Surgeon-Level Variability in Outcomes, Cost, and Comorbidity Adjusted-Cost for Elective Lumbar Decompression and Fusion

Surgeon-Level Variability in Outcomes, Cost, and Comorbidity Adjusted-Cost for Elective Lumbar... Abstract BACKGROUND The costs and outcomes following degenerative spine surgery may vary from surgeon to surgeon. Patient factors such as comorbidities may increase the health care cost. These variations are not well studied. OBJECTIVE To understand the variation in outcomes, costs, and comorbidity-adjusted cost for surgeons performing lumbar laminectomy and fusions surgery. METHODS A total of 752 patients undergoing laminectomy and fusion, performed by 7 surgeons, were analyzed. Patient-reported outcomes and 90-d cost were analyzed. Multivariate regression model was built for high-cost surgery. A separate linear regression model was built to derive comorbidity-adjusted 90-d costs. RESULTS No significant differences in improvement were found across all the patient-reported outcomes, complications, and readmission among the surgeons. In multivariable model, surgeons #4 (P < .0001) and #6 (P = .002) had higher odds of performing high-cost fusion surgery. The comorbidity-adjusted costs were higher than the actual 90-d costs for surgeons #1 (P = .08), #3 (P = .002), #5 (P < .0001), and #7 (P < .0001), whereas they were lower than the actual costs for surgeons #2 (P = .128), #4 (P < .0001), and #6 (P = .44). CONCLUSION Our study provides valuable insight into variations in 90-d costs among the surgeons performing elective lumbar laminectomy and fusion at a single institution. Specific surgeons were found to have greater odds of performing high-cost surgeries. Adjusting for preoperative comorbidities, however, led to costs that were higher than the actual costs for certain surgeons and lower than the actual costs for others. Patients’ preoperative comorbidities must be accounted for when crafting value-based payment models. Furthermore, designing intervention targeting “modifiable” factors tied to the way the surgeons practice may increase the overall value of spine care. Spine, 90-d cost, Outcomes, Surgeon variability, Bundled payments ABBREVIATIONS ABBREVIATIONS ASA American Society of Anesthesiologists AUC area under the curve BMI body mass index CI confidence interval CMS Centers for Medicare and Medicaid Services CPT current procedural terminology DRG diagnosis-related group EQ-5D EuroQol-5D NRS numeric rating scale ODI Oswestry Disability Index OR odds ratio PRO patient-reported outcome US United States The current trajectory of the health care expenditure in the United States (US) is unsustainable. According to the Centers for Medicare and Medicaid Services (CMS), projected health spending in the US will be as high as 20% of the gross domestic product by 2021. Low back pain associated with lumbar degenerative pathologies is highly prevalent and an economic burden.1-5 It is important to investigate costs and to examine the real-world clinical and cost benefit of commonly performed spinal surgeries. In an effort to curb escalating costs, CMS has proposed a pay-for-performance episode-based bundled payment model for hospital and physician reimbursement. The current fee-for-service model, whereby the Medicate reimbursement to the hospital is based on discounted rates payment rates established under the Inpatient Prospective Payment System, is considered a potential source of increasing health care spending. With the proposed bundled payment initiative, the individual physician, hospital, and other providers will be accountable for the quality of care and associated costs from surgery through 90 d after discharge.6-8 This model mandates all stakeholders to collaborate and investigate optimal strategies for arriving at a target price for a bundled payment. To determine a sustainable bundled cost, it is important to understand the biggest contributors to and variability in the cost at each level of patient care. The inpatient hospital cost, cost associated with readmission following surgery, and surgeon's professional fee cost are the 3 most important contributors of the cost for spine surgery.8-11 A number of previous studies have focused on the variations in hospital cost and total cost during the 90-d postdischarge period.8,9,11,12 Less well studied is the individual surgeon variation in cost, specifically from the same institution and given fairly standardized techniques. Variation in cost, if extant, provides an opportunity to understand the differences and learn from the better practice patterns. Given the wide variation in patient profile, disease process, surgical techniques, and practice patterns, it is prudent to assess the cost variability and to determine the factors driving high cost surgery. Furthermore, it is generally believed that the sicker patients account for higher cost associated with spine care.13-20 It is vital to understand the variation in cost associated with comorbidities. In this regard, the aims of this study were to define variability in total 90-d cost among the surgeons, to define factors associated with high-cost surgery and to determine the difference in total 90-d cost and cost adjusted for comorbidities for patients undergoing laminectomy and fusion surgery for degenerative spine diseases. METHODS Patients undergoing elective decompression and fusion surgery for degenerative spine pathology between 2011 and 2015 at a single comprehensive spine center were enrolled into a single-center prospective longitudinal spine registry. A retrospective review of prospectively collected data was conducted. An approval for the study and wavier of informed consent was obtained from the institutional review board for all the patients entered into the registry. The inclusion criteria were (1) patients age >18 yr; (2) presenting with leg and/or back pain; (3) the correlative imaging findings for the diagnosis of disc herniation, stenosis, and spondylolisthesis; and (4) failed 3-mo of multimodal nonoperative care or patients with progressive neurological deficit. The exclusion criteria were (1) pathological spine disease including tumor, infection, and trauma; (2) any extra-spinal cause of back or leg pain; (3); patients who were unable or unwilling to complete the follow-up questionnaire. Patients operated on by 7 surgeons, who are participating in the registry, were analyzed. Patient demographics, comorbidities (diabetes, hypertension, coronary artery disease, myocardial infarction, preoperative anticoagulation, congestive heart failure, chronic obstructive pulmonary disease, and osteoporosis), clinical presentation, operative variables, and postoperative morbidity were reviewed through electronic medical records. The following validated patient-reported outcomes (PROs) were recorded at baseline and 3-mo after surgery: (1) back-related disability: Oswestry Disability Index (ODI)21; (2) numeric rating scale (NRS) for back pain and leg pain22; and (3) quality of life—EuroQol-5D (EQ-5D).23 Cost Data Total 90-d costs were derived as sum of inpatient hospital stay (hospital cost), surgeons’ professional fee (derived based on current procedural terminology [CPT] codes), and postdischarge health care resource utilization. The costs were derived based on Medicare national payment amounts. To standardize and eliminate geographic variations, a unit multiplier was used. The costs were recorded based on resource utilization, derived from patient-reported use and institutional records. Such calculations have been reported previously.16,17,24-26 The hospital cost was based on the type of surgery performed, the severity of the individual case, and whether in-hospital complications occurred, which collectively determined the diagnosis-related group (DRG). Surgeons’ professional fees were derived based on CPT codes. Ancillary postdischarge resource utilization was derived from CPT codes assigned for patient self-reported resource utilization. Low back-related outpatient visits to surgeons or other physicians, chiropractors, physical and occupational therapists, and acupuncturists were captured. The postdischarge need for X-rays, computed tomography scans, magnetic resonance imaging, and electromyography were tracked to derive diagnostic cost. Postoperative devices (braces, canes, and walkers), emergency department visits, epidural steroid injections, back-specific medications (nonsteroidal anti-inflammatory drugs, oral steroids, narcotics, muscle relaxants, antidepressants), and inpatient and outpatient rehabilitation days were assessed. The costs incurred due to readmissions to our institution during the 90-d period were also recorded. Statistical Analysis Descriptive data including mean (standard deviation) for continuous variables, and frequency (proportions) for categorical variables were computed. Bivariate analysis was conducted to compare the preoperative, operative, and postoperative variables as well as costs among the surgeons. Baseline and 3-mo PROs were compared using paired t-test. Chi-square test and Fisher's exact test for nominal variables and 1-way ANOVA was used for continuous variables. The post hoc test (Bonferroni test) was used to determine the differences in cost between each individual surgeon. The median and standard deviation for total 90-d cost was derived. High cost was defined as total 90-d cost higher than third quartile. Multivariable logistic regression model was built to determine the factors associated with high-cost fusion surgery. The preoperative patient-specific and surgery-specific variables, surgeons as well as complication and readmission, which were selected a priori, were included in the model. The model performance was examined using area under the curve (AUC) for model's receiver operating characteristics curve. To demonstrate the effect of comorbidities on total 90-d cost, we compared the difference in actual 90-d cost and comorbidity adjusted 90-d cost for each surgeon. A separate linear regression model was built for total 90-d cost to derive the comorbidity-adjusted 90-d costs. The variables including age (old age ≥65 yr), obesity (body mass index [BMI] ≥ 35), ASA (American Society of Anesthesiologists scale) grades, number of comorbidities, each comorbidity including diabetes, hypertension, myocardial infarction, atrial fibrillation, arthritis, osteoporosis, chronic pulmonary disease, congestive heart failure, coronary artery disease, and history of preoperative anticoagulation were included in the model. The difference between the actual 90-d cost and comorbidity adjusted 90-d cost was compared for each surgeon. The analyses were performed using SPSS version 22 (IBM, Armonk, New York) analysis software. RESULTS A total of 752 patients undergoing decompression and fusion surgery operated on by 7 surgeons were analyzed. The mean age of 332 male and 418 female patients was 60.8 ± 12.0 yr. Figure 1 demonstrates a bar diagram for frequency of patients operated on by each surgeon. FIGURE 1. View largeDownload slide Bar-diagram demonstrating number of patients operated on by each individual surgeon. FIGURE 1. View largeDownload slide Bar-diagram demonstrating number of patients operated on by each individual surgeon. Variability in Patient Characteristics Table 1 compares the preoperative patient characteristics and surgery variables among surgeons. There was a significant variability in patient-specific variables including age at the time of surgery (P < .0001), smoking status (P = .004), insurance