Enhanced recovery after surgery (ERAS ) programs are increasingly utilized across multiple clinical domains to standardize care in the perioperative period.1,2 ERAS programs have consistently demonstrated a reduction in complications after surgery, shorter hospital lengths of stay and a commensurate reduction in hospital costs.3 ERAS were initially demonstrated after colorectal and abdominal surgery, neurological-specific procedures including spine and intracranial surgery have demonstrated similar benefits after implementation of these programs.4,5
Successful development, implementation, and maintenance of ERAS programs requires an investment in both hospital resources and ERAS personnel.6 ERAS program. On the basis of specific institutional needs, this may involve the need to employ several part-time and/or full-time ERAS coordinators and staff.6
A recent systematic review of ERAS programs for spine surgery reported significant reductions in opioid consumption, length of hospital stay (LOS) and cost. However, the authors noted significant variability in the components of ERAS protocols and the reported outcomes among the different studies included in the review.7 ERAS programs only by monetizing the reductions in surgical-related complications or LOS.8–10 ERAS when accounting for the costs associated with establishing and operationalizing ERAS programs and including a natural unit of outcome, such as hospital LOS. By calculating an incremental cost effectiveness ratio (cost/utility) and using a predefined “willingness to pay” threshold value, health care decision makers can determine whether ERAS programs are cost-effective for their institution.
The aim of this study was to model the incremental cost-effectiveness of a single-center, spine surgery ERAS program. The cost effectiveness of the ERAS program was evaluated using both patient-level hospital costs and estimated costs of personnel required to develop and maintain the program, with LOS as the utility of interest.
METHODS
Approval for this project was provided by the University of Virginia Institutional Review Board (UVA IRB HSR # 22367, 5/14/2020); the requirement for written consent was waived by the institutional review board. We followed the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines for reporting of observational data and Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement for the design and reporting of cost-effectiveness studies throughout this study.11,12
Study Design and Sample
We performed a single center, retrospective, cohort study at the University of Virginia before and following implementation of an ERAS program for spine surgery . Propensity matching was performed between ERAS and standard care groups to create a quasi-experimental design.
Adult patients (18 y of age or above) undergoing cervical (anterior or posterior), thoracic (posterior) or lumbar (anterior or posterior) spine surgery were included for analysis, before (April 2018 to March 2019) and after (October 2019 to May 2020) implementation of the ERAS program. ERAS cases were not included in the first 7 months after introduction of the program because of the phased implementation of the program, and to exclude any iterative changes made in the early phase of the program. ERAS cases were limited to procedures involving 6 or fewer levels of surgery and included both fusion and nonfusion cases. Cases requiring intensive care unit admission after surgery were not included in the ERAS pathway; generally, patients with multiple comorbidities, anticipated intraoperative blood loss >1500 mL and those requiring more than 7 levels of spinal fusion were scheduled for postoperative intensive care unit admission. Missing cases were excluded from the analysis. No a priori power calculation was performed as our study sample included all available complete data in the database.
The protocol used for the ERAS pathway is reported in Supplemental Digital Content 1 (https://links.lww.com/JNA/A473 ERAS group, patients receiving standard care had no oral intake 8 to 12 hours before surgery, preoperative multimodal analgesia (acetaminophen, gabapentin, and celecoxib) was not administered, and intraoperative/postoperative analgesia was at the discretion of the anesthesia and surgical team, respectively. No restrictions were placed on intravenous opioids, including the use of patient controlled intravenous analgesia.
Study Variables
Age, sex, race, weight, height, American Society of Anesthesiologist (ASA) physical status, preoperative composite comorbidity (hypertension or smoking or obstructive sleep apnea or coronary artery disease or diabetes or chronic obstructive airway disease or hyperlipidemia or chronic kidney disease), surgery location and type, anesthesia duration, total intraoperative morphine equivalent, total intraoperative fluid balance, total postoperative morphine equivalent, 11-point numerical pain rating (0 to 10), hospital LOS, composite postoperative complications within 30 days of the index admission (readmission, surgical site infection, new neurological deficits postoperatively, stroke, myocardial infarction, and pulmonary embolus). All data were extracted from the hospital electronic medical record system and the data registry compiled for the ERAS program.
