continuous variables are expressed using
mean and standard deviation, and nonnormally distributed continuous variables are expressed using median and interquartile range (IQR). When comparing demographics, continuous variables were compared using the Student t test or Mann-Whitney test, as appropriate, and categorical variables were compared using the chi-square test. Patients with missing variables were excluded from the analysis using listwise deletion. The number of excluded patients can be found in the Supplemental Tables. Examination of the distribution of values for hospitalization charges and average claims payment per physician per state revealed significant positive skewness (5.1 and 1.4, respectively). To normalize the distribution of these values and provide a better fit for the data, log transformation was performed using the natural logarithm (ln) of the values. After ln transformation, skewness improved to 0.2 for hospitalization charges and 0.6 for claims payments per physician per state. The association of mortality with the variable of interest [number of awarded claims per 100 physicians per state and ln (average claims payment per physician per state)] was examined using a multivariable logistic regression model. The assumptions for using this model were met because we had a binary outcome, and independent observations, and error terms. The relationship of mortality with the exposure variable of interest remained nonsignificant. The association of LOS with the exposure variables of interest was examined using a generalized linear regression model using gamma distribution, which provided better fit of the data. The assumptions for using this model were met with normal distribution of residuals, homoscedasticity, and linear fit of the data. The association of hospitalization charges with the variables of interest was examined using a generalized linear model after ln transformation of the charges, which significantly improved fit of the data. The covariates included in all the models were age, sex, race, Charlson Comorbidity Index, payment, income, state density of neurosurgeons, state density of orthopedic surgeons, hospital region, hospital location, and hospital bed size. We included all the potential confounder variables in the models. Our very large sample size minimizes the possibility of misspecification by including these variables. Statistical analyses were performed using SPSS, version 20 (IBM, Armonk, NY, USA), and XLSTAT, version 2013.6.02 (Addinsoft, New York, NY, USA). All probability values are the results of two-sided tests, and the level of significance was set at p!.05.
Results Patient characteristics In the selected study period, there were 709,951 patients (Figure) undergoing spine surgery (mean age, 54.4 years, with 49.7% females), who were registered in NIS. Of these, 186,629 were operated in states that belong to the
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
1605
claims per physician was associated with a 0.08% increase in average hospitalization charges after spine surgery.
Discussion
Figure. Cohort selection for the study.
highest quartile of claims’ volume and 182,870 in states that belong to the lowest quartile of claims’ volume. The relative distribution of the exposure variables is demonstrated in Table 1.
Clinical outcomes Nationwide mortality for spinal procedures in our cohort was 0.3%. The median LOS was 2.0 days (IQR, 3.0 days). The median hospitalization charges were $36,633 (IQR, $50,706). Patients in the highest quartile of claims’ volume produced higher hospitalization charges (Table 2).
Multivariable analysis In a multivariable regression analysis (Table 3, Table SDC1, SDC2, SDC3), higher number of claims per 100 physicians per state was associated with increased lntransformed hospitalization charges (b50.14; 95% CI, 0.13–0.14) and LOS (b50.041; 95% CI, 0.036–0.047). On the contrary, number of claims was not associated with mortality (OR, 0.99; 95% CI, 0.87–1.12) after spine surgery. After back transformation, one more paid claim in a state per 100 physicians was associated with a 15.03% increase in average hospitalization charges after spine surgery. Similarly (Table 3, Table SDC4, SDC5, SDC6), increased size of claims awarded per physician per state (ln transformed for better fit of data) was associated with increased ln-transformed hospitalization charges (b50.08; 95% CI, 0.075–0.09) and LOS (b50.02; 95% CI, 0.016– 0.031) but demonstrated no association with mortality (OR, 0.95; 95% CI, 0.82–1.11), after spine surgery. After back transformation, a 1% increase in the size of paid
