Patient travel burden and postoperative outcomes after lumbar fusion

Abstract

Background

Geographic barriers to surgical care have been linked to poor postoperative outcomes. Patients living farther from surgical centers may experience delays in care, which may adversely impact postoperative outcomes. While the influence of travel time and distance, or travel burden, on patient outcomes has been examined in other surgical populations, its impact on lumbar fusion is understudied. The authors hypothesize that travel burden represents an independent risk factor for adverse postoperative events following lumbar fusion.

Methods

A retrospective cohort study was performed at an academic Level I trauma center examining adult patients who underwent lumbar fusion (L1–L5) from 2013 to 2025. Patients with spinal malignancy, pregnancy, rheumatologic disease, polytrauma, and prior failed fusion were excluded. Travel distance and driving time to the hospital were calculated using Google Maps. Outcomes of interest included postoperative length of stay (LOS) readmissions and surgical, infectious, neurologic, pulmonary, cardiovascular, and thromboembolic complications. Analyses included multivariate Firth penalized logistic regression and multivariate linear regression.

Results

The final cohort included 403 patients (mean age 51.4 ± 10.7 years; BMI 30.5 ± 5.9). Mean travel distance and time were 44.4 ± 47.9 miles and 60.2 ± 50.1 minutes respectively. Travel burden was not significantly associated with any postoperative complication or LOS.

Conclusions

In this cohort, travel burden was not associated with postoperative complications or length of stay following lumbar fusion. These findings suggest that travel burden alone does not independently constitute a risk factor for poor lumbar fusion outcomes, which may serve to reassure both patients and surgeons, though further validation is warranted.

Introduction

Geographic barriers to surgical care, including increased travel time and distance, have been associated with adverse outcomes across multiple surgical specialties. Patients with greater travel burden may experience delays in preoperative evaluation, less access to postoperative follow-up, and challenges in coordinating rehabilitation services, all of which can contribute to higher rates of perioperative medical and surgical complications [[1], [2], [3]]. Additionally, poor geographic access to healthcare has consistently been linked to low socioeconomic status in both rural and nonrural settings, suggesting that a structural disparity is partially responsible for increased travel burden [[4], [5], [6]]. These associations are often described through the concept of distance decay, whereby health outcomes worsen as distance from care increases [[7], [8], [9]]. However, prior literature has also identified a competing phenomenon of distance bias in which patients traveling longer distances demonstrated equivalent or even improved outcomes, potentially reflecting a selective referral to high-volume centers or patient-level selection factors [[10], [11], [12]].

The consideration of travel burden is relevant in spine surgery, where perioperative management is complex and postoperative care is resource-intensive. Lumbar fusion patients often necessitate closer early follow-up and timely wound surveillance by the surgical team to ensure rapid recognition of specialty-specific neurologic or infectious complications—a process that may be especially sensitive to geographic barriers. Concurrently, the regionalization of complex spine care and the continued growth of the elderly surgical population have resulted in an increasing proportion of patients traveling substantial distances to receive surgical treatment [[13], [14], [15]]. Despite these trends, the relationship between travel burden and postoperative outcomes following lumbar fusion remains poorly defined.

Prior investigations examining travel distance and surgical outcomes have largely focused on nonspine populations, including colorectal surgery, oncologic resections, and total joint arthroplasty (TJA), with overall inconsistent findings [2,[16], [17], [18], [19], [20], [21]]. In previous orthopedic literature, travel distance did not affect patient-reported outcomes after primary total knee arthroplasty [22]. Similar results were also reported in shoulder arthroplasty patients, where the overall outcomes at 1-year follow-up were not affected by travel burden [23]. While these studies suggest that increased travel time and/or distance may not adversely affect outcomes when patient care is delivered at high-volume or specialized centers, this association has yet to be defined for lumbar fusion.

The purpose of our study was to examine the association between travel burden, measured by both travel distance and travel time, and postoperative outcomes following lumbar fusion surgery at a single academic level I trauma center. We hypothesized that increased travel time and distance to the treating hospital would represent independent risk factors for adverse postoperative events, including complications, nonhome discharge, readmission, and prolonged length of stay following lumbar fusion.

Methods

Study design and data source

A retrospective cohort study was conducted following Institutional Review Board (IRB) approval (#5250657) from a single academic Level I trauma center in San Bernardino County, Southern California. Electronic medical records were retrospectively reviewed for patients who underwent lumbar fusion surgery between January 1, 2013 and November 15, 2025.