Outcomes
Primary Outcome
The primary outcome was incremental cost effectiveness analysis of the ERAS program, modeling patient cost data, cost of personnel and an effectiveness utility.
Patient Cost Data
Individual patient-level data was matched to hospital finance records. Total hospital cost with an associated cost valuation derived from both direct and indirect costs was totaled for an entire admission. Hospital costs were adjusted to 2018 equivalent United States dollars using the Centers for Medicare and Medicaid Services Inpatient Prospective Payment System adjustment for medical-related inflation. Total hospital costs were calculated for the index hospitalization; however, if a patient was readmitted, the additional cost of readmission was included in the total cost.
Cost of Personnel to Develop and Support ERAS Program
Yearly total institutional personnel costs of the multispecialty ERAS program, which included 2 medical co-directors, one ERAS nurse leader, 1 pain management advanced practice nurse, 1 data analyst, and 1 nurse coordinator, were calculated. A single, fixed personnel cost to develop and implement the program during the year before implementation was estimated. This accounted for 40% of ERAS personnel effort for literature review, electronic medical system order set build and modification, staff education, and implementation sessions. This cost was averaged over a 5-year horizon and then standardized to a single patient cost by dividing the estimated total number of patients over the 5-year period. The estimated total number of cases over the 5-year period was calculated by projecting the yearly case volume during the study period with no adjustment for additional growth in the program.
The yearly ongoing cost attributed to the spine surgery ERAS program was calculated as the total ERAS cost divided equally by the number of ERAS programs at our institution; during the study period there were 11 additional ERAS programs. Single patient cost was then calculated by dividing yearly ERAS spine cost by the number of patients enrolled in the program for that particular year.
Effectiveness Utility
Hospital LOS was reported as the utility of interest. LOS was calculated from the day of surgery. Discharge on the day of surgery (defined as discharge <24 h after admission for surgery) was modeled as a 0.5 LOS. To perform an incremental cost effectiveness analysis , a weighted LOS utility was calculated (mean LOS/LOS); with this weighted LOS, shorter LOS had a higher utility while longer LOS had a lower utility.
Statistical Analysis
Continuous variables are presented as median and interquartile range and categoric variables as number (%). All cost and utility data are presented as mean±SD. Univariate analysis was initially performed between the ERAS and non-ERAS (standard care) groups. Propensity score analysis was performed between the ERAS and non-ERAS groups to control for the imbalances in baseline characteristics using a 3-step procedure: development of propensity score, matching, and assessment imbalance. Using logistic regression, the predicted probability of being in an ERAS program, the propensity score, was calculated. Factors included in the logistic regression model included age, body mass index, sex, race, ASA physical status, anesthesia duration, and procedure type. Matching was performed using a greedy matching algorithm without replacement and a caliper distance 0.03. C statistic for the logistic model was calculated. Numeric and visual standardized mean difference (SMD) for continuous variables were evaluated to assess the balance after matching. A SMD of <0.2 is considered a small effect size. After matching, groups were compared by using a paired univariate analysis. Categoric variables were compared using the McNemar test, while continuous variables were compared by the signed rank test or paired t test for nonparametric and parametric data, respectively. All statistical analysis was performed in SAS 9.4 software (SAS Institute Inc., Cary, NC) and R (version 4.0.3; R Foundation for Statistical Computing, Austria. Available at: www.r-project.org/