An environment of increased malpractice liability was not associated with higher mortality but was related to rising hospitalization charges among patients undergoing spine surgery, in this large population-level study. We used the number and size of paid claims as surrogates of the state-level malpractice environment. Prior research [4] has identified these as the central determinants of the perceived threat of malpractice among physicians. This burden can be particularly heavy for spine surgeons, not only in cost of malpractice insurance but also in time, stress, and additional work [4]. Neurosurgeons and orthopedic surgeons face the most hostile liability environment among the surgical specialties [2]. In this setting, the study of the association of litigation with negative outcomes and its impact on the practice of spine surgery is of major significance. Mortality after spinal procedures, the most common reason for litigation overall, was not associated with an aggressive malpractice environment. This paradoxical imbalance between negative outcomes and litigation has been detected before in other medical specialties [4] and brings the orientation of this system into question. Several other studies have demonstrated a disconnect between negligence and litigation, without specific analysis of spine surgery [5,23,24]. The Harvard Malpractice Study [5] and the California Medical Association’s Medical Insurance Feasibility Study [25] found negligence rates ranging from 0.8% to 1%. In addition, Studdert et al. [26] have shown that nearly 40% of claims were not associated with medical errors. It appears that local factors (likely financial), rather than medical outcomes, are driving the aggressiveness of litigation. Despite the lack of association of the malpractice environment with negative outcomes in spine surgery, the former was associated with increased hospitalization charges and increased LOS. Specifically, every additional paid claim was associated with a 15% increase in average hospitalization charges after spine surgery. Several studies have identified the practice of defensive medicine, resulting in more procedures and tests ordered, which might be associated with lengthier hospitalizations, as major contributors to this phenomenon [6,11,18]. Specifically, increased charges for imaging have been associated with hostile litigation environments in other areas of medicine [4]. The impact of this system on hospitalization charges for spine surgery has not been highlighted before in the literature. In the context of the national efforts for cost containment through accountable care [27], malpractice claims can become a target for cost optimization. In view of the impact of malpractice on the economics of health care, the need to regulate the medical litigation system becomes very significant.
1606
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
Table 1 Patient and hospital characteristics
Variables
All patients
Sample size (N) 707,951 Age (mean6SD) 54.41616.18 Sex, N (%) F 349,863 (49.66) M 354,607 (50.34) Unreported data 3,481 Quartiles of median income based on zip code, N (%) 1st Quartile 151,716 (21.96) 2nd Quartile 179,974 (26.05) 3rd Quartile 184,513 (26.71) 4th Quartile 174,679 (25.28) Unreported data 17,069 Payer, N (%) Medicare 224,803 (31.83) Medicaid 40,696 (5.76) Private payer 354,251 (50.17) Self-payer 13,086 (1.85) Other 73,332 (10.38) Unreported data 1,783 Charlson Comorbidity Index (CCI), N (%) Low (0–3) 657,046 (92.81) Moderate/high ($4) 50,905 (7.19) Race, N (%) Caucasian 443,783 (82.62) African American 38,552 (7.18) Hispanic 31,098 (5.79) Asian 7,144 (1.33) Other 16,557 (3.08) Unreported cases 170,817 Region, N (%) Northeast 111,640 (15.77) Midwest 169,448 (23.93) South 276,845 (39.11) West 150,018 (21.19) Location, N (%) Rural 32,645 (4.61) Urban nonteaching 295,284 (41.71) Urban teaching 380,022 (53.68) Bed size, N (%) Small 94,755 (13.38) Medium 150,273 (21.23) Large 462,923 (65.39) Neurosurgeons per 100,000 1.7260.32 population (mean6SD) Orthopedic surgeons per 100,000 7.7561.32 population (mean6SD)
States at the 4th quartile of paid claims’ volume
States at the 3rd quartile of paid claims’ volume