Calculation of travel time and distance

Travel distance and travel time were quantified using the Google Maps Application Programming Interface. Patient home addresses and corresponding surgical center locations were extracted from the electronic medical record (EMR). To ensure consistency across measurements, a standardized retrospective date and time were entered into the platform’s departure time function for all route calculations. When a range of estimated travel times was generated, the upper and lower values were averaged to produce a single representative travel time estimate. Travel time was collected in addition to distance in order to assess both objective and perceived measures of travel burden under the assumption that patients traveled to the treating surgical center by car, as geographic and traffic-related factors can result in differing travel times even with similar travel distances.

Study population

Eligible patients included all adults aged 18–85 years who underwent lumbar fusion surgery at the academic center of interest involving levels L1 through L5 for degenerative or traumatic spine pathology. Surgical procedures were identified using International Classification of Diseases, Tenth Revision (ICD-10) and Current Procedural Terminology (CPT) codes, including anterior lumbar interbody fusion (ALIF), oblique lateral interbody fusion (OLIF), extreme lateral interbody fusion (XLIF), direct lateral interbody fusion (DLIF), transforaminal lumbar interbody fusion (TLIF), and posterior lumbar interbody fusion (PLIF). ICD-9 codes were utilized for the identification of procedures taking place prior to October 2015.

Patients were excluded if they were under 18 or over 85 years of age, had less than 1 year of outpatient follow-up, or had a history of spinal malignancy. Duplicate records and incomplete observations were identified and removed during data cleaning. The final analytic cohort consisted of 403 patients.

Study variables

Travel burden measures

The primary exposures of interest were travel time and travel distance from the patient’s home to the treating hospital. Travel distance (miles) and average travel time (minutes) were calculated using the patient’s home address recorded at the time of surgery and the surgical center location.

Demographic and clinical covariates

Patient demographic characteristics included age at surgery, sex, race/ethnicity, Social Vulnerability Index (SVI), and body mass index (BMI). Baseline comorbidities were extracted from the medical record and included chronic kidney disease, degenerative spine disease, diabetes mellitus, hyperlipidemia, history of cancer, pulmonary disease (including chronic obstructive pulmonary disease and obstructive sleep apnea), and cardiovascular or cerebrovascular disease (including coronary artery disease, prior coronary artery bypass grafting, coronary stent placement, cerebrovascular disease, transient ischemic attack, or stroke). Overall SVI scores were assigned by geocoding patients' ZIP codes at the time of surgery to their U.S. Census tract and matching them to the CDC’s 2022 SVI database [24]. SVI was analyzed as a continuous variable, with higher values indicating greater social vulnerability.

Postoperative outcomes

Postoperative outcomes were assessed using a combination of 90-day and 1-year follow-up windows, depending on variable availability. All medical complications (0–90 days) included sepsis, respiratory failure, pneumonia, pulmonary embolism/deep vein thrombosis (PE/DVT), myocardial infarction (MI), cerebrovascular accident/stroke, renal failure, neurogenic bladder or bowel dysfunction, urinary tract infection/urinary incontinence, and seizure. All surgical complications (0–1 year) included wound complications, surgical site infections, neurologic complications, and revision surgery/hardware removal. Additional outcome categories included cardiovascular/thromboembolic complications (0–90 days), defined as MI or PE/DVT; pulmonary complications (0–90 days), defined as respiratory failure or pneumonia; infectious complications (composite; mixed follow-up windows), defined as surgical site infection (0–1 year), sepsis (0–90 days), or urinary tract infection/urinary incontinence (0–90 days); and neurologic complications (composite; mixed follow-up windows), defined as neurologic complication (0–1 year), neurogenic bladder or bowel dysfunction (0–90 days), seizure (0–90 days), or cerebrovascular accident/stroke (0–90 days). Any complication was defined as a composite outcome including complications assessed at both 90 days and 1 year. Readmission outcomes included medical readmission (0–90 days), surgical readmission (0–1 year), and any readmission (composite of medical and surgical readmission across mixed follow-up windows). Revision/hardware removal (0–1 year) and wound complication (0–1 year) were also analyzed individually. Discharge disposition was dichotomized as home versus nonhome discharge disposition.

