Hospital andSurgeon Variation in
Following Incident Lumbar Fusion for
Brook I. Martin, Sohail K. Mirza, Gary M.Franklin,Jon D. Lurie,
Todd A.MacKenzie, and Richard A.Deyo
Objective. To identify factors that account for variation in complication rates across
hospitals and surgeonsperforming lumbar spinal fusionsurgery.
Data Sources. Discharge registry including all nonfederal hospitals in Washington
State from 2004 to2007.
Study Design. We identified adults (n = 6,091) undergoing an initial inpatient lumbar
fusion for degenerative conditions. We identified whether each patient had a subse-
quent complication within 90 days. Logistic regression models with hospital and sur-
geon random effects were used to examine complications, controlling for patient
characteristics and comorbidity.
Principal Findings. Complications within90 daysofafusionoccurredin4.8percent
of the total variability, and surgeon effects account for 14.4 percent. Surgeon factors
account for 54.5 percent of the variation in hospital reoperation rates, and 47.2 percent
of the variation in hospital complication rates. The discretionary use of operative fea-
tures, such as the inclusion of bone morphogenetic proteins, accounted for 30 and 50
Conclusions. To improve the safety of lumbar spinal fusion surgery, quality improve-
ment efforts that focus on surgeons’ discretionary use of operative techniques may be
more effective than those that target hospitals.
Key Words. Lumbar, spine surgery, fusion, repeatsurgery, safety, complications
Low back pain is a condition for which expanding treatments and surgical
innovation have outpaced supporting scientific evidence of their effectiveness
(Deyo et al. 2009). Policy makers are questioning the value of lumbar fusion
surgery for certain indications, and insurance companies have recently
tightened coverage for this common procedure (BlueCross & BlueShield of
Health Services Research
North Carolina 2010). Recent reviews of surgical efficacy suggest that fusion
surgery is no better than multidisciplinary, intensive nonsurgical treatment for
discogenic back pain (pain due to degenerative discs, without sciatica), but
with a worse safety profile and greater cost (Mirza and Deyo 2007; Washing-
ton State Health Care Authority 2007). Poor outcomes of lumbar fusion may
be particularly pronounced in workers’ compensation populations (Maghout-
Juratli et al. 2006). Lumbar fusion may have a clearer role for treating defor-
mities such as degenerative spondylolisthesis and scoliosis (Herkowitz and
for less invasive surgery, such as decompressive laminectomy for spinal steno-
sis, complex fusion procedures that increase the risk of a complication may be
performed (Deyo et al. 2010). Multilevel fusions and circumferential
approaches are often performed without strong evidence of corresponding
improvements in pain or physical functioning. A greater understanding of fac-
tors associatedwith lumbar fusion would helpinformcurrent debates.
Postoperative complications may be influenced by the choice of surgical
technique (Fritzell, Hagg et al. 2002; Deyo et al. 2010; Cizik et al. 2012),
underscoring the imperative to rigorously evaluate the safety of surgical treat-
ments. However, population-based measures of safety have been only spar-
sely reported, and little is currently known about hospital and surgeon
variationin ratesof postoperative complications following spinal fusion.
Using a statewide inpatient discharge registry that allowed us to link suc-
cessive episodes of care for the same patient across multiple years and institu-
tions, we sought to determine the rates of postoperative complications
following fusion for degenerative disease, assess the variation in these rates
across individual hospitals and surgeons, and identify how much of the varia-
tion is accountedfor by operative features.
Address correspondence to Brook I. Martin, Ph.D., M.P.H., The Geisel School of Medicine at
Dartmouth and the Department of Orthopaedics at Dartmouth-Hitchcock Medical Center, One
Medical CenterDrive,HB 7541, Lebanon, NH 03756-0001; e-mail: Brook.I.Martin@Dartmouth.
edu. Sohail K. Mirza, M.D., M.P.H. is with the Department of Orthopaedics at Dartmouth-Hitch-
cock Medical Center. Gary M. Franklin, M.D., M.P.H., is with the Department of Environmental
and Occupational Health Sciences, Neurology, and Health Services, University of Washington
School of Public Health and Community Medicine, and the Washington State Department of
Labor & Industries, Seattle, WA. Jon D. Lurie, M.D., M.S., is with Medicine, and Community and
Family Medicine, The Geisel School of Medicine at Dartmouth and Dartmouth-Hitchcock Medi-
cal Center, Lebanon, NH. Todd A. MacKenzie, Ph.D., is with the Community & Family Medi-
cine, The Geisel School of Medicine at Dartmouth and Dartmouth-Hitchcock Medical Center,
Lebanon, N.H. Richard A. Deyo ,M.D.,M.P.H is with Family Medicine, Medicine, and Public
Health and Preventive Medicine, Kaiser Center for Health Research, Oregon Health & Science
2 HSR:Health Services Research 48:1 (February2013)
The Comprehensive Hospital Abstract Reporting System (CHARS) is an
inpatient discharge database of all nonfederal hospitals in Washington State
(Washington State Department of Health 2011). Hospitals submitting data to
CHARS receive quality reports that they then certify as being at least 95 per-
cent accurate for reporting discharges. We examined CHARS data from 2004
through 2007 for all hospitals and attending surgeons who performed at least
one inpatient lumbar fusion for a degenerative spinal diagnosis. A small num-
ber of cases were dropped because they appeared to bea duplicate record.