type (P = .003), revision surgery (P = .003), neurogenic claudication (P < .0001), motor deficits (P = .005), higher ASA grades (>3, P = .02), history of hypertension (P = .02), diagnoses (P < .0001), estimated blood loss (P < .0001), and length of hospital stay (P < .0001). The baseline ODI score (ranging from 38.2 ± 12.6 to 53.8 ± 15.7, P < .0001) and NRS-leg pain score (LP; ranging from 5.3 ± 2.4 to 7.6 ± 2.7, P = .02) was significantly different among the participating surgeons. There were no significant differences in EQ-5D, and NRS-back pain (BP; Table 1). TABLE 1. Variation in Patient-Specific and Surgery-Specific Factors Among the Participating Surgeons   1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Age [mean ± SD]  60.4 ± 13.2  61.1 ± 11.4  63.5 ± 10.3  57.5 ± 11.3  71.1 ± 9.1  52.0 ± 11.9  66.1 ± 11.4  <.0001  Gender: male  36 (43%)  42 (39%)  75 (36%)  51 (54%)  4 (67%)  4 (57%)  15 (52%)  .07  BMI [mean ± SD]  30.1 ± 6.4  30.8 ± 6.8  31.8 ± 7.7  31.2 ± 6.1  28.9 ± 6.2  28.4 ± 5.4  31.0 ± 6.9  .47  Smoker  18 (22%)  23 (21%)  27 (13%)  20 (21%)  0  0  1 (3%)  .004  Insurance                .03   Medicaid/uninsured  17 (20%)  24 (22%)  40(20%)  25 (27%)  1 (17%)  0  4 (14%)     Medicare  29 (35%)  36 (34%)  91 (44%)  26 (28%)  4 (67%)  0  13(45%)     Private  37 (45%)  47 (44%)  75 (36%)  43 (46%)  1(16%)  7 (100%)  12 (41%)    Prior surgery  23 (28%)  53 (50%)  58 (28%)  36 (38%)  2 (33%)  3 (43%)  13 (45%)  .003  Neurogenic claudication  10 (12%)  17 (16%)  72 (34%)  18 (19%)  1 (17%)  0  1 (4%)  <.0001  Motor deficits  12 (14%)  30 (28%)  68 (33%)  40 (42%)  3 (50%)  2 (29%)  7 (24%)  .005  Duration of symptoms ≥ 12 m  50 (60%)  77 (72%)  132 (64%)  60 (64%)  5 (83%)  3 (42%)  15 (52%)  .31  Duration of preoperative opioid use [mean ± SD]  425.7 ± 1048  470 ± 1045  364 ± 1082  585 ± 1139  451 ± 1449  304 ± 716  224 ± 455  .55  Comorbidities                   ASA grades > 3  56 (67%)  68 (64%)  159 (77%)  66 (70%)  4 (67%)  4 (57%)  23 (79%)  .02   Diabetes  22 (27%)  21 (20%)  57 (28%)  22 (23%)  1 (17%)  1 (14%)  10 (34%)  .59   Hypertension  47 (57%)  65 (61%)  142 (69%)  59 (63%)  6 (100%)  1 (14%)  20 (69%)  .02   MI  3 (4%)  7 (7%)  12 (6%)  6 (6%)  1 (17%)  0  2 (7%)  .85   CAD  14 (17%)  28 (26%)  46 (22%)  22 (23%)  2 (33%)  0  9 (31%)  .43   COPD  3 (4%)  2 (2%)  8 (4%)  3 (3%)  2 (33%)  0  1 (3%)  .59   Osteoporosis  3 (4%)  2 (2%)  9 (4%)  2 (2%)  0  0  1 (3%)  .88   Preoperative anticoagulation  2 (2%)  2 (2%)  5 (2%)  2 (2%)  0  0  4 (14%)  .03  Primary diagnosis                <.0001   Disc herniation  20 (24%)  7 (7%)  14 (7%)  14 (16%)  0  1 (14%)  0     Stenosis  29 (35%)  63 (59%)  89 (43%)  29 (31%)  1 (17%)  2 (29%)  0     Spondylolisthesis  34 (41%)  37 (35%)  103 (50%)  51 (54%)  5 (83%)  4 (57%)  16 (100%)    Number of levels [mean ± SD]  1.8 ± 0.83  1.9 ± 0.87  1.9 ± 0.85  1.7 ± 0.86  2.0 ± 1.1  1.3 ± 0.76  1.8 ± 0.97  .31  EBL [mean ± SD]  469 ± 302  553 ± 434  709 ± 513  512 ± 363  320 ±124  471 ± 496  509 ± 342  < .0001  Length of surgery [mean ± SD]  240 ± 63  245 ± 79  229 ± 72  235 ± 65  274 ± 76  239 ± 53  253 ± 61  .001  Length of hospital stay [mean ± SD]  4.3 ± 2.2  4.7 ± 3.1  3.6 ± 1.7  4.1 ± 1.8  3.5 ± 1.9  3.7 ± 1.9  2.7 ± 1.1  .29  Interbody graft  61 (44%)  70 (42%)  119 (45%)  69 (65%)  4 (31%)  6 (67%)  15 (47%)  .001    1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Age [mean ± SD]  60.4 ± 13.2  61.1 ± 11.4  63.5 ± 10.3  57.5 ± 11.3  71.1 ± 9.1  52.0 ± 11.9  66.1 ± 11.4  <.0001  Gender: male  36 (43%)  42 (39%)  75 (36%)  51 (54%)  4 (67%)  4 (57%)  15 (52%)  .07  BMI [mean ± SD]  30.1 ± 6.4  30.8 ± 6.8  31.8 ± 7.7  31.2 ± 6.1  28.9 ± 6.2  28.4 ± 5.4  31.0 ± 6.9  .47  Smoker  18 (22%)  23 (21%)  27 (13%)  20 (21%)  0  0  1 (3%)  .004  Insurance                .03   Medicaid/uninsured  17 (20%)  24 (22%)  40(20%)  25 (27%)  1 (17%)  0  4 (14%)     Medicare  29 (35%)  36 (34%)  91 (44%)  26 (28%)  4 (67%)  0  13(45%)     Private  37 (45%)  47 (44%)  75 (36%)  43 (46%)  1(16%)  7 (100%)  12 (41%)    Prior surgery  23 (28%)  53 (50%)  58 (28%)  36 (38%)  2 (33%)  3 (43%)  13 (45%)  .003  Neurogenic claudication  10 (12%)  17 (16%)  72 (34%)  18 (19%)  1 (17%)  0  1 (4%)  <.0001  Motor deficits  12 (14%)  30 (28%)  68 (33%)  40 (42%)  3 (50%)  2 (29%)  7 (24%)  .005  Duration of symptoms ≥ 12 m  50 (60%)  77 (72%)  132 (64%)  60 (64%)  5 (83%)  3 (42%)  15 (52%)  .31  Duration of preoperative opioid use [mean ± SD]  425.7 ± 1048  470 ± 1045  364 ± 1082  585 ± 1139  451 ± 1449  304 ± 716  224 ± 455  .55  Comorbidities                   ASA grades > 3  56 (67%)  68 (64%)  159 (77%)  66 (70%)  4 (67%)  4 (57%)  23 (79%)  .02   Diabetes  22 (27%)  21 (20%)  57 (28%)  22 (23%)  1 (17%)  1 (14%)  10 (34%)  .59   Hypertension  47 (57%)  65 (61%)  142 (69%)  59 (63%)  6 (100%)  1 (14%)  20 (69%)  .02   MI  3 (4%)  7 (7%)  12 (6%)  6 (6%)  1 (17%)  0  2 (7%)  .85   CAD  14 (17%)  28 (26%)  46 (22%)  22 (23%)  2 (33%)  0  9 (31%)  .43   COPD  3 (4%)  2 (2%)  8 (4%)  3 (3%)  2 (33%)  0  1 (3%)  .59   Osteoporosis  3 (4%)  2 (2%)  9 (4%)  2 (2%)  0  0  1 (3%)  .88   Preoperative anticoagulation  2 (2%)  2 (2%)  5 (2%)  2 (2%)  0  0  4 (14%)  .03  Primary diagnosis                <.0001   Disc herniation  20 (24%)  7 (7%)  14 (7%)  14 (16%)  0  1 (14%)  0     Stenosis  29 (35%)  63 (59%)  89 (43%)  29 (31%)  1 (17%)  2 (29%)  0     Spondylolisthesis  34 (41%)  37 (35%)  103 (50%)  51 (54%)  5 (83%)  4 (57%)  16 (100%)    Number of levels [mean ± SD]  1.8 ± 0.83  1.9 ± 0.87  1.9 ± 0.85  1.7 ± 0.86  2.0 ± 1.1  1.3 ± 0.76  1.8 ± 0.97  .31  EBL [mean ± SD]  469 ± 302  553 ± 434  709 ± 513  512 ± 363  320 ±124  471 ± 496  509 ± 342  < .0001  Length of surgery [mean ± SD]  240 ± 63  245 ± 79  229 ± 72  235 ± 65  274 ± 76  239 ± 53  253 ± 61  .001  Length of hospital stay [mean ± SD]  4.3 ± 2.2  4.7 ± 3.1  3.6 ± 1.7  4.1 ± 1.8  3.5 ± 1.9  3.7 ± 1.9  2.7 ± 1.1  .29  Interbody graft  61 (44%)  70 (42%)  119 (45%)  69 (65%)  4 (31%)  6 (67%)  15 (47%)  .001  MI, myocardial infarction; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; EBL, estimated blood level; SD, standard deviation. View Large Variability in Outcomes There was significant improvement in ODI: 45.9 ± 13.4 vs 25.8 ± 18.3, EQ-5D: 0.55 ± 0.20 vs 0.78 ± 0.17, NRS: BP: 6.6 ± 2.5 vs 3.0 ± 2.7, and NRS: LP: 6.6 ± 3 vs 2.6 ± 3.2 (P < .0001) for all patients from baseline to 3 mo after surgery. There were no significant differences in improvement across all the PROs (change score) among the surgeons (Table 2). No significant differences in the complication and readmission rates within 90-d after surgery among the surgeons were observed. TABLE 2. Variation in Complication, Readmission Within 90-day After Surgery, and PROs Among the Participating Surgeons   1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Complication  13 (16%)  22 (21%)  30 (15%)  16 (17%)  1 (17%)  0  4 (14%)  .75  Readmission (spine related)  4 (5%)  3 (3%)  8 (4%)  3 (3%)  0  0  2 (7%)  .93  Baseline PROs                   ODI [mean ± SD]  45.9 ± 13.5  45.6 ± 13.9  47.4 ± 14.3  53.8 ± 15.7  45.7 ± 15.9  38.2 ± 12.6  43.6 ± 13.5  <.001   EQ-5D [mean ± SD]  0.55 ± 0.20  0.57 ± 0.19  0.55 ± 0.21  0.49 ± 0.22  0.56 ± 0.22  0.62 ± 0.16  0.58 ± 0.18  .06   NRS-BP [mean ± SD]  6.5 ± 2.6  6.9 ± 2.3  6.9 ± 2.4  7.2 ± 1.8  6.4 ± 2.4  7.3 ± 0.87  7.5 ± 1.8  .34   NRS-LP [mean ± SD]  6.6 ± 3.0  6.2 ± 2.9  6.9 ± 2.6  6.5 ± 2.8  6.6 ± 2.4  5.3 ± 2.4  7.6 ± 2.7  .02  3-mo PROs                   ODI [mean ± SD]  26.01 ± 18.5  29.8 ± 16.8  28.4 ±15.9  36.2 ± 19.5  25.6 ±17.3  29.3 ± 16.1  26.2 ± 16.7  <.001   EQ-5D [mean ± SD]  0.77 ± 0.17  0.74 ± 0.17  0.75 ± 0.16  0.68 ± 0.21  0.77 ± 0.13  0.75 ± 0.17  0.78 ± 0.16  <.001   NRS-BP [mean ± SD]  3.0 ± 2.6  3.7 ± 2.5  3.5 ± 2.6  4.1 ± 2.8  2.7 ± 2.1  4.0 ± 2.7  3.3 ± 2.8  .06   NRS-LP [mean ± SD]  2.6 ± 3.3  2.9 ± 3.3  2.6 ± 3.0  3.1 ± 3.3  1.7 ± 2.7  3.3 ±3.0  3.0 ± 3.1  .03  Change score for PROs                   ODI [mean ± SD]  20.2 ± 17.2  15.4 ± 16.4  18.4 ± 16.7  17.1 ± 21.6  20.1 ± 17.7  8.8 ± 13.2  16.3 ± 17.1  .18   EQ-5D [mean ± SD]  0.23 ± 0.22  0.17 ± 0.20  0.20 ± 0.22  0.18 ± 0.24  0.22 ± 0.20  0.13 ± 0.24  0.20 ± 0.23  .45   NRS-BP [mean ± SD]  3.6 ± 3.1  3.2 ± 2.7  3.4 ± 3.1  3.1 ± 3.1  3.6 ± 2.9  3.3 ± 3.0  3.9 ± 3.3  .71   NRS-LP [mean ± SD]  3.9 ± 4.1  3.2 ± 3.9  4.3 ± 3.8  3.3 ± 3.9  4.9 ± 3.2  2.0 ± 4.4  4.7 ± 3.4  .15    1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Complication  13 (16%)  22 (21%)  30 (15%)  16 (17%)  1 (17%)  0  4 (14%)  .75  Readmission (spine related)  4 (5%)  3 (3%)  8 (4%)  3 (3%)  0  0  2 (7%)  .93  Baseline PROs                   ODI [mean ± SD]  45.9 ± 13.5  45.6 ± 13.9  47.4 ± 14.3  53.8 ± 15.7  45.7 ± 15.9  38.2 ± 12.6  43.6 ± 13.5  <.001   EQ-5D [mean ± SD]  0.55 ± 0.20  0.57 ± 0.19  0.55 ± 0.21  0.49 ± 0.22  0.56 ± 0.22  0.62 ± 0.16  0.58 ± 0.18  .06   NRS-BP [mean ± SD]  6.5 ± 2.6  6.9 ± 2.3  6.9 ± 2.4  7.2 ± 1.8  6.4 ± 2.4  7.3 ± 0.87  7.5 ± 1.8  .34   NRS-LP [mean ± SD]  6.6 ± 3.0  6.2 ± 2.9  6.9 ± 2.6  6.5 ± 2.8  6.6 ± 2.4  5.3 ± 2.4  7.6 ± 2.7  .02  3-mo PROs                   ODI [mean ± SD]  26.01 ± 18.5  29.8 ± 16.8  28.4 ±15.9  36.2 ± 19.5  25.6 ±17.3  29.3 ± 16.1  26.2 ± 16.7  <.001   EQ-5D [mean ± SD]  0.77 ± 0.17  0.74 ± 0.17  0.75 ± 0.16  0.68 ± 0.21  0.77 ± 0.13  0.75 ± 0.17  0.78 ± 0.16  <.001   NRS-BP [mean ± SD]  3.0 ± 2.6  3.7 ± 2.5  3.5 ± 2.6  4.1 ± 2.8  2.7 ± 2.1  4.0 ± 2.7  3.3 ± 2.8  .06   NRS-LP [mean ± SD]  2.6 ± 3.3  2.9 ± 3.3  2.6 ± 3.0  3.1 ± 3.3  1.7 ± 2.7  3.3 ±3.0  3.0 ± 3.1  .03  Change score for PROs                   ODI [mean ± SD]  20.2 ± 17.2  15.4 ± 16.4  18.4 ± 16.7  17.1 ± 21.6  20.1 ± 17.7  8.8 ± 13.2  16.3 ± 17.1  .18   EQ-5D [mean ± SD]  0.23 ± 0.22  0.17 ± 0.20  0.20 ± 0.22  0.18 ± 0.24  0.22 ± 0.20  0.13 ± 0.24  0.20 ± 0.23  .45   NRS-BP [mean ± SD]  3.6 ± 3.1  3.2 ± 2.7  3.4 ± 3.1  3.1 ± 3.1  3.6 ± 2.9  3.3 ± 3.0  3.9 ± 3.3  .71   NRS-LP [mean ± SD]  3.9 ± 4.1  3.2 ± 3.9  4.3 ± 3.8  3.3 ± 3.9  4.9 ± 3.2  2.0 ± 4.4  4.7 ± 3.4  .15  SD, standard deviation. View Large Variability in Cost The mean total 90-d direct cost for laminectomy and fusion surgery was $28 947 ± $9484 (median: $27 565, interquartile range: $22 952, $32 837; Figure 2). The DRG-based hospital cost for these patients was $24 399 ± $8190. There were significant differences in the hospital cost, surgeons’ professional fee, and costs associated with postdischarge resource utilization among the surgeons (Table 3). FIGURE 2. View largeDownload slide Box-plot representing total 90-day cost among the participating surgeons. The empty circle and asterisk represent the outliers above third quartile or below first quartile. The outliers above third quartile were defined as high-cost patients. FIGURE 2. View largeDownload slide Box-plot representing total 90-day cost among the participating surgeons. The empty circle and asterisk represent the outliers above third quartile or below first quartile. The outliers above third quartile were defined as high-cost patients. TABLE 3. Variations in Total 90-day Cost Including Hospital Cost, Surgeon Profession Costs, Postdischarge Health Care Resource Utilization (Health Care Visits, Medication Costs, Diagnostic Imaging Costs), and Readmission Costs Among the Participating Surgeons Mean (SD)  1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Direct cost 90 d  $28 345  $32 272  $26 810  $33 674  $26 310  $29 805  $23 103  <.0001    ($9472)  ($11 272)  ($7536)  ($9776)  ($7665)  ($7665)  ($5912)      (P < .0001)  (P = .328)  (P < .0001)    (P = .067)  (P = .884)  (P < .0001)    Hospital cost  $24 134  $26 464  $22 481  $28 256  $23 746  $25 495  $18 529  <.0001    ($8676)  ($9713)  ($6236)  ($8118)  ($4545)  ($7651)  ($5535)      (P = .128)  (P = .390)  (P = 0.001)    (P = .1)  (P = .999)  (P = .004)    Surgeon professional fee  $3138  $3342  $3053  $3355  $3068  $2993  $3046  .001    ($706)  ($812)  ($663)  ($872)  ($547)  ($341)  ($685)      (P = .291)  (P = .1)  (P = 0.007)    (P = .939)  (P = .836)  (P = .354)    Post-discharge health care visits  $703  $966  $834  $908  $337  $1207  $977  .002    ($701)  ($1133)  ($1186)  ($1053)  ($610)  ($795)  ($760)      (P = .729)  (P = .999)  (P = 0.996)    (P = .01)  (P = .987)  (P = .1)    Medication cost  $376  $470  $351  $478  $254  $442  $263  <.0001    ($282)  ($310)  ($303)  ($329)  ($198)  ($313)  ($244)      (P = .109)  (P = 1.0)  (P = 0.004)    (P = .122)  (P = 1.0)  (P = .006)    Diagnostic costs  $149  $160  $71  $244  $219  $132  $53  <.0001    ($282)  ($309)  ($179)  ($413)  ($341)  ($281)  ($95)      (P = .1)  (P = .089)  (P = 0.104)    (P = .972)  (P = .1)  (P = .545)    Readmission costs  $7822  $14 800  $11 980  $16 260  –  –  $7708 (−)  .479    ($5255)  ($10 740)  ($7904)  ($4103)          Mean (SD)  1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Direct cost 90 d  $28 345  $32 272  $26 810  $33 674  $26 310  $29 805  $23 103  <.0001    ($9472)  ($11 272)  ($7536)  ($9776)  ($7665)  ($7665)  ($5912)      (P < .0001)  (P = .328)  (P < .0001)    (P = .067)  (P = .884)  (P < .0001)    Hospital cost  $24 134  $26 464  $22 481  $28 256  $23 746  $25 495  $18 529  <.0001    ($8676)  ($9713)  ($6236)  ($8118)  ($4545)  ($7651)  ($5535)      (P = .128)  (P = .390)  (P = 0.001)    (P = .1)  (P = .999)  (P = .004)    Surgeon professional fee  $3138  $3342  $3053  $3355  $3068  $2993  $3046  .001    ($706)  ($812)  ($663)  ($872)  ($547)  ($341)  ($685)      (P = .291)  (P = .1)  (P = 0.007)    (P = .939)  (P = .836)  (P = .354)    Post-discharge health care visits  $703  $966  $834  $908  $337  $1207  $977  .002    ($701)  ($1133)  ($1186)  ($1053)  ($610)  ($795)  ($760)      (P = .729)  (P = .999)  (P = 0.996)    (P = .01)  (P = .987)  (P = .1)    Medication cost  $376  $470  $351  $478  $254  $442  $263  <.0001    ($282)  ($310)  ($303)  ($329)  ($198)  ($313)  ($244)      (P = .109)  (P = 1.0)  (P = 0.004)    (P = .122)  (P = 1.0)  (P = .006)    Diagnostic costs  $149  $160  $71  $244  $219  $132  $53  <.0001    ($282)  ($309)  ($179)  ($413)  ($341)  ($281)  ($95)      (P = .1)  (P = .089)  (P = 0.104)    (P = .972)  (P = .1)  (P = .545)    Readmission costs  $7822  $14 800  $11 980  $16 260  –  –  $7708 (−)  .479    ($5255)  ($10 740)  ($7904)  ($4103)          View Large Multivariable Model for High-Cost Fusion Surgery Twenty-five percent (n = 188) of patients were above the third quartile of the total 90-d cost (Figure 2), and were defined as high-cost fusion surgery. In a multivariable logistic regression analysis, the length of hospital stay (odds ratio [OR]: 1.3, 95% confidence interval [CI]: 1.2-1.5, P < .0001), length of surgery (OR: 1.013, 95% CI: 1.01-1.02, P < .0001, P < .0001), number of levels operated on (OR: 1.4, 95% CI: 1.2-1.5, P = .023), occurrence of 90-d complications (OR: 1.8, 95% CI: 1.01-3.5, P = .0045) and readmission (OR: 3.7, 95% CI: 1.2-10.9, P = .019) were associated with high 90-d costs for fusion surgery. After controlling for all the aforementioned variables, surgeon # 4 (OR: 21.4, 95% CI: 5.2-88.7, P < .0001) and surgeon # 6 (OR: 20.5, 95% CI: 2.7-155.7, P = .003) had higher odds of having high-cost fusion patients (Table 4). The AUC for models’ receiver operating characteristics curve was 0.885. TABLE 4. Multivariable Logistic Regression Analysis for High-Cost Fusion Surgery       95% CI for OR    P-value  OR  Lower  Upper  Age  .975  0.99  0.97  1.03  BMI  .451  1.01  0.98  1.05  Smoker  .246  1.40  0.79  2.49  Insurance           Medicaid vs private  .427  1.25  0.72  2.2   Medicare vs private  .711  0.89  0.47  1.67  Neurogenic claudication  .811  0.93  0.53  1.64  Duration of symptoms  .134  0.71  0.45  1.11  Revision surgery  .226  1.33  0.84  2.09  Number of comorbidities  .309  1.1  0.92  1.32  ASA grades >3  .116  0.65  0.38  1.11  Length of surgery (minutes)  <.0001  1.01  1.01  1.02  EBL (mL)  .051  1.0  1  1.01  Number of levels  .023  1.41  1.04  1.78  Interbody fusion  .105  1.43  0.93  2.21  Length of hospital stay  <.0001  1.3  1.2  1.5  Surgeon           Surgeon #1 vs #7  .092  3.36  0.82  13.77   Surgeon #2 vs #7  .073  3.59  0.89  14.49   Surgeon #3 vs #7  .258  2.21  0.56  8.77   Surgeon #4 vs #7  <.0001  21.4  5.2  88.7   Surgeon #5 vs #7  .998  0  0  –   Surgeon #6 vs #7  .003  20.5  2.7  155.7  90-day complication  .045  1.8  1.01  3.5  90-day readmission  .019  3.7  1.2  10.9        95% CI for OR    P-value  OR  Lower  Upper  Age  .975  0.99  0.97  1.03  BMI  .451  1.01  0.98  1.05  Smoker  .246  1.40  0.79  2.49  Insurance           Medicaid vs private  .427  1.25  0.72  2.2   Medicare vs private  .711  0.89  0.47  1.67  Neurogenic claudication  .811  0.93  0.53  1.64  Duration of symptoms  .134  0.71  0.45  1.11  Revision surgery  .226  1.33  0.84  2.09  Number of comorbidities  .309  1.1  0.92  1.32  ASA grades >3  .116  0.65  0.38  1.11  Length of surgery (minutes)  <.0001  1.01  1.01  1.02  EBL (mL)  .051  1.0  1  1.01  Number of levels  .023  1.41  1.04  1.78  Interbody fusion  .105  1.43  0.93  2.21  Length of hospital stay  <.0001  1.3  1.2  1.5  Surgeon           Surgeon #1 vs #7  .092  3.36  0.82  13.77   Surgeon #2 vs #7  .073  3.59  0.89  14.49   Surgeon #3 vs #7  .258  2.21  0.56  8.77   Surgeon #4 vs #7  <.0001  21.4  5.2  88.7   Surgeon #5 vs #7  .998  0  0  –   Surgeon #6 vs #7  .003  20.5  2.7  155.7  90-day complication  .045  1.8  1.01  3.5  90-day readmission  .019  3.7  1.2  10.9  BMI, body mass index; EBL, estimated blood loss. View Large Multivariable Model to Derive Comorbidity-Adjusted Cost Table 5 summarizes the comorbidities included in the model to derive the comorbidity-adjusted cost. To adjust for the patients’ preoperative health state, we compared the comorbidity-adjusted 90-d cost to the actual 90-d cost for each participating surgeon. Figure 3 demonstrates the 90-d adjusted cost vs actual 90-d cost for each participating surgeon. The vertical axis represents actual median 90-d cost ($27 565). The black dots represent average comorbidity-adjusted 90-d cost for each surgeon and the colored dot represents the average actual 90-d costs. The line joining the colored and black dots represents the difference between the actual and the adjusted average cost. The adjusted cost was higher than the actual cost for surgeons # 1 (P = .08), #3 (P = .002), #5 (P < .0001), and #7 (P < .0001), whereas the adjusted cost was lower for surgeons # 2 (P = .128), #4 (P < .0001), and #6 (P = .44). This suggests that based on the health state of their respective patients, the surgeons #1, #3, #5, and #7 spent lower than expected and surgeons #2, #4, and #6 were costlier than expected. FIGURE 3. View largeDownload slide The 90-day adjusted cost vs actual 90-day cost for each participating surgeon. The vertical axis represents actual median 90-day cost ($27 565). The black dots represent mean comorbidity-adjusted 90-day cost for each surgeon and the colored dot represents the mean actual 90-day costs. The line joining the colored and black dots represents the difference between the actual and the adjusted cost. The adjusted cost was higher than the actual cost for surgeons #1 (P = .08), #3 (P = .002), #5 (P < .0001), and #7 (P < .0001), suggesting these surgeons were less costly than predicted. In contrast, the adjusted cost was lower for surgeons #2 (P = .128), #4 (P < .0001), and #6 (P = .44), which suggests that these surgeons were more costly than predicted after adjusting for comorbidities. FIGURE 3. View largeDownload slide The 90-day adjusted cost vs actual 90-day cost for each participating surgeon. The vertical axis represents actual median 90-day cost ($27 565). The black dots represent mean comorbidity-adjusted 90-day cost for each surgeon and the colored dot represents the mean actual 90-day costs. The line joining the colored and black dots represents the difference between the actual and the adjusted cost. The adjusted cost was higher than the actual cost for surgeons #1 (P = .08), #3 (P = .002), #5 (P < .0001), and #7 (P < .0001), suggesting these surgeons were less costly than predicted. In contrast, the adjusted cost was lower for surgeons #2 (P = .128), #4 (P < .0001), and #6 (P = .44), which suggests that these surgeons were more costly than predicted after adjusting for comorbidities. TABLE 5. Multivariable Linear Regression Model to Derive Comorbidity-Adjusted Cost       95% CI for beta    Beta coefficient  P-value  Lower bound  Upper bound  Intercept  26832  <.001  22871  30793  Number of comorbidities  321  .736  –1550  2191  Age > 65 yr  1566  .028  165  2966  Smoker  1952  .037  116  3789  ASA grade  343  .652  –1148  1833  Diabetes  712  .58  –1810  3234  Hypertension  –856  .506  –3386  1673  Myocardial infarction  2698  .23  –1712  7108  Congestive heart failure  –1119  .701  –6841  4603  Chronic obstructive pulmonary disease  –117  .955  –4206  3972  Atrial fibrillation  –1974  .325  –5906  1958  Obesity (BMI > 65)  1044  .407  –1425  3513  Arthritis  –503  .697  –3033  2027  Osteoporosis  –1755  .41  –5933  2424  Preoperative anticoagulation  –75  .976  –5016  4866        95% CI for beta    Beta coefficient  P-value  Lower bound  Upper bound  Intercept  26832  <.001  22871  30793  Number of comorbidities  321  .736  –1550  2191  Age > 65 yr  1566  .028  165  2966  Smoker  1952  .037  116  3789  ASA grade  343  .652  –1148  1833  Diabetes  712  .58  –1810  3234  Hypertension  –856  .506  –3386  1673  Myocardial infarction  2698  .23  –1712  7108  Congestive heart failure  –1119  .701  –6841  4603  Chronic obstructive pulmonary disease  –117  .955  –4206  3972  Atrial fibrillation  –1974  .325  –5906  1958  Obesity (BMI > 65)  1044  .407  –1425  3513  Arthritis  –503  .697  –3033  2027  Osteoporosis  –1755  .41  –5933  2424  Preoperative anticoagulation  –75  .976  –5016  4866  BMI, body mass index. View Large DISCUSSION Current trends in strategies to target cost-containment are specially aimed at reducing variability in cost and outcomes. A number of studies have defined the variability in outcomes following spine surgery and a handful of studies have defined the variability in cost within each individual spine-related DRG.8,9,11,12,27 None of the prior studies have demonstrated the variations in cost and outcomes at the individual surgeon level. In these analyses, utilizing prospectively collected registry data from a single-center, we demonstrate that there were significant differences in total 90-d costs and patient-specific factors among the surgeons. Despite these differences, no significant differences in improvement in the disability, pain, and quality of life outcomes 90 d after surgery were observed. Some surgeons had higher odds of performing high-cost fusion surgery. There was a significant variability in comorbidity-adjusted cost vs actual cost among the surgeons. After adjusting for preoperative comorbidities, the adjusted costs were higher than actual cost for some surgeons and lower than the actual cost for others. Our study provides valuable insights into variations in patient characteristics, outcomes, and