RESULTS
Four hundred and twelve patients were initially included in the analysis from 2018 to 2020. Three patients were removed because of duplication and missing data, leaving 409 patients for inclusion in the final analysis. Demographic, intraoperative, and postoperative variables in the unmatched groups are shown in Table 1 . Body mass index, procedure type, anesthesia duration, total morphine equivalents in the operating room, LOS and readmission demonstrate a SMD >0.2. Of note, there was imbalance in the median ([interquartile range] LOS between the 2 groups (standard care, 2.0 [1.0, 5.0] d vs. ERAS , 5.0 [3.0, 6.0] d; SMD, 0.36). After propensity matching, the only imbalance between the groups, defined by a SMD >0.2, were race (SMD = 0.21), ASA physical status (SMD = 0.32) and fluid balance (SMD = 0.21) in the operating room (Table 2 ). Interestingly after matching, differences between groups were larger for length of stay (Standard care: Median [IQR] 2.0 [1.0, 3.75], ERAS : 4.0 [3.0, 5.0], SMD=0.81) with both higher charges (Standard care: Mean (SD) 64,503 + 37,172.02, ERAS : Mean (SD) $81, 956.8 +94, 985.9, SMD = 0.24) and cost (Standard care: Mean (SD) $19,291.57 + 13,572.24, ERAS : $24,363.45 + 26352.45, SMD = 0.24). The SMD before and after propensity matching is demonstrated in Supplemental Digital Content 2 (https://links.lww.com/JNA/A474
TABLE 1 -
Demographic, Intraoperative, and Postoperative Variables in the Standard Care Versus
ERAS Groups Before Matching
Variables
Standard Care (N=283)
ERAS (N=126)SMD
Age
59.3 (13.3)
57.6 (13.1)
0.12
Male sex
154 (54.4)
73 (57.9)
0.07
Race
White
244 (86.2)
113 (89.7)
0.19
Black
33 (11.7)
9 (7.1)
Asian
1 (0.4)
1 (0.8)
American Indian
1 (0.4)
0
Other
4 (1.4)
3 (2.4)
BMI (kg/m2 )
30.7 (7.2)
32.5 (6.16)
0.27
Composite comorbidity
91 (32.2)
34 (27.0)
0.11
ASA physical status
0.15
I
3 (1.1)
3 (2.4)
II
133 (47.0)
63 (50.0)
III
142 (50.2)
59 (46.8)
IV
5 (1.8)
1 (0.8)
Procedure type
0.63
Anterior cervical 1 level fusion
28 (9.9)
5 (4.0)
Anterior cervical multilevel fusion
20 (7.1)
22 (17.5)
Posterior cervical 1 level fusion
1 (0.4)
1 (0.8)
Posterior cervical 1 level, no fusion
41 (14.5)
11 (8.7)
Posterior lumbar 1 level fusion
33 (11.7)
23 (18.3)
Posterior lumbar 1 level, no fusion
94 (33.2)
30 (23.8)
Posterior lumbar 2-3 level, no fusion
3 (1.1)
0
Posterior lumbar 3-6 level fusion
60 (21.2)
24 (19.0)
Anterior lumbar 1 level fusion
2 (0.7)
5 (4.0)
Thoracic 1 level fusion
1 (0.4)
2 (1.6)
Anesthesia duration (minutes)
264.9 (98.0)
236.6 (99.1)
0.29
Total ME in OR (mg)
5.0 (0.0, 10.0)
5.3 (3.2, 10.0)
0.21
Fluid balance in the OR (mL)
815.0 (500.0, 1250.0)
800.0 (550.0, 1165.0)
0.08
Highest pain rating postoperatively
9.0 (7.0, 10.0)
8.0 (7.0, 10.0)
0.04
Total ME postoperatively (mg)
35.00 (18.33, 69.00)
32.50 (20.00, 64.25)
0.08
Composite complication
24 (8.5)
7 (5.6)
0.12
Length of stay (d)
2.0 (1.0, 5.0)
5.0 (3.0, 6.0)
0.36
Readmission
24 (9.3)
7 (5.9)
0.41
Total charges (US dollars)
92,553.82±79,085.77
89,471.19±77,226.26
0.04
Total cost (US dollars)
27,600.16±24,727.73
27,778.42±22,744.50
0.01
Data shown as mean±SD, median (interquartile range), or n (%).
ASA indicates American Society of Anesthesiologists; BMI, body mass index; ERAS , enhanced recovery after surgery ; ME, morphine equivalents; OR, operating room; SMD, standardized mean difference.