States at the 2nd quartile of paid claims’ volume
States at the 1st quartile of paid claims’ volume
p Value
165,681 53.36615.87
146,474 55.02616.62
156,429 54.71616.17
239,367 54.57616.10
!.0001
82,265 (49.65) 83,409 (50.35) 7
72,779 (49.69) 73,692 (50.31) 3
77,989 (49.86) 78,434 (50.14) 6
116,830 (49.52) 119,072 (50.48) 3,465
35,346 38,008 39,026 47,720 5,581
(22.08) (23.74) (24.38) (29.81)
35,023 42,139 39,838 25,677 3,797
(24.55) (29.53) (27.92) (18.00)
29,362 41,811 46,977 35,713 2,566
(19.08) (27.17) (30.53) (23.21)
51,985 58,016 58,672 65,569 5,125
(22.19) (24.77) (25.05) (27.99)
!.0001 !.0001 !.0001 !.0001
46,066 9,809 86,273 2,817 19,983 733
(27.93) (5.95) (52.30) (1.71) (12.11)
52,210 9,071 67,153 3,050 14,823 167
(35.69) (6.20) (45.90) (2.08) (10.13)
50,554 8,562 79,851 2,826 14,479 157
(32.35) (5.48) (51.10) (1.81) (9.27)
75,973 13,254 120,974 4,393 24,047 726
(31.84) (5.55) (50.69) (1.84) (10.08)
!.0001 !.0001 !.0001 !.0001 !.0001
.239
154,712 (93.38) 10,969 (6.62)
134,801 (92.03) 11,673 (7.97)
145,286 (92.88) 11,143 (7.12)
222,247 (92.85) 17,120 (7.15)
!.0001
114,350 11,319 5,821 1,471 6,049 26,671
(82.26) (8.14) (4.19) (1.06) (4.35)
100,601 7,461 6,224 911 3,355 27,922
(84.86) (6.29) (5.25) (0.77) (2.83)
79,381 4,954 1,861 650 1,914 67,669
(89.43) (5.58) (2.10) (0.73) (2.16)
149,451 14,818 17,192 4,112 5,239 48,555
(78.32) (7.77) (9.01) (2.16) (2.75)
!.0001 !.0001 !.0001 !.0001 !.0001
89,524 28,944 37,044 10,169
(54.03) (17.47) (22.36) (6.14)
18,426 19,034 98,161 10,853
(12.58) (12.99) (67.02) (7.41)
0 75,269 23,498 57,662
(0) (48.12) (15.02) (36.86)
3,690 46,201 118,142 71,334
(1.54) (19.30) (49.36) (29.80)
!.0001 !.0001 !.0001 !.0001
6,195 (3.74) 53,930 (32.55) 105,556 (63.71)
8,723 (5.96) 69,364 (47.36) 68,387 (46.69)
7,758 (4.96) 67,193 (42.95) 81,478 (52.09)
9,969 (4.16) 104,797 (43.78) 124,601 (52.05)
!.0001 !.0001 !.0001
25,570 (15.43) 43,916 (26.51) 96,195 (58.06) 1.7560.35
13,062 (8.92) 30,083 (20.54) 103,329 (70.54) 1.6460.29
22,636 (14.47) 35,793 (22.88) 98,000 (62.65) 1.9660.32
33,487 (13.99) 40,481 (16.91) 165,399 (69.10) 1.6060.21
!.0001 !.0001 !.0001 !.0001
8.1161.50
7.2861.57
8.1660.84
7.5461.16
!.0001
SD, standard deviation; CCI, Charlson Comorbidity Index.
Table 2 Primary outcomes for patients undergoing spinal surgery in the United States
Variables
All patients
States at the 4th quartile of paid claims’ volume
Mortality, N (%) Length of stay, median (IQR) Hospitalization charges, median (IQR)
2,277 (0.32) 2.0 (3.0) 36,633 (50,706)
553 (0.33) 2.0 (3.0) 35,015 (49,106)
IQR, interquartile range.
States at the 3rd quartile of paid claims’ volume
States at the 2nd quartile of paid claims’ volume
States at the 1st quartile of paid claims’ volume
p Value
500 (0.34) 2.0 (3.0) 40,364 (53,233)
441 (0.28) 2.0 (3.0) 29,781 (42,448)
783 (0.33) 2.0 (3.0) 40,486 (55,016)
.015 !.0001 !.0001
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
1607
Table 3 Regression models for primary outcomes and variables of interest Number of paid claims per 100 physicians per state*
ln-Transformed size of paid claims per physician per statey
Variables
OR
95% CI
p Value
OR
95% CI
p Value
Mortalityz Length of stayx Ln-transformed hospitalization chargesk
0.99 0.041 0.138
0.87–1.12 0.036–0.047 0.13–0.14
.852 !.0001 !.0001
0.95 0.02 0.08
0.82–1.11 0.016–0.031 0.075–0.09
.541 !.0001 !.0001
OR, odds ratio; 95% CI, 95% confidence interval. * The odds ratio and confidence interval for this regression were large, and rescaling was applied for the exposure variable so that it reflects the number of paid claims per 100 physicians per state. y The payment amount was ln transformed to provide a better fit for the data. z Based on a logistic regression model including age, sex, income, payer, Charlson Comorbidity Index (CCI), race, hospital region, hospital location, hospital bedsize, density of neurosurgeons, and orthopedic surgeons as covariates. x Based on a linear regression model including age, sex, income, payer, CCI, race, hospital region, hospital location, hospital bedsize, density of neurosurgeons, and orthopedic surgeons as covariates. k Based on a linear regression including age, sex, income, payer, CCI, race, hospital region, hospital location, hospital bedsize, density of neurosurgeons, and orthopedic surgeons as covariates. Hospitalization charges underwent a ln transformation because this provided the best fit for our data.