columbar, lumbar, or lumbosacral fusion operation for degenerative spinal
conditions from 2004 through 2007. This starting year (2004) was selected
because it corresponds to the first calendar year that codes for three or more
disc levels fused (four or more vertebrae) became available. Patients were
identified using relevant diagnosis and procedure codes from the October
2010 update of the International Classification of Diseases, 9th Revision, Clin-
ical Modification (ICD-9-CM) (Centers for Disease Control & Prevention
2010). Details of our algorithm for classifying spine-related medical encoun-
ters, surgical characteristics, and safety can be obtained from the correspond-
for each admission. We searched all codes to identify patients undergoing
fusion surgery for common degenerative spinal conditions, including disc
degeneration (e.g., spondylosis), herniated discs, stenosis, spondylolisthesis,
and scoliosis. We did not include patients who had nondegenerative spinal
admissions in the previous year, such as spinal fracture, vertebral dislocation,
spinal cord injury, or inflammatory spondylopathy. We also excluded patients
who, in the previous year, had inpatient admission codes for accidents,
All lumbar fusions were included, whether they were combined with a
discectomy or laminectomy. Patients who had other types of spine-related
procedures, including artificial disc replacement, corpectomy, osteotomy,
kyphectomy, insertion of spacers, and insertion of dynamic stabilizing
Safetyof Lumbar Fusion3
devices, were excluded even if these were performed in conjunction with a
fusion operation. We also excluded patients who, in the preceding 10 years,
had any type of prior lumbar spine surgery or had diagnosis or procedure
codes that implied a previous lumbar operation (such as “reopening of lamin-
ectomy site,” “refusion,” or “removal of an internal fixation device”). Previous
surgery has been shown to be an important predictor of higher complication
and repeat surgery rates(Deyo et al. 2011).
We created a composite indicator of 90-day major surgical complications,
defined as postoperative device complication, life-threatening complication,
wound problem, or death occurring within 90 days. We separately examined
the rates of repeat lumbar spine surgery. The presence of each complication
(or repeat surgery) was coded as a dichotomous variable based on ICD-9-CM
tal and attending surgeon performing the initial fusion and was not counted as
an index case or acomplicationfor another hospital or surgeon.
Repeat surgery was identified as the first instance of any second lumbar
spine operation (i.e., “reoperation”) and not necessarily a repeat of the same
procedure or performed at the same vertebral level. Device complications
were based onICD-9-CMdiagnosisandprocedurecodesthatindicateaprob-
lem with an internal orthopedic device (e.g., “malfunction of orthopedic
device”). We did not count device complications or repeat spinal surgeries
coded during the index admission, and we also required that they be co-coded
with a lumbar spine-specific diagnosis and/or procedure code.
We counted wound problems, life-threatening complications, and death
if they occurred during the 90-day postoperative surveillance or during the
index admission. We followed Deyo et al.'s (2010) approach for identifying
life-threatening and wound complications in spine surgery (Deyo et al. 2010).
Life-threatening complications included major medical events such as acute
respiratory failure, cardiopulmonary resuscitation, endotracheal intubation,
pneumonia, stroke, and mechanical ventilation. Life-threatening events have
major consequences on health, and their ICD-9-CM coding is more reliable
than those of minor complications (Lawthers et al. 2000). Postoperative
wound problems were identified using codes for hemorrhages, excisional
debridement of infection, postoperative wound disruption, seroma, and
hematoma complicating a procedure. Mortality was identified by linking each
index caseto the Washington State vital records.
4HSR:Health Services Research 48:1 (February2013)
Potentially Confounding Factors
Patient characteristics and operative features may explain variation in rates of
complications and repeat surgery across hospitals and surgeons. We adjusted
the rates of complications for differences due to patient age, sex, comorbidity,
diagnosis, and primary insurance. We used Quan’s “enhanced” version of the
Charlson index (categorized “none,” “one,” and “two or more”) to adjust for
comorbidity, applying it to admissions occurring at or during the year preced-
ing each index visit (Quan et al. 2005). However, as the comorbidity index
includes myocardial infarctions and strokes, and these are among the life-
these items in the comorbidity score if they occurred during the previous year
The primary source of payment was coded into the following groups:
“Medicare,” “Medicaid,” “Health maintenance organization” (e.g., Kaiser,
Group Health Cooperative), “Commercial insurance” (e.g., Mutual of
Omaha, United Health Plan, Safeco), “Workers’ compensation,” “Health ser-
vice contractor” (e.g., Premera, Premera/Blue Cross), and “Other.” The latter
category included charity cases, self-pay, and “other government sponsored
patients” (e.g., Tri-care, CHAMPUS, Indian Health, Corrections) that com-
bine to account for 2.5 percent of the total cases. Variables for race or ethnicity
were not providedin CHARSduringthe study years.
Some operative features that may be associated with outcomes are iden-
tifiable from ICD-9 codes, including use of bone morphogenetic protein
(BMP), surgical approach (anterior, posterior, or combined/circumferential),
an indicator of whether fusion was combined with a decompression, and an
indicator of whether 3+ disc levels were fused(Fritzell, Hagg et al.2002; Deyo
et al. 2010, 2012; Mirza 2011; Cizik et al. 2012). We also included the attend-
ing surgeon’s fusion volume in the preceding 365 days to account for any
potential association between-surgeon experience and outcomes (Silber et al.
2010). This measure of recent volume may be a better indicator of surgeon
experience than the annual or the cumulative volume, both of which may be
prone to misclassification (French and Heagerty 2011). We grouped the index
cases into quartiles based on the ranking of the surgeon’s volume.
Bivariate associations between patient characteristics and complications were
assessed using chi-square or t-test comparisons. We examined the risk for
Safetyof Lumbar Fusion5
complications using logistic regression analysis, including only patients who
had a minimum of 90 days of surveillance available to assess each outcome.