TABLE 2 -
Demographic, Intraoperative, and Postoperative Variables in the Standard Care Versus
ERAS Group After Matching
Variables
Standard Care (N=54)
ERAS (N=54)SMD
P
Age
59.4 (12.4)
57.9 (14.6)
0.10
0.59
Male sex
31 (57.4)
31 (57.4)
<0.001
1.0
Race
0.21
0.77
White
48 (88.9)
48 (88.9)
Black
4 (7.4)
3 (5.6)
Asian
0
1 (1.9)
Other
2 (3.7)
2 (3.7)
BMI (kg/m2 )
30.7 (5.2)
32.5 (6.2)
0.08
0.67
Composite comorbidity
19 (35.2)
14 (25.9)
0.20
0.40
ASA physical status
0.32
0.27
I
0
2 (3.7)
II
33 (61.1)
28 (51.9)
III
21 (38.9)
24 (44.4)
Procedure type
<0.001
1.0
Anterior cervical 1 level fusion
1 (1.9)
1 (1.9)
Anterior cervical multilevel fusion
4 (7.4)
4 (7.4)
Posterior cervical 1 level, no fusion
7 (13.0)
7 (13.0)
Posterior lumbar 1 level fusion
11 (20.4)
11 (20.4)
Posterior lumbar 1 level, no fusion
25 (46.3)
25 (46.3)
Posterior lumbar 3-6 level fusion
6 (11.1)
6 (11.1)
Anesthesia duration (minutes)
230.5 (75.6)
240.8 (87.7)
0.13
0.52
Total ME in OR (mg)
5.0 (0.0, 7.6)
5.0 (2.7, 10.0)
0.20
0.24
Fluid balance in the OR (mL)
687.5 (492.5, 997.5)
750.0 (532.5, 1002.5)
0.21
0.41
Highest pain rating postoperatively
8.0 (7.0, 10.0)
8.0 (7.0, 10.0)
0.18
0.88
Total ME postoperatively (mg)
26.3 (7.1, 52.9)
32.1 (20.2, 65.0)
0.17
0.08
Composite complication
6 (11.1)
3 (5.6)
0.20
0.49
Length of stay (d)
2.0 (1.0, 3.75)
4.0 (3.0, 5.0)
0.81
<0.001
Readmission
6 (12.5)
3 (5.9)
0.48
Total charges (dollars)
64,503±37,172.02
81,956.8±94,985.9
0.24
0.30
Total cost (dollars)
19,291.57±13,572.24
24,363.45±26,352.45
0.24
0.08
Data shown as mean±SD, median (interquartile range), or n (%).
ASA indicates American Society of Anesthesiologists; BMI, body mass index; ERAS , enhanced recovery after surgery ; ME, morphine equivalents; OR, operating room; SMD, standardized mean difference.
Cost-effectiveness Estimations
A simple model was developed to analyze the incremental cost effectiveness of the ERAS program on LOS in a cohort of patients undergoing spine surgery Supplemental Digital Content 3 (https://links.lww.com/JNA/A475 ERAS arm. The utility for the incremental cost effectiveness analysis was the weighted LOS. A 2-way sensitivity analyses was performed by allowing the model to range between varying weighted LOS utilities. An a priori willingness to pay value of $50,000 was chosen based on a previous cost-utility analysis.13
Cost-effectiveness Modeling
The cost-effectiveness of the ERAS program was modeled using the estimates of cost and LOS derived from our propensity matched cohort, as described above. The TreeAge cost-effectiveness model which was utilized for this analysis is shown in Supplemental Digital Content 3 (https://links.lww.com/JNA/A475 ERAS protocol is dominated by the standard care protocol. In a 1-way sensitivity analysis for net monetary benefit, the standard care protocol dominated the ERAS group across a broad range of ERAS patient cost (Fig. 1 ). The 2-way sensitivity analysis varying both ERAS cost and ERAS utility (weighted LOS) is shown in Figure 2 ; this demonstrates that the standard care protocol dominates the ERAS protocol across a broad range of cost and utility. ERAS only becomes cost effective when the ERAS utility is >2.2.
FIGURE 1: One-way sensitivity analysis demonstrates that the model is not sensitive to any variation in the cost of the enhanced recovery after surgery (ERAS ) program. Willingness to pay=$50,000.
FIGURE 2: Two-way sensitivity analysis varying both the cost of the enhanced recovery after surgery (ERAS ) program and the weighted length of stay demonstrates that standard care is cost-effective at most estimates. Blue indicates when ERAS is more cost-effective, while red indicates when standard care is more cost-effective. The 95% confidence intervals were used to vary the axes. Willingness to pay=$50,000.