Statistical analysis

All statistical analyses were conducted using R version 4.5.1 (R Core Team) [25]. Statistical significance was defined as a 2-sided p-value <.05. Baseline demographic and clinical characteristics were summarized using means and standard deviations for continuous variables and counts with percentages for categorical variables. Binary variables were reported as event counts and percentages.

Associations between travel burden measures (travel time and travel distance) and binary postoperative outcomes, including Discharge disposition (nonhome vs. home), were assessed using multivariate Firth penalized logistic regression, adjusted for SVI, BMI, and age at the time of surgery. All assumptions for Firth Logistic regression were met. Length of stay was analyzed using multivariate linear regression adjusted for SVI, BMI, and age at the time of surgery. Because LOS exhibited right-skewness and non-normal residuals, LOS was log-transformed as log(LOS + 1) to improve model assumptions. All assumptions for linear regression (Linearity, homoscedasticity, and normality of residuals) were met.

In addition, a sensitivity analysis was conducted. Travel distance and travel time were categorized into quartiles, creating 4 groups with approximately equal numbers of participants. Q1 represents the lowest travel burden, and Q4 represents the highest travel burden.

Multivariable Firth penalized logistic regression models were then repeated using quartile-based travel distance and travel time predictors, with Q1 serving as the reference group, while adjusting for age at surgery, BMI at surgery, and overall SVI score. Length of stay was similarly evaluated using multivariable linear regression with quartile-based travel distance and travel time predictors, adjusting for the same covariates.

Baseline demographic characteristics and comorbidities across quartile cohorts were summarized using means and standard deviations for continuous variables and counts with percentages for categorical variables. To evaluate whether baseline characteristics differed across quartile cohorts, continuous variables were compared using Kruskal-Wallis tests, while categorical variables were compared using chi-square or Fisher’s exact tests, as appropriate.

The 3 covariates (age, BMI, and overall SVI) were selected a priori based on their clinical relevance to postoperative outcomes and their potential association with travel burden. Overall, SVI was added to account for neighborhood-level social vulnerability, which may influence both travel burden and postoperative recovery. Although additional demographic characteristics and comorbidities were collected, they were not included as covariates to preserve model stability, given the limited number of events for several outcomes. Expanding the covariate set would have led to unstable effect estimates Figure.

Figure.

Forest plots illustrating the adjusted analyses for: (A) Association between distance to hospital and postoperative outcomes; (B) Association between average time to hospital and postoperative outcomes; (C) Association between travel burden and log-transformed length of stay.

Results

Patient demographics and baseline characteristics

A total of 403 patients were included in the final analytic cohort (Table 1). The mean age at surgery was 51.42 ± 10.66 years, and the mean body mass index (BMI) was 30.51 ± 5.89. Patients traveled a mean distance of 44.35 ± 47.94 miles to the hospital, with an average travel time of 60.16 ± 50.09 minutes. All included patients underwent a minimum of 1 year of recorded outpatient postoperative follow-up.

Table 1.

Baseline patient characteristics.

General
Age	51.42 ± 10.66
BMI	30.51 ± 5.89
Distance to hospital (miles)	44.35 ± 47.94
Average time to hospital (min)	60.16 ± 50.09
Sex
Female	221 (54.8%)
Male	182 (45.2%)
Race/Ethnicity
Asian/Pacific Islander	8 (2%)
Black or African American	32 (7.9%)
Hispanic or Latino	101 (25.1%)
Other/Unknown/Refused	17 (4.2%)
White or Caucasian	245 (60.8%)
Cormorbidities
Renal disease	300 (74.4%)
Degenerative spine disease	262 (65%)
Diabetes mellitus	96 (23.8%)
Hyperlipidemia	137 (34%)
History of cancer	78 (19.4%)
Cardiovascular/cerebrovascular disease	239 (59.3%)
Pulmonary disease	169 (41.9%)

Total number of patients = 403.

The cohort was 54.8% female (n = 221) and 45.2% male (n = 182). Racial and ethnic composition was predominantly White or Caucasian (60.8%, n = 245), followed by Hispanic or Latino (25.1%, n = 101), Black or African American (7.9%, n = 32), Other/Unknown/Refused (4.2%, n = 17), and Asian/Pacific Islander (2.0%, n = 8). Common comorbidities included chronic kidney disease (74.4%), degenerative spine disease (65.0%), hyperlipidemia (34.0%), and diabetes mellitus (23.8%). Additionally, 59.3% of patients had a history of cardiovascular or cerebrovascular disease, and 41.9% had pulmonary disease.