For example, patients who had an initial operation in December 2007 were
not eligible to be assessed for the 90-day outcomes because the data only
We used a logistic regression model that allowed a random-intercept for
each hospital (Rabe-Hesketh and Skrondal 2008). These models prevent hos-
pitals with a relatively small volume of cases from being misclassified when
high sampling variability may account for apparently poor (or excellent) per-
formance (i.e., through the use of “shrinkage factors”). The models also adjust
the standard errors to account for similarities occurring among patients “clus-
To compare risk-adjusted outcomes across hospitals, we used the results
from the logistic regression models to estimate the risk of complication for an
“average” patient. This was accomplished by setting the covariates for age,
sex, comorbidity, insurance, and diagnosis to the mean statewide distributions
and then reporting the mean rate of complications within each hospital, along
with a 95 percent empirical Bayes confidence interval.
To examine the influence of surgeon factors on the rates of complica-
tions, we then added an additional random-effect parameter to the previous
(“hospital only”) model. This parameter represents the variability attributed
to surgeons’ “nested”withineachof the hospitals he/she operated in.
Model construction was based on the principles by Hosmer and Leme-
show (2000), including checking for interactions. We used likelihood ratio
and Aikake’s Information Criteria (AIC) to identify the models with the best
fit. An advantage of AIC is that it allows comparisons of models with the
same number of parameters, a useful consideration when comparing
random-effects models. The best-fitted model, with the lowest AIC, included
patient and operative features along with surgeon random-effect. The specifi-
cation for this final model, along with 95 percent Bayesian coverage inter-
1 ? lKk
¼ a þ Kkþ bXik
95%CIk¼ ð^Kk? 1:96
In this specification, l represents the probability of a complication for
patient i, whose initial operation was performed by surgeon k. The terms
6 HSR:Health Services Research 48:1 (February2013)
a combined with Kkrepresents a surgeon-specific intercept. X represents a
vector of covariates with corresponding beta-coefficients for patient character-
istics (age, sex, comorbidity, primary diagnosis and insurance) and operative
features (surgical approach, multilevel fusion, BMP, surgeon volume, instru-
We found no evidence of a poorly fitted model for complication
(p = .5615) using Hosmer–Lemeshow goodness-of-fit statistic. We confirmed
linearity between the logarithmic odds for reoperations with each of the
variables in the model, and no issues were identified that forced us to trans-
form any variables. No changes in parameter coefficients as variables were
added to the model suggested evidence of multicollinearity; this was con-
firmed by examining variance inflation factors and tolerance statistics in the
mally distributed. Empirical Bayes predictions of the random-effects were
assessed for normality using a Q-Q plot. There were no serious violations in
the random-effect distribution through the bulk of the data. However, at the
high end of the random-effect spectrum, several surgeons’ complication rates
deviated higher than predicted, indicating that they had unusually high com-
plication rates. This finding translates into a conservative estimation (underes-
timate of their actual rates) of the complication rates for these particular
surgeons, because they are “over-fitted” toward the overall state mean. Our
final model had a C-statistic of 0.662 for the complication model and 0.660 for
We then estimated the proportion of total variation that was explained
by the inclusion of hospital (or surgeon) effects. The total variability for a ran-
dom-effect model of a dichotomous outcome is calculated as the sum of hospi-
tal (surgeon)-specific variance(s) (w) plus the mathematical constant p2/3
(Rabe-Hesketh and Skrondal 2008). The proportion of total variation
explained by covariates that are added to the previously null model (a model
with random-effects parameter but no explanatory variables) is calculated
using the following:
1 ? varðFullModelÞ
An intraclass correlation coefficient (ICC) was calculated for hospital
(surgeon) effects to estimate the proportion of variation explained within the
hospitals (surgeons) by patient characteristics and operative features. The ICC
Safetyof Lumbar Fusion7
differences not accounted for by covariates in the model. For a dichotomous
outcome, the ICC is calculated using the followingequation:
w þ p2=3
To measure the proportion of hospital (surgeon) variation that was
explained by the addition of covariates, a model with only random-effects
(null) was compared with subsequent models with covariates. A reduction in
the ICCascovariates are added tothe modelrepresentsthe proportion ofhos-
pital (surgeon)variation that is explained at that level by the addedcovariates.
We also examined our findings after excluding six (24 percent) hospitals
that included only one surgeon who performed fusions because, in these
instances, the random-effect models cannot differentiate the variation in com-
plications at the hospital level from that of the surgeon level. All analyses were
performed using Stata-MP, version 11 (College Station, TX), with hypothesis
testing performed usingatwo-sided alpha levelset at 0.05.
Atotal of 7,680 patients were identified as having an initial inpatient fusion for
lumbar degenerative conditions between 2004 and 2007. Atotal of 1,589 (20.7
percent) were excluded for reasons shown in Table 1. The predominant rea-
son for exclusion was lumbar surgery in the previous 10 years. Excluded
patients were statistically older (57.7 years vs. 56.4 years, p = .001), but this
was not clinically different. They were also less likely to be female (55.1 per-
cent vs. 60.7 percent, p < .001), more likely to be insured by workers’ com-
pensation (16.6 percent vs. 11.3 percent, p < .001), to have a comorbidity
score greater than one (41.8 percent vs. 27.2 percent, p < .001), less likely to
have spondylolisthesis (33.1 percent vs. 54.1 percent), and were more likely to
have degenerative disc disease (24.9 percent vs. 18.0 percent) and herniated
disc (16.2 percentvs. 9.4 percent).