DISCUSSION
In this single-center, cost-effectiveness analysis in patients undergoing spine surgery , standard care was more cost-effective than an ERAS program when factoring the cost of developing and operationalizing the ERAS program. The primary reason for the difference in cost-effectiveness in our study was the longer LOS in the ERAS group and its associated higher cost. The longer LOS following implementation of the ERAS program was evident in both the unmatched and matched cohorts when accounting for preoperative and intraoperative predictors likely to increase LOS. This is unusual as most studies have reported a reduction in LOS utility with ERAS . One possible reason for the longer LOS in the ERAS group in our study was the presence of unmeasured confounders in the study cohort. However, we used a robust statistical tool, propensity score matching, to create a quasi-experimental design to equate the standard care and ERAS groups. We achieved good balance between the groups after matching, suggesting that confounding and bias were appropriately addressed and that the identified imbalance in LOS was a real phenomenon. In addition, practice changes introduced by the ERAS program, such as a requirement for ambulation before discharge, might have accounted for the longer LOS and were not controlled for in our model. Before introduction of the ERAS program, no standardized postoperative ambulation targets were needed before discharge, whereas in the ERAS program patients were required to ambulate with the supervision of a physiotherapist before discharge. There was also a higher incidence of ASA III patients in the ERAS group (44.4%) compared with the standard care group (38.9%), which might also have contributed to the increased LOS in the ERAS group. However, it is interesting to note that there was a slightly higher complication rate in the standard care group (11.1%) compared with the ERAS group (5.6%), which did not translate into longer LOS.
The findings of our study are in contrast to those of other studies and a meta-analysis, which demonstrated reductions in cost and postoperative complications, and an improvement in quality of life utilities, associated with a broad range of ERAS programs.14,15 16 P <0.001) for ERAS compared with standard care, respectively, and that the ERAS program was associated with a shorter LOS (ERAS , 1.23±0.8 d vs. standard care, 3.9±1.1 d; P =0.009); in that study the benefits of the ERAS program were monetized by the difference in total cost between the standard of care and ERAS . Similarly, Tarikci Kilic et al17 ERAS programs for simple, nonfusion spine surgery .
The major limitations of the aforementioned studies that purport cost benefits of ERAS programs are that the cost of implementing and maintaining the program were not included in a comprehensive cost-effectiveness analysis. Implementation costs of new ERAS programs include factors such as site visits, training courses, surgeon/anesthesiology/nursing champions, ERAS staff, education materials and data analysis. Stone et al6 ERAS program at a quaternary institution in the United States at ∼$552,783, with the cost per patient within the first year varying between $1179 and $1106 based on enrolling 100 or 500 patients, respectively, into the program. The cost per patient after the first year of implementation of the ERAS program decreased to $1079 and $714 for 100 or 500 patients, respectively. In their sensitivity analyses, Stone et al6 ERAS programs when factoring in the cost of implementation and maintenance are only realized with progressive reductions in LOS and improved cost savings. In our study, LOS was longer in the ERAS program and associated with higher cost, therefore making standard care the dominant strategy across a broad range of cost and LOS.
Our study has several limitations which may limit its broader applicability. The study was confined to a single, quaternary academic center in the United States, with the financial model for cost and reimbursement difficult to extrapolate to other institutions, regions or countries. We utilized the mean cost and LOS in this analysis, which is more sensitive to outliers. We choose this approach because the majority of cost-effectiveness analyses performed in the United States are based on a parametric assumption of cost and LOS. This was a retrospective study with potential unmeasured confounders not accounted for in our propensity matching. Furthermore, estimations regarding potential numbers of patients and the 5-year time horizon were made for the analysis. We used only LOS as the utility of interest. Patient reported outcomes such as the Oswestry Disability Index or the Patient Reported Outcome Measurement Information System measure were not reported in this study and may have yielded a different incremental cost-effectiveness analysis. Finally, chronic pain defined by preoperative opioid use was not reported in the analysis. However, it is important to note that in both the unadjusted and adjusted cohorts, total morphine equivalent in the operating room, highest pain rating after surgery and total morphine equivalent after surgery were not different between the standard care and ERAS groups.
CONCLUSIONS
In summary, we report a real-world, cost-effectiveness analysis following implementation of an ERAS program for spine surgery at a quaternary medical center in the United States. Our study demonstrated that considering LOS as the sole determinant, standard care is the dominant cost-effective strategy compared with the ERAS protocol.