The average travel time- and distance-based quartile analyses of patient demographics and comorbidities identified overall SVI score (p = <.001), race/ethnicity (p = <.001), renal disease (p = <.001), hyperlipidemia (p = .002), and history of cancer (p = .023) as potential confounding variables for travel time; and overall SVI score (p = <.001), race/ethnicity (p = <.001), and renal disease (p = <.001) as potential confounding variables for travel distance (Supplemental 1, 2).

Travel burden versus postoperative outcomes

Our adjusted analyses demonstrate an association between shorter travel time and distance and a greater likelihood of nonhome discharge (OR 0.9941, 95% CI 0.9875–0.9997; p = .04). This association reached statistical significance but was of very small magnitude and close to the null. Travel burden was not associated with any other postoperative outcomes (Table 2, Table 3). Distance to hospital and travel time were also not significantly associated with length of stay (Table 4). Three of the measured outcomes had a low events-per-variable (EPV) ratio (<5) in the multivariable model, indicating limited data relative to the number of model parameters. Thus, these outcomes may include wider uncertainty and should be interpreted with caution.

Table 2.

Association between distance to hospital and postoperative outcomes: Multivariable firth logistic regression adjusted for age, BMI, and overall SVI score.

Outcome	Events (n)	OR (per mile) (95% CI)	p-value
All medical complications	23 (5.8%)	1.00 (0.99–1.01)	.84
All surgical complications	70 (17.5%)	1.00 (0.99–1.00)	.87
Any complication	93 (23.2%)	1.00 (0.99–1.00)	.55
Cardiovascular/Thromboembolic complications*	8 (2.0%)	1.01 (1.00–1.02)	.14
Nonhome discharge disposition	101 (25.2%)	0.99 (0.99–1.00)	.04
Infectious complications	24 (6.0%)	1.00 (0.99–1.01)	.89
Neurologic complications	19 (4.8%)	0.99 (0.98–1.01)	.39
Pulmonary complications*	1 (0.2%)	1.02 (1.00–1.04)	.07
Medical readmission*	11 (2.8%)	0.99 (0.97–1.01)	.44
Surgical readmission	39 (9.8%)	1.00 (0.99–1.01)	.98
Any readmission	49 (12.2%)	1.00 (0.99–1.00)	.62
Revision/Hardware removal	37 (9.2%)	1.00 (0.99–1.01)	.86
Wound complication	20 (5.0%)	1.00 (0.99–1.01)	.75

Bolded symbol in the p-value column represents significant values (p<0.05).

All medical complications, cardiovascular/thromboembolic complications, pulmonary complications, and medical readmission included outcomes within 1 year of surgery. All other outcomes were measured at 90 days.

^⁎Represents outcomes with EPV < 5.

Table 3.

Association between average time to hospital and postoperative outcomes: Multivariable firth logistic regression adjusted for age, BMI, and overall SVI score.

Outcome	Events (n)	OR (per min) (95% CI)	p-value
All medical complications	23 (5.8%)	1.00 (0.99–1.01)	.82
All surgical complications	70 (17.5%)	1.00 (0.99–1.00)	.84
Any complication	93 (23.2%)	1.00 (0.99–1.00)	.53
Cardiovascular/Thromboembolic complications*	8 (2.0%)	1.01 (1.00–1.01)	.15
Nonhome discharge disposition	101 (25.2%)	0.99 (0.99–1.00)	.04
Infectious complications	24 (6.0%)	1.00 (0.99–1.01)	.75
Neurologic complications	19 (4.8%)	1.00 (0.98–1.00)	.39
Pulmonary complications*	1 (0.2%)	1.02 (1.00–1.05)	.07
Medical readmission*	11 (2.8%)	1.00 (0.98–1.01)	.63
Surgical readmission	39 (9.8%)	1.00 (0.99–1.01)	.86
Any readmission	49 (12.2%)	1.00 (0.99–1.00)	.61
Revision/Hardware removal	37 (9.2%)	1.00 (0.99–1.00)	.71
Wound complication	20 (5.0%)	1.00 (0.99–1.01)	.69

Bolded symbol in the p-value column represents significant values (p<0.05).