The study population consistedof 6,091 patients who had an initial inpa-
tient spinal fusion during the study period (Table 2). The mean age of the
cohort was 56.4 years (SD 13.9); 31.3 percent were insured by Medicare, 6.1
percent by Medicaid, 11.3 percent by workers’ compensation; and 27.2 per-
cent had a comorbidity index greater than zero. Adiagnosis of spondylolisthe-
sis accounted for the largest fraction of the fusion operations (54.1 percent),
8 HSR:Health Services Research 48:1 (February2013)
followed by degenerative disc disease (18.0 percent), herniated disc without
myelopathy (9.4 percent), spinal stenosis (9.0 percent), scoliosis (7.1 percent),
and herniated disc with myelopathy (2.3 percent). The index operations were
performed by 298 surgeons, who on average performed 20 fusions each
(range: 1–333) during the study period. These were performed within 43 hos-
pitals that performed an average of 142 (range: 3–533) fusions during the
Age between 61 and 80, greater comorbidity, and Medicare and Medicaid
insurance were associated with higher risk of having a complication within
90 days (Table 2). Workers’ compensation (a public payer in Washington
State) and health maintenance organizations had the lowest 90-day rates of
complications among all types of insurance. Having 3+ disc levels fused, use
of bone morphogenetic proteins, and anterior surgical approaches were asso-
ciated withhighercomplication rates.
percent) were lower than for disc degeneration (4.2 percent), spondylolisthesis
(4.5 percent), spinal stenosis (6.4 percent), herniated disc with myelopathy (8.6
percent), or scoliosis (9.7 percent). Complications were not mutually exclusive.
For example, the same patient may have had a device complication and a reop-
eration. Wound problems were the most common type of complications (3.1
ingtonState ComprehensiveAbstract Reporting System, 2004–2007
Reasons for Exclusion from Lumbar Fusion Safety Study: Wash-
ExclusionFactors(Not MutuallyExclusive)NumberExcluded(n = 1,589)
HIVor immune deficiencyinpreviousyear
Lumbarspinesurgeryin previous10 years
Safetyof Lumbar Fusion9
Rates for Any Postoperative Adverse Events Following Lumbar Spinal Fusion Surgery for Common
Degenerative Diagnoses (CHARS2004–2007)
Complications (ordeath) within90 Days,n (%)
Number eligible for surveillance
10HSR: Health Services Research 48:1 (February2013)
Table 2. Continued
Complications(ordeath) within90 Days, n(%)
Herniated discwith myelopathy
Fus + deco
Safety of Lumbar Fusion 11
Table 2. Continued
Complications (ordeath)within 90 Days,n(%)
Number (Column %)
*Significantdifference withincategoriesofvariable (<.05).
†Surgeon1-yearfusionvolume groupis based on fusion volume foranytype offusionor re-fusion priortoapplyingexclusion criteria.
12HSR: Health Services Research 48:1 (February2013)
within 90 Days of an Initial Inpatient Lumbar Fusion in Washington State
Multivariate Analysis of Repeat Surgery and Complications and
Reoperationwithin90 Days Other MajorComplication
Safetyof Lumbar Fusion 13
percent), followed by repeat surgery (2.2 percent), life-threatening complica-
tions (2.0 percent), device problems (0.8 percent), and death (0.2 percent).
Those with a diagnosis of scoliosis had a higher rate of complications than all
Table 3 presents the multivariate logistic regression models with sur-
geon random-effects used to estimate the risk-adjusted rates for complications
within 90 days. The overall 90-day complication rates for surgeons, adjusted
for patient age, sex, insurance type, comorbidity, and diagnosis was 4.1 per-
cent (95% CI 2.7–6.4), and the adjusted mean 90-day reoperation rate was 1.7
percent (95% CI0.9, 3.7).
Anterior operative approaches were associated with a significantly
higher risk for 90-day repeat surgery (OR 3.33; 95% CI 2.03–5.44), even after
adjusting for patient characteristics, comorbidity, diagnosis, insurance status,
and other operative features. Circumferential fusions were associated with a
higher90-dayrisk forcomplications (OR1.24;95%CI0.82–1.89)andreoper-
ations (OR 1.13; 95% CI 0.59–2.17), but they did not reach statistical signifi-
cance in this small subset of patients. Patients who had 3+ disc levels fused had
a higher risk for complications (OR 1.64; 95% CI1.12–2.40).
Higher surgeon case volume was not associated with a significantly
lower 90-day repeat surgery or complication rate. The use of bone morphoge-
netic proteins (BMP) was associated with a nonsignificantly higher risk for
complication (OR 1.20; 95% CI 0.88–1.63) and repeat surgery (OR 1.78; 95%
CI 1.17–2.69). Having 3+ disc levels fused was also associated with an
increased risk for complication (OR 1.64; 95% CI 1.12–2.40) and repeat sur-
gery (OR 1.78; 95% CI1.17–2.69).
Scoliosis was associated with a higher risk for complications within a
given surgeon, while the risk for stenosis, spondylolisthesis,and herniated disc
Reoperationwithin90 DaysOther MajorComplication
†Odds ratio based on generalized linear and latent mixed models using Stata-MP command
14 HSR: Health Services Research 48:1 (February2013)
did not significantly differ from that of disc degeneration. Stenosis was associ-
ated with higher reoperation rates within a given surgeon, compared with
degenerative disc disease. Adjusting for operative features lowered the odds
ratio for complications among those with scoliosis from 1.94 (95% CI 1.23–
3.09, not shown) to 1.58(1.25–3.87, final model.)