On average, physicians spend nearly 11% of their 40year careers with an open, unresolved malpractice claim [28]. The burden is heavier for litigation-prone specialists, such as neurosurgeons and orthopedic surgeons. In addition to this cost for providers at a personal level, the economic burden of the current malpractice system to physicians and society is extremely heavy, especially considering the liability outcomes dissociation. The impact of this system on the quality of care is questionable in the context of practicing defensive medicine. Unnecessary tests and procedures, in response to high litigation risk, expose patients to increased radiation risk and the possibility of adverse events, while increasing hospitalization charges. Finally, there is some evidence [12] for avoidance behavior on the part of the physicians, in high malpractice settings, resulting in increased travel times for patients needing to undergo high-risk procedures. Therefore, the need for reform of the malpractice system is imperative. The present study has several limitations common to administrative databases. Residual confounding could account for some of the observed associations. In addition, some coding inaccuracies will undoubtedly occur and can affect our estimates. This is no different than other studies involving the NIS. The NIS during the years studied did not include hospitals from all states. However, the creation of the 20% random sample to be included in NIS is done in such a way by the Healthcare Cost and Utilization Project that the hospitals included are still representative of the nation. In addition, some data categories were not available for all patients. To avoid the introduction of further bias, we excluded those patients from any analysis. Additionally, we were lacking disease severity, as well as posthospitalization and long-term data, including functional outcomes on our cohort. However, because two-thirds of paid claims involve death [20], we believe that the use of mortality as a metric is representative. The NIS does not provide a breakdown of the charges, and we cannot identity if the higher charges are the result of more coding for surgical
procedures or other hospitalization aspects, such as postoperative imaging studies. Finally, we are analyzing the observed associations, but causality is very hard to establish based on ecologic data. An important limitation of the NPDB is the inability to analyze claims that do not result in payments. However, the literature supports that paid claims are the most important determinants of perceived aggressiveness of the litigation system [20]. In addition, the NPDB underestimates the number of malpractice payments because settlements paid on behalf of corporate entities instead of physicians are exempt from reporting [29]. However, this phenomenon is not expected to be unevenly distributed throughout the United States [30]. Conclusions Spine surgeons are faced with a large number of malpractice claims during their careers. The potential imbalance between malpractice liability cost and quality of care has been an issue of debate in several areas of medicine. In the present national study merging three large databases (NIS, NPDB, and ARF), aggressive malpractice environment was not correlated with mortality but was associated with higher hospitalization charges after spine surgery. In view of the impact of malpractice on the economics of health care, the need to regulate the medical litigation system becomes a priority. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.spinee.2015.03.041. References [1] Kane C. Policy research perspectives—medical liability claim frequency: a 2007-2008 snapshot of physicians. Chicago, IL: American Medical Association, 2010.