costs among the participating surgeons at a single institution. This study can form the basis to stimulate action to improve uniformity and cost-containment for lumbar fusion surgery. Number of factors can result in variations in the cost associated with fusion surgery. It is possible that the interplay of patient-specific factors and surgery-specific factors may also induce variation in the utilization of health care services and therefore induce variability in total 90-d costs. Prior studies have demonstrated that the patients’ comorbidity burden influence 90-d costs.13-17 In this study, we have sought to take one step further, however, and simulate a workflow for surgeon-level process improvement. Figure 3 reveals which surgeons incurred average costs exceeding risk-adjusted predictions, and which surgeons were less costly than predicted. This finding highlights that there may be several exogenous factors, in addition to the patient-specific comorbidities that are tied to the way surgeons practice that influences the total 90-d costs. We can now turn our attention to a particular surgeon who was costlier than expected and investigate his or her specific cost distribution. For example, surgeons #4 and #6 had higher frequency of performing interbody graft fusion compared to other surgeons. However, after adjusting for interbody graft and other surgery-specific variables the surgeons #4 and #6 were high-cost surgeons. This suggests that there might be confounders beyond the variables included in the model. For example, surgeon #6 had higher postdischarge health care visit associated costs and surgeon #4 had higher hospital cost. Further, granular data with details on type of implant used, additional intraoperative costs, and breakdown of the cost associated with inpatient hospital stay is needed to accurately determine the practice pattern of high-cost surgeons. A habitual practice of examining surgeon-level cost variability in this manner, when coupled with evaluations of modifiable risk factors within surgeons’ high-cost patient populations, can provide the makings of a true “learning health care system” as imagined by the Institute of Medicine.28 Consistent with previous studies, in our analysis, the high cost for fusion surgery was also associated with postoperative complication and readmission, increased surgery duration, and extended length of hospital stay. Complications and readmission within 90 d global period occur at a consistent frequency, specifically when analyzing the larger data sets.20,29-32 The factors including patient age, obesity, associated comorbidities, primary diagnosis, and surgical invasiveness and complexity are associated with higher likelihood of developing complications and also influence the cost and outcomes following surgery.18,19,30,33-38 We demonstrate that complications do contribute to higher cost for fusion surgery; however, there was no difference in the complication rate among the participating surgeons. This suggests that occurrence of complication alone might not explain the variation in cost among surgeons. Clearly, measures focused on prevention of complications will be able to decrease the cost and therefore increase the cost-benefit ratio. Analogous to the previous studies, extended duration of surgery and length of hospital stay were significant drivers of the total 90-d cost. The surgery duration and length of hospital stay were significantly different among the participating surgeons. Identifying the factors associated with this variability might be able to precisely determine and remedy the modifiable factors and contain the escalating costs associated with fusion. In our study, surgeons #4 and #6 had higher odds of performing high-cost fusion surgery. There is a tendency for a surgeon to account for the high cost by stating that they care for sicker patients. In a retrospective study, Walid et al13 demonstrated that the comorbidities additively increase the cost associated with spine surgery. Several other authors have emphasized the importance of considering the patients’ comorbidities as a driver of cost and outcomes following spine surgery.13-20 None of the previous studies have demonstrated the surgeon level variability in cost accounted for the comorbidities. In our study after adjusting for comorbidities, surgeon #4 was found to be costlier than expected and surgeon #6 was found to be less costly than expected, based on the health state of their respective patients. Furthermore, surgeon #2 was not a high-cost surgeon overall; however, this surgeon was found to be costlier than expected, when adjusting for comorbidities. This suggests that there is no doubt that preoperative comorbidities should be accounted for when deriving the aggregate total cost. However, as mentioned above, there are other factors beyond comorbidities that influence the total 90-d cost. The value-based pay-per-performance for an episode of care (30- or 90-d) initiatives are conceptually simple and have potential to improvise the health-care quality and contain escalating costs.39-45 The ideal episode of care model should be well designed to accommodate the highs and lows in each component of total 90-d cost. Given the heterogeneity of patient-specific and surgeon-specific factors, implementation of these models in real-world practice is complex and challenging. As demonstrated in our analyses, there is significant variation in the cost and characteristics of the patients managed by each surgeon at a single institution. Without fuller accountability of these variations in each component of the cost, the episode of care model can potentially be seen as an infringement of the individual surgeons’ autonomy. Therefore, it is imperative to involve surgeons in the decision-making and to consider variability in cost at the individual surgeon level after adjusting for complexity of the patient population managed by each surgeon. Limitations There are several limitations associated with this study. This is a single-institution study, and the costs were adjusted based for Medicare allowable 2013 national payments amount. The institution costs may vary for private payers and if a different geographical multiplier is used. We used unit multiple to eliminate any geographical variation. Furthermore, we do not have granular data on different types of instrumentation, use of infuse, or type of biologics used by each surgeon. This will certainly be important as hospitals try to better understand cost saving measures to improve profit margin. This will require a partnership with surgeons, so the costs are cut but care is not compromised. Our institution, similar to other institutions will not allow detailed information regarding unique cost arrangements with companies to be published. The data on postdischarge utilization is gleaned from the electronic medical records and augmented by patient interview (to capture care outside of the facility) and is subject to recall bias. Furthermore, the indirect cost from societal perspective including patient workday losses, family work day loss, caregiver cost was not included in the cost calculation, as the goal of this study was to define variability in cost from payers’ and providers’ perspective. The confounding variables included in the analysis may not be exhaustive, therefore adding more variables might account for other variation in the 90-d costs. Finally, the number of patients operated on by each surgeon was different. Sensitivity analysis was performed. A separate model was fitted excluding patients operated on by surgeons #5 and #6. There were no differences in the effect size and model performance (AUC—0.881) between this model and all-inclusive model (AUC—0.885; Tables, Supplemental Digital Content 1 and 2). The generalized application of these data needs to consider these limitations. Nonetheless, utilizing comprehensive list of variables captured in a single-center prospective longitudinal spine registry, we present the variations in 90-d cost and outcomes for laminectomy and fusion for lumbar degenerative pathology. CONCLUSION Our study provides valuable insight into variations in 90-d costs among the surgeons performing elective lumbar laminectomy/fusions at a single institution. Specific surgeons were found to have greater odds of performing high-cost surgeries. Adjusting for preoperative comorbidities, however, led to costs that were higher than the actual costs for certain surgeons and lower than the actual costs for others. Patient's preoperative comorbidities must therefore be accounted for when crafting value-based bundled payment models. More broadly, this study demonstrates that the designing intervention targeting “modifiable” factors tied to the way surgeons practice may increase the overall value of lumbar laminectomy and fusion. Disclosure The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. REFERENCES 1. Deyo RA, Gray DT, Kreuter W, Mirza S, Martin BI. United States trends in lumbar fusion surgery for degenerative conditions. Spine . 2005; 30( 12): 1441- 1445; discussion 1446-1447. Google Scholar CrossRef Search ADS PubMed  2. Dagenais S, Caro J, Haldeman S. A systematic review of low back pain cost of illness studies in the United States and internationally. Spine J . 2008; 8( 1): 8- 20. Google Scholar CrossRef Search ADS PubMed  3. Frymoyer JW, Cats-Baril WL. An overview of the incidences and costs of low back pain. Orthop Clin North Am . 1991; 22( 2): 263- 271. Google Scholar PubMed  4. Martin BI, Deyo RA, Mirza SK et al.   Expenditures and health status among adults with back and neck problems. JAMA . 2008; 299( 6): 656- 664. Google Scholar CrossRef Search ADS PubMed  5. Martin BI, Turner JA, Mirza SK, Lee MJ, Comstock BA, Deyo RA. Trends in health care expenditures, utilization, and health status among US adults with spine problems, 1997-2006. Spine . 2009; 34( 19): 2077- 2084. Google Scholar CrossRef Search ADS PubMed  6. Get ready: bundled payments are in your future. Hosp Case Manag . 2015; 23( 9): 109- 112. PubMed  7. Bozic KJ, Ward L, Vail TP, Maze M. Bundled payments in total joint arthroplasty: targeting opportunities for quality improvement and cost reduction. Clin Orthop Relat Res . 2014; 472( 1): 188- 193. Google Scholar CrossRef Search ADS PubMed  8. Ugiliweneza B, Kong M, Nosova K et al.   Spinal surgery: variations in health care costs and implications for episode-based bundled payments. Spine . 2014; 39( 15): 1235- 1242. Google Scholar CrossRef Search ADS PubMed  9. Birkmeyer JD, Gust C, Baser O, Dimick JB, Sutherland JM, Skinner JS. Medicare payments for common inpatient procedures: implications for episode-based payment bundling. Health Serv Res . 2010; 45( 6 pt 1): 1783- 1795. Google Scholar CrossRef Search ADS PubMed  10. Grenda TR, Pradarelli JC, Thumma JR, Dimick JB. Variation in hospital episode costs with bariatric surgery. JAMA Surg . 2015; 150( 12): 1109- 1115. Google Scholar CrossRef Search ADS PubMed  11. Schoenfeld AJ, Harris MB, Liu H, Birkmeyer JD. Variations in Medicare payments for episodes of spine surgery. Spine J . 2014; 14( 12): 2793- 2798. Google Scholar CrossRef Search ADS PubMed  12. Miller DC, Gust C, Dimick JB, Birkmeyer N, Skinner J, Birkmeyer JD. Large variations in Medicare payments for surgery highlight savings potential from bundled payment programs. Health Aff . 2011; 30( 11): 2107- 2115. Google Scholar CrossRef Search ADS   13. Walid MS, Robinson JS Jr. Economic impact of comorbidities in spine surgery. J Neurosurg Spine.  