All medical complications, cardiovascular/thromboembolic complications, pulmonary complications, and medical readmission included outcomes within 1 year of surgery. All other outcomes were measured at 90 days.

^⁎Represents outcomes with EPV < 5.

Table 4.

Association between travel burden and log-transformed length of stay: Multivariate linear regression adjusted for age and BMI.

Outcome	Predictor	β (95% CI)	p-value
log(Length of stay [d] + 1)	Distance to hospital (per 10 miles)	−0.003 (−0.013 to 0.008)	.604
log(Length of stay [d] + 1)	Average travel time (per 30 min)	−0.006 (−0.036 to 0.023)	.686

Of our sensitivity analyses, the one based on travel time quartiles showed no statistically significant differences in all medical complications, all surgical complications, or overall complication rates across quartiles. Patients in Quartile 3 had lower odds of nonhome discharge compared to those in Quartile 1; however, this finding should be interpreted conservatively due to insufficient outcome events. Quartile 2 was associated with higher odds of surgical readmission within 1 year, but statistical significance trended downwards as quartile comparison became more extreme, suggesting a marginally meaningful difference (Supplemental 3). The analysis that utilized travel distance-based quartiles identified a higher risk of surgical complication (Quartile 4 vs. Quartile 1) and lower risk of nonhome discharge (Quartile 3 vs. Quartile 1, Quartile 4 vs. Quartile 1), but these outcomes also demonstrated insufficient EPV (Supplemental 4). Length of stay had no significant association with either travel time or travel distance across all quartile comparisons (Supplemental 5).

Discussion

This retrospective cohort study evaluated whether geographic travel burden (travel time and/or travel distance) was associated with postoperative outcomes following lumbar fusion at a high-volume academic Level I trauma center. Contrary to our hypothesis, patients with increased travel burden were not associated with postoperative complications, readmissions, discharge disposition, or length of stay compared to those with shorter travel time and distance. To our knowledge, this study is the first to specifically evaluate the association between geographic travel burden and postoperative outcomes following lumbar fusion surgery.

While these findings appear to conflict with portions of the existing broader medical literature suggesting a distance decay effect, they are consistent with prior orthopedic surgery research. A prior systematic review by Kelly et al. included 108 studies across a variety of medical specialties and determined mixed results, with 77% of studies showing evidence of distance decay. Nearly 6% identified the reverse effect of distance bias, and the remaining 17% of studies determined no relationship between distance and outcomes [7]. Another study in elective colorectal surgery found a positive association between travel distance and postoperative LOS and suggested that travel distance be considered as a risk adjustment factor [20]. Conversely, in orthopedic literature, a retrospective review of nearly 39,000 patients undergoing total joint arthroplasty determined no significant increase in complication risk with higher patient travel distance [26]. Similarly, a study by Mehta et al. [23] found that a travel distance of greater than 50 miles had no effect on patient-reported functional outcomes following total shoulder arthroplasty. These results were recently corroborated with additional patient-reported outcomes by Jones et al. [27] in a single-center study of 413 patients, which again determined no differences in either outcomes or patient-reported outcomes. Collectively, these studies suggest that, within structured referral networks and high-volume centers, increased travel time and distance alone may not adversely impact early postoperative outcomes in these study populations.

A notable finding is the association between shorter travel time and distance and increased likelihood of nonhome discharge, despite its marginal statistical significance. Given the well-described relationship between geographic barriers to healthcare and low socioeconomic status, our findings suggest that discharge disposition may be mediated by factors other than travel burden alone and underscore the complex interplay between social determinants of health and patient care [[4], [5], [6]]. In a study evaluating predictors of postoperative discharge disposition, the Risk Assessment and Predictor Tool (RAPT), which takes into account age, gender, use of walking aid, ambulatory status, community support, and caregiver dependence, was found to accurately predict discharge disposition based on risk stratification in TJA and spinal fusion patients [28]. A survey study of primary TJA patients found that nearly 85% of patients discharged to home had caregiver support compared to less than 45% of patients discharged to skilled nursing facilities [29]. Considering the adjustment of our analyses for age, BMI, and Social Vulnerability Index score, travel burden alone appears less significant of a predictor of discharge disposition than other factors such as caregiver presence and functional dependence.