REFERENCES
1. Varadhan KK, Neal KR, Dejong CH, et al. The
enhanced recovery after surgery (
ERAS ) pathway for patients undergoing major elective open colorectal surgery: a meta-analysis of randomized controlled trials. Clin Nutr. 2010;29:434–440. doi:10.1016/j.clnu.2010.01.004
2. Zhang X, Yang J, Chen X, et al.
Enhanced recovery after surgery on multiple clinical outcomes: Umbrella review of systematic reviews and meta-analyses. Medicine (Baltimore). 2020;99:e20983. doi:10.1097/MD.0000000000020983
3. Noba L, Rodgers S, Chandler C, et al.
Enhanced recovery after surgery (
ERAS ) reduces hospital costs and improve clinical outcomes in liver surgery: a systematic review and meta-analysis. J Gastrointest Surg. 2020;24:918–932. doi:10.1007/s11605-019-04499-0
4. Dietz N, Sharma M, Adams S, et al.
Enhanced recovery after surgery (
ERAS ) for
spine surgery : a systematic review. World Neurosurg. 2019;130:415–426. doi:10.1016/j.wneu.2019.06.181
5. Lampilas A, Bouyer B, Ferrero E, et al. Evaluation of enhanced recovery after
spine surgery : specificities in an academic public hospital. Orthop Traumatol Surg Res. 2021;107:103027. doi:10.1016/j.otsr.2021.103027
6. Stone AB, Grant MC, Pio Roda C, et al. Implementation costs of an
enhanced recovery after surgery program in the United States: a financial model and sensitivity analysis based on experiences at a Quaternary Academic Medical Center. J Am Coll Surg. 2016;222:219–225. doi:10.1016/j.jamcollsurg.2015.11.021
7. Tong Y, Fernandez L, Bendo JA, et al.
Enhanced recovery after surgery trends in adult
spine surgery : a systematic review. Int J Spine Surg. 2020;14:623–640. doi:10.14444/7083
8. Luzuy-Guarnero V, Gronnier C, Figuereido S, et al. Cost-benefit analysis of an enhanced recovery program for gastrectomy: a retrospective controlled analysis. World J Surg. 2021;45:3249–3257. doi:10.1007/s00268-021-06220-2
9. Martin LW, Sarosiek BM, Harrison MA, et al. Implementing a thoracic enhanced recovery program: lessons learned in the first year. Ann Thorac Surg. 2018;105:1597–1604. doi:10.1016/j.athoracsur.2018.01.080
10. Thiele RH, Rea KM, Turrentine FE, et al. Standardization of care: impact of an enhanced recovery protocol on length of stay, complications, and direct costs after colorectal surgery. J Am Coll Surg. 2015;220:430–443. doi:10.1016/j.jamcollsurg.2014.12.042
11. von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med. 2007;147:573–577. doi:10.7326/0003-4819-147-8-200710160-00010
12. Husereau D, Drummond M, Petrou S, et al. Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement. BMJ. 2013;346:f1049. doi:10.1136/bmj.f1049
13. Muurinen J-M. Demand for health: a generalised Grossman model. J Health Econ. 1982;1:5–28. doi:10.1016/0167-6296(82)90019-4
14. Ren Y, Sun D, Pei L, et al. A full
enhanced recovery after surgery program in gynecologic laparoscopic procedures: a randomized controlled trial. J Minim Invasive Gynecol. 2021;28:1610.e6–1617.e6. doi:10.1016/j.jmig.2021.01.024
15. Wang LH, Zhu RF, Gao C, et al. Application of enhanced recovery after gastric cancer surgery: an updated meta-analysis. World J Gastroenterol. 2018;24:1562–1578. doi:10.3748/wjg.v24.i14.1562
16. Wang MY, Chang HK, Grossman J. Reduced acute care costs with the
ERAS (R) minimally invasive transforaminal lumbar interbody fusion compared with conventional minimally invasive transforaminal lumbar interbody fusion. Neurosurgery. 2018;83:827–834. doi:10.1093/neuros/nyx400
17. Tarikci Kilic E, Demirbilek T, Naderi S. Does an
enhanced recovery after surgery protocol change costs and outcomes of single-level lumbar microdiscectomy? Neurosurg Focus. 2019;46:E10. doi:10.3171/2019.1.FOCUS18665