2011; 14( 3): 318- 321. Google Scholar CrossRef Search ADS PubMed  14. Walid MS, Robinson EC, Robinson JS Jr. Higher comorbidity rates in unemployed patients may significantly impact the cost of spine surgery. J Clin Neurosci . 2011; 18( 5): 640- 644. Google Scholar CrossRef Search ADS PubMed  15. Kalanithi PA, Arrigo R, Boakye M. Morbid obesity increases cost and complication rates in spinal arthrodesis. Spine . 2012; 37( 11): 982- 988. Google Scholar CrossRef Search ADS PubMed  16. Devin CJ, Chotai S, Parker SL, Tetreault L, Fehlings MG, McGirt MJ. A cost-utility analysis of lumbar decompression with and without fusion for degenerative spine disease in the elderly. Neurosurgery . 2015; 77( suppl 4): S116- S124. Google Scholar CrossRef Search ADS PubMed  17. Chotai S, Sielatycki JA, Parker SL et al.   Effect of obesity on cost per quality-adjusted life years gained following anterior cervical discectomy and fusion in elective degenerative pathology. Spine J . 2016; 16( 11): 1342- 1350. Google Scholar CrossRef Search ADS PubMed  18. Wang MY, Green BA, Shah S, Vanni S, Levi AD. Complications associated with lumbar stenosis surgery in patients older than 75 years of age. Neurosurgical Focus . 2003; 14( 2): e7. Google Scholar CrossRef Search ADS PubMed  19. Campbell PG, Yadla S, Nasser R, Malone J, Maltenfort MG, Ratliff JK. Patient comorbidity score predicting the incidence of perioperative complications: assessing the impact of comorbidities on complications in spine surgery. J Neurosurg Spine.  2012; 16( 1): 37- 43. Google Scholar CrossRef Search ADS PubMed  20. Peter GC, Sanjay Y, Rani N, Jennifer M, Mitchell GM, John KR. Patient comorbidity score predicting the incidence of perioperative complications: assessing the impact of comorbidities on complications in spine surgery. J Neurosurg Spine . 2012; 16( 1): 37- 43. Google Scholar CrossRef Search ADS PubMed  21. Fairbank JC, Pynsent PB. The oswestry disability index. Spine . 2000; 25( 22): 2940- 2952; discussion 2952. Google Scholar CrossRef Search ADS PubMed  22. Langley GB, Sheppeard H. The visual analogue scale: its use in pain measurement. Rheumatol Int . 1985; 5( 4): 145- 148. Google Scholar CrossRef Search ADS PubMed  23. EuroQol. –a new facility for the measurement of health-related quality of life. Health Pol . 1990; 16( 3): 199- 208. CrossRef Search ADS   24. Parker SL, Fulchiero EC, Davis BJ et al.   Cost-effectiveness of multilevel hemilaminectomy for lumbar stenosis-associated radiculopathy. Spine J . 2011; 11( 8): 705- 711. Google Scholar CrossRef Search ADS PubMed  25. Adogwa O, Parker SL, Shau D et al.   Cost per quality-adjusted life year gained of revision fusion for lumbar pseudoarthrosis: defining the value of surgery. J Spinal Disord Tech . 2015; 28( 3): 101- 105. Google Scholar CrossRef Search ADS PubMed  26. Parker SL, Godil SS, Mendenhall SK, Zuckerman SL, Shau DN, McGirt MJ. Two-year comprehensive medical management of degenerative lumbar spine disease (lumbar spondylolisthesis, stenosis, or disc herniation): a value analysis of cost, pain, disability, and quality of life: clinical article. J Neurosurg Spine.  2014; 21( 2): 143- 149. Google Scholar CrossRef Search ADS PubMed  27. McGirt MJ, Sivaganesan A, Asher AL, Devin CJ. Prediction model for outcome after low-back surgery: individualized likelihood of complication, hospital readmission, return to work, and 12-month improvement in functional disability. Neurosurgical Focus . 2015; 39( 6): E13. Google Scholar CrossRef Search ADS PubMed  28. Smith M. Available at: http://www.nationalacademies.org/hmd/Activities/Quality/LearningHealthCare.aspx Accessed September 2016. 29. Carreon LY, Puno RM, Dimar JR 2nd, Glassman SD, Johnson JR. Perioperative complications of posterior lumbar decompression and arthrodesis in older adults. J Bone Joint Surg Am . 2003; 85-a( 11): 2089- 2092. Google Scholar CrossRef Search ADS PubMed  30. Cassinelli EH, Eubanks J, Vogt M, Furey C, Yoo J, Bohlman HH. Risk factors for the development of perioperative complications in elderly patients undergoing lumbar decompression and arthrodesis for spinal stenosis: an analysis of 166 patients. Spine . 2007; 32( 2): 230- 235. Google Scholar CrossRef Search ADS PubMed  31. Hoffman RM, Wheeler KJ, Deyo RA. Surgery for herniated lumbar discs: a literature synthesis. J Gen Int Med . 1993; 8( 9): 487- 496. Google Scholar CrossRef Search ADS   32. Turner JA, Loeser JD, Deyo RA, Sanders SB. Spinal cord stimulation for patients with failed back surgery syndrome or complex regional pain syndrome: a systematic review of effectiveness and complications. Pain . 2004; 108( 1-2): 137- 147. Google Scholar CrossRef Search ADS PubMed  33. Bekelis K, Desai A, Bakhoum SF, Missios S. A predictive model of complications after spine surgery: the National Surgical Quality Improvement Program (NSQIP) 2005-2010. Spine J . 2014; 14( 7): 1247- 1255. Google Scholar CrossRef Search ADS PubMed  34. Faciszewski T, Winter RB, Lonstein JE, Denis F, Johnson L. The surgical and medical perioperative complications of anterior spinal fusion surgery in the thoracic and lumbar spine in adults. A review of 1223 procedures. Spine . 1995; 20( 14): 1592- 1599. Google Scholar CrossRef Search ADS PubMed  35. Glassman SD, Alegre G, Carreon L, Dimar JR, Johnson JR. Perioperative complications of lumbar instrumentation and fusion in patients with diabetes mellitus. Spine J . 2003; 3( 6): 496- 501. Google Scholar CrossRef Search ADS PubMed  36. Guzman JZ, Iatridis JC, Skovrlj B et al.   Outcomes and complications of diabetes mellitus on patients undergoing degenerative lumbar spine surgery. Spine . 2014; 39( 19): 1596- 1604. Google Scholar CrossRef Search ADS PubMed  37. Schwender JD, Casnellie MT, Perra JH et al.   Perioperative complications in revision anterior lumbar spine surgery: incidence and risk factors. Spine . 2009; 34( 1): 87- 90. Google Scholar CrossRef Search ADS PubMed  38. Kalanithi PS, Patil CG, Boakye M. National complication rates and disposition after posterior lumbar fusion for acquired spondylolisthesis. Spine . 2009; 34( 18): 1963- 1969. Google Scholar CrossRef Search ADS PubMed  39. Cutler DM, Ghosh K. The potential for cost savings through bundled episode payments. N Engl J Med . 2012; 366( 12): 1075- 1077. Google Scholar CrossRef Search ADS PubMed  40. Delisle DR. Big things come in bundled packages: implications of bundled payment systems in health care reimbursement reform. Am J Med Qual . 2013; 28( 4): 339- 344. Google Scholar CrossRef Search ADS PubMed  41. Iorio R. Strategies and tactics for successful implementation of bundled payments: bundled payment for care improvement at a large, urban, academic medical center. J Arthroplasty . 2015; 30( 3): 349- 350. Google Scholar CrossRef Search ADS PubMed  42. Kazberouk A, McGuire K, Landon BE. A survey of innovative reimbursement models in spine care. Spine . 2016; 41( 4): 344- 352. Google Scholar CrossRef Search ADS PubMed  43. Mechanic RE. Opportunities and challenges for episode-based payment. N Engl J Med . 2011; 365( 9): 777- 779. Google Scholar CrossRef Search ADS PubMed  44. Rossi VJ, Ahn J, Bohl DD, Tabaraee E, Singh K. Economic factors in the future delivery of spinal healthcare. World J Orthop . 2015; 6( 5): 409- 412. Google Scholar CrossRef Search ADS PubMed  45. Sood N, Huckfeldt PJ, Escarce JJ, Grabowski DC, Newhouse JP. Medicare's bundled payment pilot for acute and postacute care: analysis and recommendations on where to begin. Health Aff . 2011; 30( 9): 1708- 1717. Google Scholar CrossRef Search ADS   Supplemental digital content is available for this article at www.neurosurgery-online.com. COMMENTS In this manuscript, the authors have presented a cost-utilization and outcome analysis of inter-surgeon variability for elective lumbar decompression and fusion. The authors have outlined the variations in patient profile, cost, and outcomes for each surgeon performing laminectomy and fusion for degenerative spine disease at a single center. The authors found that when adjusted for preoperative comorbidities, the costs for certain surgeons were found to be higher than the comorbidity-adjusted 90-day cost. As a per surgeon utilization study, the manuscript represents an important addition to the literature. Single center studies such as this are uniquely situated to assess per surgeon costs as they control for institutional variables. With that said, a multi-center study would provide insight into cost factors that remain out of the surgeons control. In one of the supplemental tables, the authors provide an analysis that suggests that the higher cost surgeons fused more levels and had longer Operating Room time, indicating that they were doing more involved cases than the low-cost surgeons. How much of the cost difference does this explain? In addition to adjusting for preoperative morbidities, the authors should consider adjusting for intraoperative factors (type of fusion surgery, length of construct, etc). Moreover, 3 of 7 surgeons did 33 or fewer cases between 2011 and 2015, with 1 surgeon doing only 9 cases in that period. The other surgeons in the study did well over 100 cases each. Although in some analyses 2 of the 3 lower-volume surgeons were excluded, it is unclear the relationship this had to the overall cost driver analysis. In future analyses, the authors may consider delving into the reasons that explain the cost variations. For instance, what aspects of patient care or operative factors (implantable devices, etc) drive the costs for each surgeon? It would also be interesting to see how the publication of the article has impacted the cost decisions of the surgeons identified as higher cost relative to their colleagues. Overall, this article highlights an important topic: surgeon variability and its relationship to costs. Mohamad Bydon Rochester, Minnesota The authors examined a longitudinal registry from a single institution for 90-day costs and outcomes for 752 patients undergoing laminectomy and fusion surgery. They found that while there were no differences in the complication rate, 90-day readmission rate, or measurements of clinical outcome, 2 of the 7 surgeons incurred significantly higher costs. The subject relates to critical analysis of the cost for the care that spinal surgeons deliver—a matter universally recognized as an area we as a society need to better understand to permit the best care that we can afford. The study is limited in providing any practical information about the value of any particular surgeon's approach to operative management of degenerative spine disease; as the authors note that they lacked any "granular data" to assess the various surgeons with regard to use of particular spinal implants, biologics, or costs associated with the inpatient stay. Still, the study is useful in providing needed data to validate the importance of considering the preoperative morbidities in creating bundled payment models. John Kenneth Houten Bronx, New York Copyright © 2017 by the Congress of Neurological Surgeons http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Neurosurgery Oxford University Press

Surgeon-Level Variability in Outcomes, Cost, and Comorbidity Adjusted-Cost for Elective Lumbar Decompression and Fusion

Loading next page...
 
/lp/ou_press/surgeon-level-variability-in-outcomes-cost-and-comorbidity-adjusted-eTpRgJJIIK
Publisher
Oxford University Press
Copyright
Copyright © 2017 by the Congress of Neurological Surgeons
ISSN
0148-396X
eISSN
1524-4040
D.O.I.