The rates for complications are shown in Figure 1, with each spike rep-
resenting a single hospital or surgeon. We found that the nearly eightfold vari-
ation in risk-adjusted repeat surgery across surgeons was substantially
attributed to a few surgeons (11/298, 3.7 percent) with rates significantly
above the state mean (represented by a solid horizontal line). Limiting our
analysis to the hospitals where more than one surgeon performed a fusion did
tion (Bottom Row) Rates Following Inpatient Lumbar Spinal Fusion Surgery
for Common Degenerative Diagnoses (Each spike represents 95 percent
Bayesian confidence interval for the rates within a hospital (left) or surgeon
nested within hospitals (right) in Washington State; the solid horizontal line
represents the statewide mean)
Risk-Adjusted 90-Day Repeat Surgery (Top Row) and Complica-
Safety of Lumbar Fusion15
not substantially change the variance estimates or lead us to alter our conclu-
Table 4 provides the random-effect variance and fit parameters for a model
without any covariates (null model, containing only random effects), as well as
for models adding patient characteristics and operative features. In the null
model for reoperations, the proportion of total variation due to hospital effects
decreased from 8.8 percent to 4.0 percent with the addition of a parameter for
the surgeon effects (a reduction of 54.5 percent). Similar reductions in hospi-
tal-level variation were observed in the models containing patient characteris-
tics and operative features. Hospital effects accounted for a smaller proportion
tially reduced by the inclusionof a surgeon parameter.
Surgeon effects accounted for 14.4 percent of the total variation in reo-
perations and 5.0 percent of the total variation for complications. Surgeon-
level variation for both repeat surgery and complications was greater than the
variation observed among hospitals. For example, the variance for the sur-
geon effects for repeat surgery was 0.341 (9.0 percent of total), compared with
0.153 (4.0 percent)for hospitals.
Explaining Hospital (Surgeon) Variation
The addition of patient characteristics did not reduce the variability in reoper-
ation rates, but it accounted for 32.6 percent of the between-surgeon variation
in complication rates. The addition of operative features accounted for 4.4
percent of the total variation in reoperations (30 percent of the variability
between surgeons) and 2.5 percent of the total variation in complications (50
percent of the between-surgeon variation).
Description of Reoperations
The unadjusted rate of repeat lumbar spine surgery in our cohort was 2.2 per-
cent at 90 days and 5.0 percent at 1 year. We found that 115/137 (84 percent)
of the reoperations were performed by the same surgeon who performed the
initial surgery. The most common diagnoses at the time of these reoperations
was spinal stenosis (21.2 percent), followed by disc degeneration (18.3 per-
cent), spondylolisthesis (17.5 percent), disc herniation with myelopathy (13.1
16 HSR: Health Services Research 48:1 (February2013)
Random-Effects Variance and Model Fit Parameters of Multivariate Models for Repeat Surgery and Compli-
cations within 90 Days of an Initial Inpatient Lumbar Fusion in Washington State (CHARS2004–2007)
Safety of Lumbar Fusion17
percent), disc herniation (6.6), and scoliosis (5.8 percent). The remaining 16.1
pedic device or surgical aftercare. Device complication codes were included
in 36/137 (26 percent) of the all repeat surgeries within 90 days, and the code
for “arthrodesis status” (implying a problem at the same vertebral level as the
index surgery) in 53/137 (38.7 percent). The most common procedures coded
at the time of reoperation were fusion or re-fusion (51.8 percent), decompres-
sion only (22.6 percent), and other spinal procedures such as removal of
orthopedic devices (25.5 percent).
Complications within 90 days of an initial lumbar fusion operation for a
degenerative diagnosis occurred in 4.9 percent of patients, and 2.2 percent of
patients had a second operation within 90 days. After adjusting for patient
characteristics and diagnosis, we found that the mean 90-day complication
rate for individual surgeons varied from 2.5 percent to 11.7 percent.The range
for reoperations varied from 0.6 percent to 9.3 percent.
Estimating complications is notoriously difficult in studies involving sur-
gical treatments, a fact that is reflective of the large unexplained variation in
our models. Our findings that the majority of the total variability (~85 percent)
occurredwithin (ratherthan between) surgeons orhospitals suggests that com-
plications and reoperations may be more encounter-specific than they are due
to systematic difference in quality across surgeonsorhospitals.
Nevertheless, the “explainable” variation among hospitals in the rate of
complications was substantially reduced after including surgeon effects in our
models. Furthermore, the discretionary use of BMP, 3+ disc levels fused, and
surgical approach accounts for 50 percent and 30 percent of the variation
among surgeons’ complication and repeat surgery rates, respectively. This
suggests that for improving safety, the addition of surgeon quality improve-
ment efforts may be more effective than those solely targeting the hospital.
However, surgeon quality improvement programs should not necessarily
replace current hospital-level quality efforts.
Among 6,091 initial fusions, 366 patients had either a complication or a
within each type of adverse event, or attributed to individual providers, the
uncertainty surrounding individual estimates became larger, potentially
18 HSR: Health Services Research 48:1 (February2013)
physician decision making may be a key to improving surgical safety, the low
precision of surgeon-level empirical performance of complications in these
with external benchmarks that are not derived from the data (Jones and
Spiegelhalter 2011). At a policy level, the question is whether acting on sur-
geon-level adverse event data, with this known imprecision, will lead to better
delivery of health care to a population. While the most serious or frequent
harms that may arise from fusion operations may lead to more detailed review
of a small number of specific surgeons, surgeon-focused quality improvement
The finding of a lower adverse event rate in the workers’ compensation
population may seem surprising. Patient-reported outcomes of surgical proce-
dures, including lumbar fusion, are consistently worse for workers’ compensa-
tion compared with nonworkers’ compensation populations (Harris et al.