In terms of clinical application, the absence of significant associations found between travel burden and most postoperative outcomes in this population may serve to both reassure patients undergoing lumbar fusion who must travel to access specialized surgical care, as well as persuade surgeons who advocate against transfer to academic centers due to concerns about the negative impact that increased travel burden may have on surgical outcomes due to factors such as barriers to follow-up compliance and family access. At our institution, postoperative management following lumbar fusion typically involves a standardized follow-up scheduled with the surgical team at specific intervals (eg 2-week, 6-week, and 12-week timepoints). While routine wound checks or basic medical concerns can occasionally be coordinated with local primary care providers for patients living at great distances, the specialized nature of spine surgery, specifically the need for nuanced neurologic examination and interpretation of early postoperative imaging, often necessitates evaluation by the operative surgical team. Our findings suggest that even when patients must return to the primary surgical center for this specialized care, the geographic burden does not inherently translate to inferior clinical outcomes.

A potential explanation for the lack of worsened outcomes seen in our study population is the status of our chosen surgical institution as a high-volume academic medical center (AMC). Prior work by Birkmeyer et al. has demonstrated that high-volume AMCs have superior resources and staffing to smaller centers, and describes both surgeon and hospital volume as predictors of postoperative mortality following major surgery [30,31]. Additionally, a retrospective review of more than 18,000 patients who underwent rectal cancer resection found that patients who traveled a longer distance to a high-volume center had better perioperative characteristics and improved 5-year survival compared to those who were treated at local low-volume centers [32]. Ghaferi et al. [33] determined that low-volume hospitals experienced minimally higher postoperative complication risk but significantly higher failure-to-rescue rates during high-risk surgery when compared to high-volume hospitals. Together, these findings suggest that the superior surgeon and hospital volume, advanced management of perioperative complications, and improved institutional resources enjoyed by high-volume AMCs may mitigate adverse effects associated with longer travel time and distance.

Limitations

This study has several limitations. Firstly, this was conducted as a single-center study with postoperative events limited to the institutional medical record and within 1 year of surgery. Therefore, complications, readmissions, or follow-up occurring at outside institutions or outside the specified time frame may not have been fully captured, potentially leading to deflated outcomes rates, specifically in patients with greater travel distances. Also, the determination of travel time relied on the assumption that patients traveled to the surgical center by car rather than any alternate forms of transportation, which may not have accurately captured true travel time. Furthermore, the retrospective design of the study limits the ability to establish a relationship between travel burden and postoperative outcomes, introducing the potential for possible unmeasured confounding such as socioeconomic status, caregiver support, health literacy, etc. Additionally, several of our multivariable regression analyses were limited by low events per variable (EPV < 5), which may have increased the risk of bias and imprecision. We utilized Firth's penalized likelihood regression to mitigate small-sample bias and address potential separation, but this method still does not completely eliminate concerns about estimation accuracy. Thus, the findings regarding the outcomes with insufficient EPV should be regarded as descriptive and be used to generate hypotheses for future investigation, rather than understood as definitive. Finally, our inclusion of covariates in the multivariate regression analyses utilized in this study were limited by our study’s single-institution sample size. Age, BMI, and overall SVI score were ultimately designated as covariates in the final analysis in order to control for the maximum number of relevant confounding factors while preserving the validity of our results. Future studies with larger sample sizes should incorporate additional baseline characteristics and comorbidities that differed across travel time and distance quartiles, including race/ethnicity, renal disease, hyperlipidemia, and history of cancer, to better account for potential residual confounding as seen in Supplemental 1 and Supplemental 2. Further studies with more outcome events and larger sample sizes are warranted to validate any associations and provide more precise estimates of effect.

Conclusion

Increased travel distance and travel time were not independently associated with postoperative complications, readmissions, discharge disposition, or length of stay following lumbar fusion. These findings do not support a distance decay association in this spine population and instead suggest that, within a high-volume academic referral center, travel burden alone may not adversely influence short-term surgical outcomes. Collectively, our results indicate that geographic travel burden may influence access to care or referral patterns more than postoperative recovery once treatment is delivered. Larger, multicenter studies incorporating socioeconomic variables and long-term follow-up are warranted to better define the interaction between travel burden, healthcare access, and outcomes after lumbar fusion.

Declaration of competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.