10.1093/neuros/nyx243
Publisher site
See Article on Publisher Site

Abstract

Abstract BACKGROUND The costs and outcomes following degenerative spine surgery may vary from surgeon to surgeon. Patient factors such as comorbidities may increase the health care cost. These variations are not well studied. OBJECTIVE To understand the variation in outcomes, costs, and comorbidity-adjusted cost for surgeons performing lumbar laminectomy and fusions surgery. METHODS A total of 752 patients undergoing laminectomy and fusion, performed by 7 surgeons, were analyzed. Patient-reported outcomes and 90-d cost were analyzed. Multivariate regression model was built for high-cost surgery. A separate linear regression model was built to derive comorbidity-adjusted 90-d costs. RESULTS No significant differences in improvement were found across all the patient-reported outcomes, complications, and readmission among the surgeons. In multivariable model, surgeons #4 (P < .0001) and #6 (P = .002) had higher odds of performing high-cost fusion surgery. The comorbidity-adjusted costs were higher than the actual 90-d costs for surgeons #1 (P = .08), #3 (P = .002), #5 (P < .0001), and #7 (P < .0001), whereas they were lower than the actual costs for surgeons #2 (P = .128), #4 (P < .0001), and #6 (P = .44). CONCLUSION Our study provides valuable insight into variations in 90-d costs among the surgeons performing elective lumbar laminectomy and fusion at a single institution. Specific surgeons were found to have greater odds of performing high-cost surgeries. Adjusting for preoperative comorbidities, however, led to costs that were higher than the actual costs for certain surgeons and lower than the actual costs for others. Patients’ preoperative comorbidities must be accounted for when crafting value-based payment models. Furthermore, designing intervention targeting “modifiable” factors tied to the way the surgeons practice may increase the overall value of spine care. Spine, 90-d cost, Outcomes, Surgeon variability, Bundled payments ABBREVIATIONS ABBREVIATIONS ASA American Society of Anesthesiologists AUC area under the curve BMI body mass index CI confidence interval CMS Centers for Medicare and Medicaid Services CPT current procedural terminology DRG diagnosis-related group EQ-5D EuroQol-5D NRS numeric rating scale ODI Oswestry Disability Index OR odds ratio PRO patient-reported outcome US United States The current trajectory of the health care expenditure in the United States (US) is unsustainable. According to the Centers for Medicare and Medicaid Services (CMS), projected health spending in the US will be as high as 20% of the gross domestic product by 2021. Low back pain associated with lumbar degenerative pathologies is highly prevalent and an economic burden.1-5 It is important to investigate costs and to examine the real-world clinical and cost benefit of commonly performed spinal surgeries. In an effort to curb escalating costs, CMS has proposed a pay-for-performance episode-based bundled payment model for hospital and physician reimbursement. The current fee-for-service model, whereby the Medicate reimbursement to the hospital is based on discounted rates payment rates established under the Inpatient Prospective Payment System, is considered a potential source of increasing health care spending. With the proposed bundled payment initiative, the individual physician, hospital, and other providers will be accountable for the quality of care and associated costs from surgery through 90 d after discharge.6-8 This model mandates all stakeholders to collaborate and investigate optimal strategies for arriving at a target price for a bundled payment. To determine a sustainable bundled cost, it is important to understand the biggest contributors to and variability in the cost at each level of patient care. The inpatient hospital cost, cost associated with readmission following surgery, and surgeon's professional fee cost are the 3 most important contributors of the cost for spine surgery.8-11 A number of previous studies have focused on the variations in hospital cost and total cost during the 90-d postdischarge period.8,9,11,12 Less well studied is the individual surgeon variation in cost, specifically from the same institution and given fairly standardized techniques. Variation in cost, if extant, provides an opportunity to understand the differences and learn from the better practice patterns. Given the wide variation in patient profile, disease process, surgical techniques, and practice patterns, it is prudent to assess the cost variability and to determine the factors driving high cost surgery. Furthermore, it is generally believed that the sicker patients account for higher cost associated with spine care.13-20 It is vital to understand the variation in cost associated with comorbidities. In this regard, the aims of this study were to define variability in total 90-d cost among the surgeons, to define factors associated with high-cost surgery and to determine the difference in total 90-d cost and cost adjusted for comorbidities for patients undergoing laminectomy and fusion surgery for degenerative spine diseases. METHODS Patients undergoing elective decompression and fusion surgery for degenerative spine pathology between 2011 and 2015 at a single comprehensive spine center were enrolled into a single-center prospective longitudinal spine registry. A retrospective review of prospectively collected data was conducted. An approval for the study and wavier of informed consent was obtained from the institutional review board for all the patients entered into the registry. The inclusion criteria were (1) patients age >18 yr; (2) presenting with leg and/or back pain; (3) the correlative imaging findings for the diagnosis of disc herniation, stenosis, and spondylolisthesis; and (4) failed 3-mo of multimodal nonoperative care or patients with progressive neurological deficit. The exclusion criteria were (1) pathological spine disease including tumor, infection, and trauma; (2) any extra-spinal cause of back or leg pain; (3); patients who were unable or unwilling to complete the follow-up questionnaire. Patients operated on by 7 surgeons, who are participating in the registry, were analyzed. Patient demographics, comorbidities (diabetes, hypertension, coronary artery disease, myocardial infarction, preoperative anticoagulation, congestive heart failure, chronic obstructive pulmonary disease, and osteoporosis), clinical presentation, operative variables, and postoperative morbidity were reviewed through electronic medical records. The following validated patient-reported outcomes (PROs) were recorded at baseline and 3-mo after surgery: (1) back-related disability: Oswestry Disability Index (ODI)21; (2) numeric rating scale (NRS) for back pain and leg pain22; and (3) quality of life—EuroQol-5D (EQ-5D).23 Cost Data Total 90-d costs were derived as sum of inpatient hospital stay (hospital cost), surgeons’ professional fee (derived based on current procedural terminology [CPT] codes), and postdischarge health care resource utilization. The costs were derived based on Medicare national payment amounts. To standardize and eliminate geographic variations, a unit multiplier was used. The costs were recorded based on resource utilization, derived from patient-reported use and institutional records. Such calculations have been reported previously.16,17,24-26 The hospital cost was based on the type of surgery performed, the severity of the individual case, and whether in-hospital complications occurred, which collectively determined the diagnosis-related group (DRG). Surgeons’ professional fees were derived based on CPT codes. Ancillary postdischarge resource utilization was derived from CPT codes assigned for patient self-reported resource utilization. Low back-related outpatient visits to surgeons or other physicians, chiropractors, physical and occupational therapists, and acupuncturists were captured. The postdischarge need for X-rays, computed tomography scans, magnetic resonance imaging, and electromyography were tracked to derive diagnostic cost. Postoperative devices (braces, canes, and walkers), emergency department visits, epidural steroid injections, back-specific medications (nonsteroidal anti-inflammatory drugs, oral steroids, narcotics, muscle relaxants, antidepressants), and inpatient and outpatient rehabilitation days were assessed. The costs incurred due to readmissions to our institution during the 90-d period were also recorded. Statistical Analysis Descriptive data including mean (standard deviation) for continuous variables, and frequency (proportions) for categorical variables were computed. Bivariate analysis was conducted to compare the preoperative, operative, and postoperative variables as well as costs among the surgeons. Baseline and 3-mo PROs were compared using paired t-test. Chi-square test and Fisher's exact test for nominal variables and 1-way ANOVA was used for continuous variables. The post hoc test (Bonferroni test) was used to determine the differences in cost between each individual surgeon. The median and standard deviation for total 90-d cost was derived. High cost was defined as total 90-d cost higher than third quartile. Multivariable logistic regression model was built to determine the factors associated with high-cost fusion surgery. The preoperative patient-specific and surgery-specific variables, surgeons as well as complication and readmission, which were selected a priori, were included in the model. The model performance was examined using area under the curve (AUC) for model's receiver operating characteristics curve. To demonstrate the effect of comorbidities on total 90-d cost, we compared the difference in actual 90-d cost and comorbidity adjusted 90-d cost for each surgeon. A separate linear regression model was built for total 90-d cost to derive the comorbidity-adjusted 90-d costs. The variables including age (old age ≥65 yr), obesity (body mass index [BMI] ≥ 35), ASA (American Society of Anesthesiologists scale) grades, number of comorbidities, each comorbidity including diabetes, hypertension, myocardial infarction, atrial fibrillation, arthritis, osteoporosis, chronic pulmonary disease, congestive heart failure, coronary artery disease, and history of preoperative anticoagulation were included in the model. The difference between the actual 90-d cost and comorbidity adjusted 90-d cost was compared for each surgeon. The analyses were performed using SPSS version 22 (IBM, Armonk, New York) analysis software. RESULTS A total of 752 patients undergoing decompression and fusion surgery operated on by 7 surgeons were analyzed. The mean age of 332 male and 418 female patients was 60.8 ± 12.0 yr. Figure 1 demonstrates a bar diagram for frequency of patients operated on by each surgeon. FIGURE 1. View largeDownload slide Bar-diagram demonstrating number of patients operated on by each individual surgeon. FIGURE 1. View largeDownload slide Bar-diagram demonstrating number of patients operated on