2005). However, most of these studies did not specifically examine surgical
complications or adjust for fusion indications or comorbidities. Because of
poor outcomes reported for fusion in Washington State, including high rates
of reoperation (Franklin et al. 1994; Martin et al. 2007), the Washington’s
workers’ compensation program had a restrictive fusion coverage policy
implemented through prospective utilization review during the years under
study. These policies included limiting initial fusions to a single level and
requiringmeasureable instability for approval (Wickizeret al. 2004). The pro-
portion of workers’ compensation patients who had 3+ disc levels fused in our
study was 3.1 percent, compared with 8.5 percent for all other payers. Our
findingof lower complication rates in these patients may suggest that the more
parsimonious use of fusion in this population may have reduced complica-
In our data, we could not corroborate Bederman’s (2009) (Bederman
et al. 2009) finding in a Canadian population or Farjoodi’s (2012) (Farjoodi,
Skolasky, and Riley 2011) study of the Nationwide Inpatient Sample that
patients operated on by high-volume surgeons had a significantly lower risk
for reoperation. Differences in the study populations, type of procedure, and
measurement of volume may explain the differences between these studies.
Administrative data generally allow for a longer duration of follow-up
than clinical studies and can often identify subsequent care even when it
occurs at a different institution. Despite these strengths, our study has a
Safety of Lumbar Fusion19
number of limitations that arise from the analysis of observational data. Such
data are susceptible to confounding factors that limit causal inference. The
presence of such an unmeasured factor would have to have a large effect and
be systematically disproportionate across providers to alter individual esti-
mates. By excluding significant comorbidity, nondegenerative spinal pathol-
ogy, and previous surgery, as well as adjusting our model for comorbidity and
diagnoses, we have accounted forsome of this potential confounding.
The complication rates that we report may be an underestimate of the
actual rates because we only counted those that were associated with a read-
mission. Some postoperative events, such as pneumonia, might have been
treated in an outpatient setting and not counted as an adverse event in our
analysis. In addition, the use of normally distributed random effects results in
a conservative estimation for a few providers with higher than expected rates.
Future efforts should use sampling techniques to fit nonnormally distributed
Although ICD-9-CM codes are commonly used in spinal research, they
lack specific clinical detail, such as disease severity or pain intensity, specifica-
tion of exact vertebral levels, and functioning. Although this may lead to some
imprecision, administrative data are reliable for ascertaining major complica-
tions (Lawthers et al. 2000; Campbell et al. 2011).
We could not account for patient migration into or out of Washington
State, which may influence the rates of complications that we observed.
In addition, we were unable to identify lumbar operations occurring in our
cohort prior to 1987, or those that occurred outside of Washington State. As a
result, some patients in our analysis may have had a previous lumbar opera-
tion that was unknown to us.
The use of administratively derived patient safety indicators following
fusion surgery has not been rigorously validated through a comparison of
chart reviews. However, readmission, infections, mortality, and life-threaten-
ing complications are part of the National Surgical Quality Improvement
Program (NSQIP), which has been used as the gold standard to improve the
validity of patient safety indicators (Romano et al. 2009). Furthermore, our
estimates for repeat surgery and mortality following fusion surgery are
similar to those reported in both administrative and clinical studies (Malter
et al. 1998; Martin et al. 2007, 2011; Juratli et al. 2009). Future research
efforts should focus on the validation and improvement of claims-based
methods and seek to understand how physicians’ judgments, technical influ-
ences, or other factors may account for the variation in rates that we
20 HSR: Health Services Research 48:1 (February2013)
Safety data regarding spine surgery are available from only a few ran-
domized trials. The Sport Patient Outcomes Research Trial (SPORT) did not
focus on differences in outcomes based on the type of operative feature, was
not limited to fusion procedures, and did not primarily focus on safety (Wein-
stein et al. 2006, 2008). Fritzell, Hagg et al. (2002) found that circumferential
fusions had a higher rate of complications compared with posterolateral
fusions. These findings havenotbeenconfirmedinalargepopulation.Popula-
tion-based studies allow us to provide empirical performance data for postop-
erative complications among surgeons performing lumbar fusion. Data on
complications may be useful to inform policies that aim at making spinal
fusion surgery safer. By helping patients to weigh the potential harms and
potential benefits, such data may be useful for soliciting informed consent and
for engaging them in “shared decision making.” Future efforts should strive to
clarify indications for complex fusion operations, where the risk for complica-
tions may be substantial.
Joint Acknowledgment/Disclosure Statement: Dr. Martin receives partial salary
support or consulting fees for work on several back pain–related research
grants funded by the NIH, AHRQ, and the National Bureau of Economic
Research (NBER). Within the past five years Dr. Martin received partial sal-
ary support through a gift to the University of Washington from Surgical
Dynamics (now acquired by Stryker Corporation), a maker of surgical
implants. Dr. Franklin is a principal investigator for CDC Grant
5R21CE001850-02 but does not receive any salary from this. Dr. Lurie
receives salary support from Dartmouth Hitchcock Medical Center and Dart-
mouth College with grant funding from NIH and AHRQ. He has also served
as a paid consultant to Blue Cross Blue Shield Corporation, the Foundation
for InformedMedical Decision Making, Baxano Inc., FzioMedInc., and New-
Vert Ltd. Dr. MacKenzie has no disclosures. Dr. Deyo holds a professorship
endowed by a gift to Oregon Health and Science University from Kaiser Per-
manente. He receives honoraria for serving on the Board of Directors for the
non-profit Foundation for Informed Medical Decision Making. He also
receives honoraria from UpToDate for authoring topics on low back pain.
He receives honoraria from the Robert Wood Johnson Foundation for serving
on a National Advisory Committee to a physician research training program.