1608
S. Missios and K. Bekelis / The Spine Journal 15 (2015) 1602–1608
[2] Jena AB, Seabury S, Lakdawalla D, Chandra A. Malpractice risk according to physician specialty. N Engl J Med 2011;365:629–36. [3] Mello MM, Studdert DM, Brennan TA. The new medical malpractice crisis. N Engl J Med 2003;348:2281–4. [4] Baicker K, Fisher ES, Chandra A. Malpractice liability costs and the practice of medicine in the Medicare program. Health Aff (Millwood) 2007;26:841–52. [5] Brennan TA, Leape LL, Laird NM, Hebert L, Localio AR, Lawthers AG, et al. Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. N Engl J Med 1991;324:370–6. [6] Studdert DM, Mello MM, Sage WM, DesRoches CM, Peugh J, Zapert K, et al. Defensive medicine among high-risk specialist physicians in a volatile malpractice environment. JAMA 2005;293: 2609–17. [7] Mukherjee S, Pringle C, Crocker M. A nine-year review of medicolegal claims in neurosurgery. Ann R Coll Surg Engl 2014;96: 266–70. [8] Eloy JA, Svider PF, Folbe AJ, Couldwell WT, Liu JK. Comparison of plaintiff and defendant expert witness qualification in malpractice litigation in neurological surgery. J Neurosurg 2014;120:185–90. [9] Emery E, Balossier A, Mertens P. Is the medicolegal issue avoidable in neurosurgery? A retrospective survey of a series of 115 medicolegal cases from public hospitals. World Neurosurg 2014;81:218–22. [10] Rovit RL, Simon AS, Drew J, Murali R, Robb J. Neurosurgical experience with malpractice litigation: an analysis of closed claims against neurosurgeons in New York State, 1999 through 2003. J Neurosurg 2007;106:1108–14. [11] Mello MM, Studdert DM, DesRoches CM, Peugh J, Zapert K, Brennan TA, et al. Effects of a malpractice crisis on specialist supply and patient access to care. Ann Surg 2005;242:621–8. [12] Dranove D, Gron A. Effects of the malpractice crisis on access to and incidence of high-risk procedures: evidence from Florida. Health Aff (Millwood) 2005;24:802–10. [13] Matsen FAr, Stephens L, Jette JL, Warme WJ, Posner KL. Lessons regarding the safety of orthopaedic patient care: an analysis of four hundred and sixty-four closed malpractice claims. J Bone Joint Surg Am 2013;95:e201–8. [14] Groff MW, Heller JE, Potts EA, Mummaneni PV, Shaffrey CI, Smith JS. A survey-based study of wrong-level lumbar spine surgery: the scope of the problem and current practices in place to help avoid these errors. World Neurosurg 2013;79:585–92. [15] Daniels EW, Gordon Z, French K, Ahn UM, Ahn NU. Review of medicolegal cases for cauda equina syndrome: what factors lead to an adverse outcome for the provider? Orthopedics 2012;35:e414–9.
[16] Quraishi NA, Hammett TC, Todd DB, Bhutta MA, Kapoor V. Malpractice litigation and the spine: the NHS perspective on 235 successful claims in England. Eur Spine J 2012;21:S196–9. [17] Epstein NE. A medico-legal review of cases involving quadriplegia following cervical spine surgery: is there an argument for a nofault compensation system? Surg Neurol Int 2010;1:3. [18] Nahed BV, Babu MA, Smith TR, Heary RF. Malpractice liability and defensive medicine: a national survey of neurosurgeons. PLoS One 2012;7:e39237. [19] Steiner C, Elixhauser A, Schnaier J. The healthcare cost and utilization project: an overview. Eff Clin Pract 2002;5:143–51. [20] Chandra A, Nundy S, Seabury SA. The growth of physician medical malpractice payments: evidence from the National Practitioner Data Bank. Health Aff (Millwood) 2005. Suppl Web Exclusives:W5-240-W5-9. [21] Charlson ME, Ales KL, Simon R, MacKenzie CR. Why predictive indexes perform less well in validation studies. Is it magic or methods? Arch Intern Med 1987;147:2155–61. [22] Romano PS, Roos LL, Jollis JG. Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: differing perspectives. J Clin Epidemiol 1993;46:1075–90. [23] Brennan TA, Sox CM, Burstin HR. Relation between negligent adverse events and the outcomes of medical-malpractice litigation. N Engl J Med 1996;335:1963–7. [24] Localio AR, Lawthers AG, Brennan TA, Laird NM, Hebert LE, Peterson LM, 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;325:245–51. [25] California Medical Association. Report on the medical insurance feasibility study. San Francisco, CA: California Medical Association, 1977. [26] Studdert DM, Mello MM, Gawande AA, Gandhi TK, Kachalia A, Yoon C, et al. Claims, errors, and compensation payments in medical malpractice litigation. N Engl J Med 2006;354:2024–33. [27] Fisher ES, McClellan MB, Safran DG. Building the path to accountable care. N Engl J Med 2011;365:2445–7. [28] Seabury SA, Chandra A, Lakdawalla DN, Jena AB. On average, physicians spend nearly 11 percent of their 40-year careers with an open, unresolved malpractice claim. Health Aff (Millwood) 2013;32:111–9. [29] General Accounting Office. Major improvements are needed to enhance data bank’s reliability. Washington, DC: General Accounting Office, 2000. [30] Bishop TF, Ryan AM, Casalino LP. Paid malpractice claims for adverse events in inpatient and outpatient settings. JAMA 2011;305: 2427–31.