by each individual surgeon. Variability in Patient Characteristics Table 1 compares the preoperative patient characteristics and surgery variables among surgeons. There was a significant variability in patient-specific variables including age at the time of surgery (P < .0001), smoking status (P = .004), insurance type (P = .003), revision surgery (P = .003), neurogenic claudication (P < .0001), motor deficits (P = .005), higher ASA grades (>3, P = .02), history of hypertension (P = .02), diagnoses (P < .0001), estimated blood loss (P < .0001), and length of hospital stay (P < .0001). The baseline ODI score (ranging from 38.2 ± 12.6 to 53.8 ± 15.7, P < .0001) and NRS-leg pain score (LP; ranging from 5.3 ± 2.4 to 7.6 ± 2.7, P = .02) was significantly different among the participating surgeons. There were no significant differences in EQ-5D, and NRS-back pain (BP; Table 1). TABLE 1. Variation in Patient-Specific and Surgery-Specific Factors Among the Participating Surgeons   1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Age [mean ± SD]  60.4 ± 13.2  61.1 ± 11.4  63.5 ± 10.3  57.5 ± 11.3  71.1 ± 9.1  52.0 ± 11.9  66.1 ± 11.4  <.0001  Gender: male  36 (43%)  42 (39%)  75 (36%)  51 (54%)  4 (67%)  4 (57%)  15 (52%)  .07  BMI [mean ± SD]  30.1 ± 6.4  30.8 ± 6.8  31.8 ± 7.7  31.2 ± 6.1  28.9 ± 6.2  28.4 ± 5.4  31.0 ± 6.9  .47  Smoker  18 (22%)  23 (21%)  27 (13%)  20 (21%)  0  0  1 (3%)  .004  Insurance                .03   Medicaid/uninsured  17 (20%)  24 (22%)  40(20%)  25 (27%)  1 (17%)  0  4 (14%)     Medicare  29 (35%)  36 (34%)  91 (44%)  26 (28%)  4 (67%)  0  13(45%)     Private  37 (45%)  47 (44%)  75 (36%)  43 (46%)  1(16%)  7 (100%)  12 (41%)    Prior surgery  23 (28%)  53 (50%)  58 (28%)  36 (38%)  2 (33%)  3 (43%)  13 (45%)  .003  Neurogenic claudication  10 (12%)  17 (16%)  72 (34%)  18 (19%)  1 (17%)  0  1 (4%)  <.0001  Motor deficits  12 (14%)  30 (28%)  68 (33%)  40 (42%)  3 (50%)  2 (29%)  7 (24%)  .005  Duration of symptoms ≥ 12 m  50 (60%)  77 (72%)  132 (64%)  60 (64%)  5 (83%)  3 (42%)  15 (52%)  .31  Duration of preoperative opioid use [mean ± SD]  425.7 ± 1048  470 ± 1045  364 ± 1082  585 ± 1139  451 ± 1449  304 ± 716  224 ± 455  .55  Comorbidities                   ASA grades > 3  56 (67%)  68 (64%)  159 (77%)  66 (70%)  4 (67%)  4 (57%)  23 (79%)  .02   Diabetes  22 (27%)  21 (20%)  57 (28%)  22 (23%)  1 (17%)  1 (14%)  10 (34%)  .59   Hypertension  47 (57%)  65 (61%)  142 (69%)  59 (63%)  6 (100%)  1 (14%)  20 (69%)  .02   MI  3 (4%)  7 (7%)  12 (6%)  6 (6%)  1 (17%)  0  2 (7%)  .85   CAD  14 (17%)  28 (26%)  46 (22%)  22 (23%)  2 (33%)  0  9 (31%)  .43   COPD  3 (4%)  2 (2%)  8 (4%)  3 (3%)  2 (33%)  0  1 (3%)  .59   Osteoporosis  3 (4%)  2 (2%)  9 (4%)  2 (2%)  0  0  1 (3%)  .88   Preoperative anticoagulation  2 (2%)  2 (2%)  5 (2%)  2 (2%)  0  0  4 (14%)  .03  Primary diagnosis                <.0001   Disc herniation  20 (24%)  7 (7%)  14 (7%)  14 (16%)  0  1 (14%)  0     Stenosis  29 (35%)  63 (59%)  89 (43%)  29 (31%)  1 (17%)  2 (29%)  0     Spondylolisthesis  34 (41%)  37 (35%)  103 (50%)  51 (54%)  5 (83%)  4 (57%)  16 (100%)    Number of levels [mean ± SD]  1.8 ± 0.83  1.9 ± 0.87  1.9 ± 0.85  1.7 ± 0.86  2.0 ± 1.1  1.3 ± 0.76  1.8 ± 0.97  .31  EBL [mean ± SD]  469 ± 302  553 ± 434  709 ± 513  512 ± 363  320 ±124  471 ± 496  509 ± 342  < .0001  Length of surgery [mean ± SD]  240 ± 63  245 ± 79  229 ± 72  235 ± 65  274 ± 76  239 ± 53  253 ± 61  .001  Length of hospital stay [mean ± SD]  4.3 ± 2.2  4.7 ± 3.1  3.6 ± 1.7  4.1 ± 1.8  3.5 ± 1.9  3.7 ± 1.9  2.7 ± 1.1  .29  Interbody graft  61 (44%)  70 (42%)  119 (45%)  69 (65%)  4 (31%)  6 (67%)  15 (47%)  .001    1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Age [mean ± SD]  60.4 ± 13.2  61.1 ± 11.4  63.5 ± 10.3  57.5 ± 11.3  71.1 ± 9.1  52.0 ± 11.9  66.1 ± 11.4  <.0001  Gender: male  36 (43%)  42 (39%)  75 (36%)  51 (54%)  4 (67%)  4 (57%)  15 (52%)  .07  BMI [mean ± SD]  30.1 ± 6.4  30.8 ± 6.8  31.8 ± 7.7  31.2 ± 6.1  28.9 ± 6.2  28.4 ± 5.4  31.0 ± 6.9  .47  Smoker  18 (22%)  23 (21%)  27 (13%)  20 (21%)  0  0  1 (3%)  .004  Insurance                .03   Medicaid/uninsured  17 (20%)  24 (22%)  40(20%)  25 (27%)  1 (17%)  0  4 (14%)     Medicare  29 (35%)  36 (34%)  91 (44%)  26 (28%)  4 (67%)  0  13(45%)     Private  37 (45%)  47 (44%)  75 (36%)  43 (46%)  1(16%)  7 (100%)  12 (41%)    Prior surgery  23 (28%)  53 (50%)  58 (28%)  36 (38%)  2 (33%)  3 (43%)  13 (45%)  .003  Neurogenic claudication  10 (12%)  17 (16%)  72 (34%)  18 (19%)  1 (17%)  0  1 (4%)  <.0001  Motor deficits  12 (14%)  30 (28%)  68 (33%)  40 (42%)  3 (50%)  2 (29%)  7 (24%)  .005  Duration of symptoms ≥ 12 m  50 (60%)  77 (72%)  132 (64%)  60 (64%)  5 (83%)  3 (42%)  15 (52%)  .31  Duration of preoperative opioid use [mean ± SD]  425.7 ± 1048  470 ± 1045  364 ± 1082  585 ± 1139  451 ± 1449  304 ± 716  224 ± 455  .55  Comorbidities                   ASA grades > 3  56 (67%)  68 (64%)  159 (77%)  66 (70%)  4 (67%)  4 (57%)  23 (79%)  .02   Diabetes  22 (27%)  21 (20%)  57 (28%)  22 (23%)  1 (17%)  1 (14%)  10 (34%)  .59   Hypertension  47 (57%)  65 (61%)  142 (69%)  59 (63%)  6 (100%)  1 (14%)  20 (69%)  .02   MI  3 (4%)  7 (7%)  12 (6%)  6 (6%)  1 (17%)  0  2 (7%)  .85   CAD  14 (17%)  28 (26%)  46 (22%)  22 (23%)  2 (33%)  0  9 (31%)  .43   COPD  3 (4%)  2 (2%)  8 (4%)  3 (3%)  2 (33%)  0  1 (3%)  .59   Osteoporosis  3 (4%)  2 (2%)  9 (4%)  2 (2%)  0  0  1 (3%)  .88   Preoperative anticoagulation  2 (2%)  2 (2%)  5 (2%)  2 (2%)  0  0  4 (14%)  .03  Primary diagnosis                <.0001   Disc herniation  20 (24%)  7 (7%)  14 (7%)  14 (16%)  0  1 (14%)  0     Stenosis  29 (35%)  63 (59%)  89 (43%)  29 (31%)  1 (17%)  2 (29%)  0     Spondylolisthesis  34 (41%)  37 (35%)  103 (50%)  51 (54%)  5 (83%)  4 (57%)  16 (100%)    Number of levels [mean ± SD]  1.8 ± 0.83  1.9 ± 0.87  1.9 ± 0.85  1.7 ± 0.86  2.0 ± 1.1  1.3 ± 0.76  1.8 ± 0.97  .31  EBL [mean ± SD]  469 ± 302  553 ± 434  709 ± 513  512 ± 363  320 ±124  471 ± 496  509 ± 342  < .0001  Length of surgery [mean ± SD]  240 ± 63  245 ± 79  229 ± 72  235 ± 65  274 ± 76  239 ± 53  253 ± 61  .001  Length of hospital stay [mean ± SD]  4.3 ± 2.2  4.7 ± 3.1  3.6 ± 1.7  4.1 ± 1.8  3.5 ± 1.9  3.7 ± 1.9  2.7 ± 1.1  .29  Interbody graft  61 (44%)  70 (42%)  119 (45%)  69 (65%)  4 (31%)  6 (67%)  15 (47%)  .001  MI, myocardial infarction; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; EBL, estimated blood level; SD, standard deviation. View Large Variability in Outcomes There was significant improvement in ODI: 45.9 ± 13.4 vs 25.8 ± 18.3, EQ-5D: 0.55 ± 0.20 vs 0.78 ± 0.17, NRS: BP: 6.6 ± 2.5 vs 3.0 ± 2.7, and NRS: LP: 6.6 ± 3 vs 2.6 ± 3.2 (P < .0001) for all patients from baseline to 3 mo after surgery. There were no significant differences in improvement across all the PROs (change score) among the surgeons (Table 2). No significant differences in the complication and readmission rates within 90-d after surgery among the surgeons were observed. TABLE 2. Variation in Complication, Readmission Within 90-day After Surgery, and PROs Among the Participating Surgeons   1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Complication  13 (16%)  22 (21%)  30 (15%)  16 (17%)  1 (17%)  0  4 (14%)  .75  Readmission (spine related)  4 (5%)  3 (3%)  8 (4%)  3 (3%)  0  0  2 (7%)  .93  Baseline PROs                   ODI [mean ± SD]  45.9 ± 13.5  45.6 ± 13.9  47.4 ± 14.3  53.8 ± 15.7  45.7 ± 15.9  38.2 ± 12.6  43.6 ± 13.5  <.001   EQ-5D [mean ± SD]  0.55 ± 0.20  0.57 ± 0.19  0.55 ± 0.21  0.49 ± 0.22  0.56 ± 0.22  0.62 ± 0.16  0.58 ± 0.18  .06   NRS-BP [mean ± SD]  6.5 ± 2.6  6.9 ± 2.3  6.9 ± 2.4  7.2 ± 1.8  6.4 ± 2.4  7.3 ± 0.87  7.5 ± 1.8  .34   NRS-LP [mean ± SD]  6.6 ± 3.0  6.2 ± 2.9  6.9 ± 2.6  6.5 ± 2.8  6.6 ± 2.4  5.3 ± 2.4  7.6 ± 2.7  .02  3-mo PROs                   ODI [mean ± SD]  26.01 ± 18.5  29.8 ± 16.8  28.4 ±15.9  36.2 ± 19.5  25.6 ±17.3  29.3 ± 16.1  26.2 ± 16.7  <.001   EQ-5D [mean ± SD]  0.77 ± 0.17  0.74 ± 0.17  0.75 ± 0.16  0.68 ± 0.21  0.77 ± 0.13  0.75 ± 0.17  0.78 ± 0.16  <.001   NRS-BP [mean ± SD]  3.0 ± 2.6  3.7 ± 2.5  3.5 ± 2.6  4.1 ± 2.8  2.7 ± 2.1  4.0 ± 2.7  3.3 ± 2.8  .06   NRS-LP [mean ± SD]  2.6 ± 3.3  2.9 ± 3.3  2.6 ± 3.0  3.1 ± 3.3  1.7 ± 2.7  3.3 ±3.0  3.0 ± 3.1  .03  Change score for PROs                   ODI [mean ± SD]  20.2 ± 17.2  15.4 ± 16.4  18.4 ± 16.7  17.1 ± 21.6  20.1 ± 17.7  8.8 ± 13.2  16.3 ± 17.1  .18   EQ-5D [mean ± SD]  0.23 ± 0.22  0.17 ± 0.20  0.20 ± 0.22  0.18 ± 0.24  0.22 ± 0.20  0.13 ± 0.24  0.20 ± 0.23  .45   NRS-BP [mean ± SD]  3.6 ± 3.1  3.2 ± 2.7  3.4 ± 3.1  3.1 ± 3.1  3.6 ± 2.9  3.3 ± 3.0  3.9 ± 3.3  .71   NRS-LP [mean ± SD]  3.9 ± 4.1  3.2 ± 3.9  4.3 ± 3.8  3.3 ± 3.9  4.9 ± 3.2  2.0 ± 4.4  4.7 ± 3.4  .15    1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Complication  13 (16%)  22 (21%)  30 (15%)  16 (17%)  1 (17%)  0  4 (14%)  .75  Readmission (spine related)  4 (5%)  3 (3%)  8 (4%)  3 (3%)  0  0  2 (7%)  .93  Baseline PROs                   ODI [mean ± SD]  45.9 ± 13.5  45.6 ± 13.9  47.4 ± 14.3  53.8 ± 15.7  45.7 ± 15.9  38.2 ± 12.6  43.6 ± 13.5  <.001   EQ-5D [mean ± SD]  0.55 ± 0.20  0.57 ± 0.19  0.55 ± 0.21  0.49 ± 0.22  0.56 ± 0.22  0.62 ± 0.16  0.58 ± 0.18  .06   NRS-BP [mean ± SD]  6.5 ± 2.6  6.9 ± 2.3  6.9 ± 2.4  7.2 ± 1.8  6.4 ± 2.4  7.3 ± 0.87  7.5 ± 1.8  .34   NRS-LP [mean ± SD]  6.6 ± 3.0  6.2 ± 2.9  6.9 ± 2.6  6.5 ± 2.8  6.6 ± 2.4  5.3 ± 2.4  7.6 ± 2.7  .02  3-mo PROs                   ODI [mean ± SD]  26.01 ± 18.5  29.8 ± 16.8  28.4 ±15.9  36.2 ± 19.5  25.6 ±17.3  29.3 ± 16.1  26.2 ± 16.7  <.001   EQ-5D [mean ± SD]  0.77 ± 0.17  0.74 ± 0.17  0.75 ± 0.16  0.68 ± 0.21  0.77 ± 0.13  0.75 ± 0.17  0.78 ± 0.16  <.001   NRS-BP [mean ± SD]  3.0 ± 2.6  3.7 ± 2.5  3.5 ± 2.6  4.1 ± 2.8  2.7 ± 2.1  4.0 ± 2.7  3.3 ± 2.8  .06   NRS-LP [mean ± SD]  2.6 ± 3.3  2.9 ± 3.3  2.6 ± 3.0  3.1 ± 3.3  1.7 ± 2.7  3.3 ±3.0  3.0 ± 3.1  .03  Change score for PROs                   ODI [mean ± SD]  20.2 ± 17.2  15.4 ± 16.4  18.4 ± 16.7  17.1 ± 21.6  20.1 ± 17.7  8.8 ± 13.2  16.3 ± 17.1  .18   EQ-5D [mean ± SD]  0.23 ± 0.22  0.17 ± 0.20  0.20 ± 0.22  0.18 ± 0.24  0.22 ± 0.20  0.13 ± 0.24  0.20 ± 0.23  .45   NRS-BP [mean ± SD]  3.6 ± 3.1  3.2 ± 2.7  3.4 ± 3.1  3.1 ± 3.1  3.6 ± 2.9  3.3 ± 3.0  3.9 ± 3.3  .71   NRS-LP [mean ± SD]  3.9 ± 4.1  3.2 ± 3.9  4.3 ± 3.8  3.3 ± 3.9  4.9 ± 3.2  2.0 ± 4.4  4.7 ± 3.4  .15  SD, standard deviation. View Large Variability in Cost The mean total 90-d direct cost for laminectomy and fusion surgery was $28 947 ± $9484 (median: $27 565, interquartile range: $22 952, $32 837; Figure 2). The DRG-based hospital cost for these patients was $24 399 ± $8190. There were significant differences in the hospital cost, surgeons’ professional fee, and costs associated with postdischarge resource utilization among the surgeons (Table 3). FIGURE 2. View largeDownload slide Box-plot representing total 90-day cost among the participating surgeons. The empty circle and asterisk represent the outliers above third quartile or below first quartile. The outliers above third quartile were defined as high-cost patients. FIGURE 2. View largeDownload slide