He is the principal investigator, co-investigator, or consultant for multiple
Safety of Lumbar Fusion21
grants from the NIH and AHRQ. No other disclosures. Dr. Mirza is the PI of
an AHRQ research grant to examine the safety of spinal surgery. He receives
royalties for a patented surgical drill, through the University of Washington’s
Technology Transfer Office. Within the past five years his research benefitted
from support of a gift to the University of Washington from Surgical Dynam-
ics (now acquired by Stryker Corporation), a maker of surgical implants.
Bederman, S. S., H. J. Kreder, I. Weller, J. A. Finkelstein, M. H. Ford, and A. J. Yee.
2009. “The Who, What and When of Surgery for the Degenerative Lumbar
Spine: A Population-Based Study of Surgeon Factors, Surgical Procedures,
Recent Trends and Reoperation Rates.” Canadian Journal of Surgery 52 (4):
BlueCross & BlueShield of North Carolina. 2010. Lumbar Spine Fusion Surgery “Notifica-
tion” [accessed on September 6, 2011]. Available at http://www.bcbsnc.com/
Campbell, P. G., J. Malone, S. Yadla, R. Chitale, R. Nasser, M. G. Maltenfort, A.
Vaccaro, and J. K. Ratliff. 2011. “Comparison of ICD-9-Based, Retrospective,
and Prospective Assessments of Perioperative Complications: Assessment of
Accuracy in Reporting.”Journal of Neurosurgery: Spine14 (1): 16–22.
Centers for Disease Control & Prevention. 2010. “International Classification of Dis-
eases, Ninth Revision, Clinical Modification” [accessed on September 6, 2011].
Available at http://www.cdc.gov/nchs/icd/icd9cm.htm
Cizik, A. M., M. J. Lee, B. I. Martin, R. J. Bransford, C. Bellabrba, J. R. Chapman, and
S. K. Mirza. 2012. “Using the Spine Surgical Invasiveness Index to Identify Risk
of Surgical Site Infection: A Multivariate Analysis.” Journal of Bone and Joint Sur-
gery.American Volume94 (4): 335–42.
Deyo, R. A., S. K. Mirza, J. A. Turner, and B. I. Martin. 2009. “Overtreating Chronic
Back Pain: Time to Back off?” Journal of the American Board of Family Medicine 22
Deyo, R. A., S. K. Mirza, B. I. Martin, W. Kreuter, D. C. Goodman, and J. G. Jarvik.
2010. “Trends, Major Medical Complications, and Charges Associated with
Surgery for Lumbar Spinal Stenosis in Older Adults.”Journal of the American Med-
ical Association 303 (13): 1259–65.
Deyo, R. A., B. I. Martin, W. Kreuter, J. G. Jarvik, H. Angier, and S. K. Mirza. 2011.
“Revision Surgery Following Operations for Lumbar Stenosis.” Journal of Bone
and Joint Surgery.American Volume 93 (21): 1979–86.
Deyo, R. A., A. Ching, L. Matsen, B. I. Martin, W. Kreuter, J. G. Jarvik, H. Angier, and
S. K. Mirza. 2012. “Use of Bone Morphogenetic Proteins in Spinal Fusion
22HSR: Health Services Research 48:1 (February2013)
Surgery for Older Adults with Lumbar Stenosis: Trends, Complications, Repeat
Surgery,and Charges.”Spine (Phila Pa1976) 37 (3): 222–30.
Farjoodi, P.R. L. Skolasky, and L. H. Riley. 2011. “The Effects of Hospital and Surgeon
Volume on Postoperative Complications after Lumbar Spine Surgery.” Spine 36
Fischgrund, J. S., M. Mackay, H. N. Herkowitz, R. Brower, D. M. Montgomery,
and L. T. Kurz. 1997. “1997 Volvo Award Winner in Clinical Studies. Degen-
erative Lumbar Spondylolisthesis with Spinal Stenosis: A Prospective,
Randomized Study Comparing Decompressive Laminectomy and Arthrode-
sis with and without Spinal Instrumentation.” Spine (Phila Pa 1976) 22 (24):
Franklin, G. M., J. Haug, N. J. Heyer, S. P. McKeefrey, and J. F. Picciano. 1994. “Out-
come of Lumbar Fusion in Washington State Workers’ Compensation.” Spine
(Phila Pa1976) 19 (17): 1897-903; discussion 1904.
French, B., and P. J. Heagerty. 2011. “A General Framework for Estimating Vol-
ume-Outcome Associations from Longitudinal Data.”Statistics in Medicine 31 (4):
Fritzell, P., O. Hagg, P. Wessberg, A. Nordwall, and the Swedish Lumbar Spine
Study Group. 2002. “Chronic Low Back Pain and Fusion: A Comparison of
Three Surgical Techniques: A Prospective Multicenter Randomized Study
from the Swedish Lumbar Spine Study Group.” Spine 27 (11): 1131–41.
Harris, I., J. Mulford, M. Solomon, J. M. van Gelder, and J. Young. 2005. “Association
between Compensation Status and Outcome after Surgery: A Meta-Analysis.”
Journal ofthe American Medical Association293 (13): 1644–52.
Herkowitz, H. N., and L. T..Kurz.1991. “Degenerative Lumbar Spondylolisthesis with
Spinal Stenosis. A Prospective Study Comparing Decompression with Decom-
pression and Intertransverse Process Arthrodesis.” Journal of Bone and Joint Sur-
gery.American Volume73 (6): 802–8.
Hosmer, D. W., and S. Lemeshow. 2000. Applied Logistic Regression. 2nd ed. New York:
Jones, H. E., and D. J. Spiegelhalter. 2011. “The Identification of “Unusual”
Health-Care Providers from a Hierarchical Model.” The American Statistician 65
Juratli, S. M., S. K. Mirza, D. Fulton-Kehoe, T. M. Wickizer, and G. M. Franklin. 2009.
“Mortalityafter Lumbar FusionSurgery.”Spine (Phila Pa1976) 34 (7): 740–7.
Lawthers, A. G., E. P. McCarthy, R. B. Davis, L. E. Peterson, R. H. Palmer, and L. I.
Iezzono. 2000. “Identification of In-Hospital Complications from Claims Data.
IsIt Valid?”Medical Care38 (8): 785–95.
Maghout-Juratli, S., G. M. Franklin, S. K. Mirza, T. M. Wickizer, and D. Fulton-Kehoe.
2006. “Lumbar Fusion Outcomes in Washington State Workers’ Compensa-
tion.”Spine (Phila Pa1976) 31 (23): 2715–23.
Malter, A. D., B. McNeney, J. D. Loeser, and R. A. Deyo. 1998. “5-Year Reoperation
Rates after Different Types of Lumbar Spine Surgery.” Spine (Phila Pa 1976) 23
Safety of Lumbar Fusion 23
Martin, B. I., S. K. Mirza, B. A. Comstock, D. T. Gray, W. Kreuter, and Deyo R. A.
2007. “Reoperation Rates Following Lumbar Spine Surgery and the Influence of
SpinalFusionProcedures.” Spine(Phila Pa1976) 32 (3): 382–7.
Martin, B. I., S. K. Mirza, D. R. Flum, T. M. Wickizer, P. J. Heagerty, A. F. Lenkoski,
andR. A. Deyo. 2011.“RepeatSurgery afterDecompressionforHerniatedDisc:
The Quality Implications of Hospital and Surgeon Variation.” The Spine Journal
12 (2): 89–97.
Mirza, S. K.. 2011. “Folly of FDA-Approval Studies for Bone Morphogenetic Protein.”
Journal ofSpine 11 (6): 495–9.
Mirza,S.K.,andR.A.Deyo.2007.“Systematic Review ofRandomized TrialsCompar-
ing Lumbar Fusion Surgery to Nonoperative Care for Treatment of Chronic
Back Pain.”Spine (Phila Pa1976) 32 (7): 816–23.
Quan,H.,V.Sundararajan,P.Halfon, A.Fong,B.Burnand, J.C.Luthi,L.D.Saunders,
C. A. Beck, T. E. Feasby, and W. A. Ghali. 2005. “Coding Algorithms for Defin-
ingComorbiditiesin ICD-9-CM andICD-10 Administrative Data.”Medical Care
43 (11): 1130–9.
Rabe-Hesketh, S., and A. Skrondal. 2008. Multilevel and Longitudinal Modeling Using
Stata,2ndEdition.College Station, TX: Stata Press.
Romano, P. S., H. J. Mull, P. E. Rivard, S. Zhao, W. G. Henderson, S. Loveland,
D. Tsilimingras, C. L. Christiansen, and A. K. Rosen. 2009. “Validity of Selected
AHRQ Patient Safety Indicators Based on VA National Surgical Quality
Improvement Program Data.”Health Services Research 44 (1): 182–204.
Silber, J. H., P. R. Rosenbaum, T. J. Brachet, R. N. Ross, L. J. Bressler, O. Even-Sho-
shan, S. A. Lorch, and K. G. Volpp. 2010. “The Hospital Compare Mortality
Model and the Volume-Outcome Relationship.” Health Services Research 45
(5 p 1): 1148–67.
Washington State Department of Health, Comprehensive Hospital Abstract Reporting
System. 2011. “Hospital and Patient Data Systems, Comprehensive Hospital
Abstract Reporting System” [accessed on September 6, 2011]. Available at
Washington State Health Care Authority. Health Technology Assessment Program.
2007. “Findings and Coverage of Lumbar Fusion & Discography” [accessed on
September 6, 2011]. Available at http://www.hta.hca.wa.gov/lumbar.html
Weinstein, J. N., T. D. Tosteson, J. D. Lurie, A. N. Tosteson, B. Hanscom, J. S. Skinner,
W. A. Abdu, A. S. Hilibrand, S. D. Boden, and R. A. Deyo. 2006. “Surgical vs
Nonoperative Treatment for Lumbar Disk Herniation: The Spine Patient Out-
comes Research Trial (SPORT): A Randomized Trial.” Journal of the American
Medical Association 296 (20): 2441–50.
Weinstein, J. N., T. D. Tosteson, J. D. Lurie, A. N. Tosteson, E. Blood, B. Hanscom,
H. Herkowitz, F. Cammisa, T. Albert, S. D. Boden, A. Hilibrand, H. Goldberg,
S. Berven, H. An, and Sport Investigators. 2008. “Surgical versus Nonsurgical
Therapy for Lumbar Spinal Stenosis.” New England Journal of Medicine 358 (8):
Wickizer, T. M., G. Franklin, J. V. Gluck, and D. Fulton-Kehoe. 2004. “Improving
Quality through Identifying Inappropriate Care: The Use of Guideline-Based
24 HSR: Health Services Research 48:1 (February2013)
Utilization Review Protocols in the Washington State Workers’ Compensation Download full-text
System.”Journal ofOccupational and Environmental Medicine 46 (3): 198–204.
Additional supporting information may be found in the online version of this
Please note: Wiley-Blackwell is not responsible for the content or func-
tionality of any supporting materials supplied by the authors. Any queries
(other than missing material) should be directed to the corresponding author
for the article.
Safety of Lumbar Fusion25