Box-plot representing total 90-day cost among the participating surgeons. The empty circle and asterisk represent the outliers above third quartile or below first quartile. The outliers above third quartile were defined as high-cost patients. TABLE 3. Variations in Total 90-day Cost Including Hospital Cost, Surgeon Profession Costs, Postdischarge Health Care Resource Utilization (Health Care Visits, Medication Costs, Diagnostic Imaging Costs), and Readmission Costs Among the Participating Surgeons Mean (SD)  1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Direct cost 90 d  $28 345  $32 272  $26 810  $33 674  $26 310  $29 805  $23 103  <.0001    ($9472)  ($11 272)  ($7536)  ($9776)  ($7665)  ($7665)  ($5912)      (P < .0001)  (P = .328)  (P < .0001)    (P = .067)  (P = .884)  (P < .0001)    Hospital cost  $24 134  $26 464  $22 481  $28 256  $23 746  $25 495  $18 529  <.0001    ($8676)  ($9713)  ($6236)  ($8118)  ($4545)  ($7651)  ($5535)      (P = .128)  (P = .390)  (P = 0.001)    (P = .1)  (P = .999)  (P = .004)    Surgeon professional fee  $3138  $3342  $3053  $3355  $3068  $2993  $3046  .001    ($706)  ($812)  ($663)  ($872)  ($547)  ($341)  ($685)      (P = .291)  (P = .1)  (P = 0.007)    (P = .939)  (P = .836)  (P = .354)    Post-discharge health care visits  $703  $966  $834  $908  $337  $1207  $977  .002    ($701)  ($1133)  ($1186)  ($1053)  ($610)  ($795)  ($760)      (P = .729)  (P = .999)  (P = 0.996)    (P = .01)  (P = .987)  (P = .1)    Medication cost  $376  $470  $351  $478  $254  $442  $263  <.0001    ($282)  ($310)  ($303)  ($329)  ($198)  ($313)  ($244)      (P = .109)  (P = 1.0)  (P = 0.004)    (P = .122)  (P = 1.0)  (P = .006)    Diagnostic costs  $149  $160  $71  $244  $219  $132  $53  <.0001    ($282)  ($309)  ($179)  ($413)  ($341)  ($281)  ($95)      (P = .1)  (P = .089)  (P = 0.104)    (P = .972)  (P = .1)  (P = .545)    Readmission costs  $7822  $14 800  $11 980  $16 260  –  –  $7708 (−)  .479    ($5255)  ($10 740)  ($7904)  ($4103)          Mean (SD)  1 (n = 141)  2 (n = 167)  3 (n = 279)  4 (n = 109)  5 (n = 14)  6 (n = 9)  7 (n = 33)  P-value  Direct cost 90 d  $28 345  $32 272  $26 810  $33 674  $26 310  $29 805  $23 103  <.0001    ($9472)  ($11 272)  ($7536)  ($9776)  ($7665)  ($7665)  ($5912)      (P < .0001)  (P = .328)  (P < .0001)    (P = .067)  (P = .884)  (P < .0001)    Hospital cost  $24 134  $26 464  $22 481  $28 256  $23 746  $25 495  $18 529  <.0001    ($8676)  ($9713)  ($6236)  ($8118)  ($4545)  ($7651)  ($5535)      (P = .128)  (P = .390)  (P = 0.001)    (P = .1)  (P = .999)  (P = .004)    Surgeon professional fee  $3138  $3342  $3053  $3355  $3068  $2993  $3046  .001    ($706)  ($812)  ($663)  ($872)  ($547)  ($341)  ($685)      (P = .291)  (P = .1)  (P = 0.007)    (P = .939)  (P = .836)  (P = .354)    Post-discharge health care visits  $703  $966  $834  $908  $337  $1207  $977  .002    ($701)  ($1133)  ($1186)  ($1053)  ($610)  ($795)  ($760)      (P = .729)  (P = .999)  (P = 0.996)    (P = .01)  (P = .987)  (P = .1)    Medication cost  $376  $470  $351  $478  $254  $442  $263  <.0001    ($282)  ($310)  ($303)  ($329)  ($198)  ($313)  ($244)      (P = .109)  (P = 1.0)  (P = 0.004)    (P = .122)  (P = 1.0)  (P = .006)    Diagnostic costs  $149  $160  $71  $244  $219  $132  $53  <.0001    ($282)  ($309)  ($179)  ($413)  ($341)  ($281)  ($95)      (P = .1)  (P = .089)  (P = 0.104)    (P = .972)  (P = .1)  (P = .545)    Readmission costs  $7822  $14 800  $11 980  $16 260  –  –  $7708 (−)  .479    ($5255)  ($10 740)  ($7904)  ($4103)          View Large Multivariable Model for High-Cost Fusion Surgery Twenty-five percent (n = 188) of patients were above the third quartile of the total 90-d cost (Figure 2), and were defined as high-cost fusion surgery. In a multivariable logistic regression analysis, the length of hospital stay (odds ratio [OR]: 1.3, 95% confidence interval [CI]: 1.2-1.5, P < .0001), length of surgery (OR: 1.013, 95% CI: 1.01-1.02, P < .0001, P < .0001), number of levels operated on (OR: 1.4, 95% CI: 1.2-1.5, P = .023), occurrence of 90-d complications (OR: 1.8, 95% CI: 1.01-3.5, P = .0045) and readmission (OR: 3.7, 95% CI: 1.2-10.9, P = .019) were associated with high 90-d costs for fusion surgery. After controlling for all the aforementioned variables, surgeon # 4 (OR: 21.4, 95% CI: 5.2-88.7, P < .0001) and surgeon # 6 (OR: 20.5, 95% CI: 2.7-155.7, P = .003) had higher odds of having high-cost fusion patients (Table 4). The AUC for models’ receiver operating characteristics curve was 0.885. TABLE 4. Multivariable Logistic Regression Analysis for High-Cost Fusion Surgery       95% CI for OR    P-value  OR  Lower  Upper  Age  .975  0.99  0.97  1.03  BMI  .451  1.01  0.98  1.05  Smoker  .246  1.40  0.79  2.49  Insurance           Medicaid vs private  .427  1.25  0.72  2.2   Medicare vs private  .711  0.89  0.47  1.67  Neurogenic claudication  .811  0.93  0.53  1.64  Duration of symptoms  .134  0.71  0.45  1.11  Revision surgery  .226  1.33  0.84  2.09  Number of comorbidities  .309  1.1  0.92  1.32  ASA grades >3  .116  0.65  0.38  1.11  Length of surgery (minutes)  <.0001  1.01  1.01  1.02  EBL (mL)  .051  1.0  1  1.01  Number of levels  .023  1.41  1.04  1.78  Interbody fusion  .105  1.43  0.93  2.21  Length of hospital stay  <.0001  1.3  1.2  1.5  Surgeon           Surgeon #1 vs #7  .092  3.36  0.82  13.77   Surgeon #2 vs #7  .073  3.59  0.89  14.49   Surgeon #3 vs #7  .258  2.21  0.56  8.77   Surgeon #4 vs #7  <.0001  21.4  5.2  88.7   Surgeon #5 vs #7  .998  0  0  –   Surgeon #6 vs #7  .003  20.5  2.7  155.7  90-day complication  .045  1.8  1.01  3.5  90-day readmission  .019  3.7  1.2  10.9        95% CI for OR    P-value  OR  Lower  Upper  Age  .975  0.99  0.97  1.03  BMI  .451  1.01  0.98  1.05  Smoker  .246  1.40  0.79  2.49  Insurance           Medicaid vs private  .427  1.25  0.72  2.2   Medicare vs private  .711  0.89  0.47  1.67  Neurogenic claudication  .811  0.93  0.53  1.64  Duration of symptoms  .134  0.71  0.45  1.11  Revision surgery  .226  1.33  0.84  2.09  Number of comorbidities  .309  1.1  0.92  1.32  ASA grades >3  .116  0.65  0.38  1.11  Length of surgery (minutes)  <.0001  1.01  1.01  1.02  EBL (mL)  .051  1.0  1  1.01  Number of levels  .023  1.41  1.04  1.78  Interbody fusion  .105  1.43  0.93  2.21  Length of hospital stay  <.0001  1.3  1.2  1.5  Surgeon           Surgeon #1 vs #7  .092  3.36  0.82  13.77   Surgeon #2 vs #7  .073  3.59  0.89  14.49   Surgeon #3 vs #7  .258  2.21  0.56  8.77   Surgeon #4 vs #7  <.0001  21.4  5.2  88.7   Surgeon #5 vs #7  .998  0  0  –   Surgeon #6 vs #7  .003  20.5  2.7  155.7  90-day complication  .045  1.8  1.01  3.5  90-day readmission  .019  3.7  1.2  10.9  BMI, body mass index; EBL, estimated blood loss. View Large Multivariable Model to Derive Comorbidity-Adjusted Cost Table 5 summarizes the comorbidities included in the model to derive the comorbidity-adjusted cost. To adjust for the patients’ preoperative health state, we compared the comorbidity-adjusted 90-d cost to the actual 90-d cost for each participating surgeon. Figure 3 demonstrates the 90-d adjusted cost vs actual 90-d cost for each participating surgeon. The vertical axis represents actual median 90-d cost ($27 565). The black dots represent average comorbidity-adjusted 90-d cost for each surgeon and the colored dot represents the average actual 90-d costs. The line joining the colored and black dots represents the difference between the actual and the adjusted average cost. The adjusted cost was higher than the actual cost for surgeons # 1 (P = .08), #3 (P = .002), #5 (P < .0001), and #7 (P < .0001), whereas the adjusted cost was lower for surgeons # 2 (P = .128), #4 (P < .0001), and #6 (P = .44). This suggests that based on the health state of their respective patients, the surgeons #1, #3, #5, and #7 spent lower than expected and surgeons #2, #4, and #6 were costlier than expected. FIGURE 3. View largeDownload slide The 90-day adjusted cost vs actual 90-day cost for each participating surgeon. The vertical axis represents actual median 90-day cost ($27 565). The black dots represent mean comorbidity-adjusted 90-day cost for each surgeon and the colored dot represents the mean actual 90-day costs. The line joining the colored and black dots represents the difference between the actual and the adjusted cost. The adjusted cost was higher than the actual cost for surgeons #1 (P = .08), #3 (P = .002), #5 (P < .0001), and #7 (P < .0001), suggesting these surgeons were less costly than predicted. In contrast, the adjusted cost was lower for surgeons #2 (P = .128), #4 (P < .0001), and #6 (P = .44), which suggests that these surgeons were more costly than predicted after adjusting for comorbidities. FIGURE 3. View largeDownload slide The 90-day adjusted cost vs actual 90-day cost for each participating surgeon. The vertical axis represents actual median 90-day cost ($27 565). The black dots represent mean comorbidity-adjusted 90-day cost for each surgeon and the colored dot represents the mean actual 90-day costs. The line joining the colored and black dots represents the difference between the actual and the adjusted cost. The adjusted cost was higher than the actual cost for surgeons #1 (P = .08), #3 (P = .002), #5 (P < .0001), and #7 (P < .0001), suggesting these surgeons were less costly than predicted. In contrast, the adjusted cost was lower for surgeons #2 (P = .128), #4 (P < .0001), and #6 (P = .44), which suggests that these surgeons were more costly than predicted after adjusting for comorbidities. TABLE 5. Multivariable Linear Regression Model to Derive Comorbidity-Adjusted Cost       95% CI for beta    Beta coefficient  P-value  Lower bound  Upper bound  Intercept  26832  <.001  22871  30793  Number of comorbidities  321  .736  –1550  2191  Age > 65 yr  1566  .028  165  2966  Smoker  1952  .037  116  3789  ASA grade  343  .652  –1148  1833  Diabetes  712  .58  –1810  3234  Hypertension  –856  .506  –3386  1673  Myocardial infarction  2698  .23  –1712  7108  Congestive heart failure  –1119  .701  –6841  4603  Chronic obstructive pulmonary disease  –117  .955  –4206  3972  Atrial fibrillation  –1974  .325  –5906  1958  Obesity (BMI > 65)  1044  .407  –1425  3513  Arthritis  –503  .697  –3033  2027  Osteoporosis  –1755  .41  –5933  2424  Preoperative anticoagulation  –75  .976  –5016  4866        95% CI for beta    Beta coefficient  P-value  Lower bound  Upper bound  Intercept  26832  <.001  22871  30793  Number of comorbidities  321  .736  –1550  2191  Age > 65 yr  1566  .028  165  2966  Smoker  1952  .037  116  3789  ASA grade  343  .652  –1148  1833  Diabetes  712  .58  –1810  3234  Hypertension  –856  .506  –3386  1673  Myocardial infarction  2698  .23  –1712  7108  Congestive heart failure  –1119  .701  –6841  4603  Chronic obstructive pulmonary disease  –117  .955  –4206  3972  Atrial fibrillation  –1974  .325  –5906  1958  Obesity (BMI > 65)  1044  .407  –1425  3513  Arthritis  –503  .697  –3033  2027  Osteoporosis  –1755  .41  –5933  2424  Preoperative anticoagulation  –75  .976  –5016  4866  BMI, body mass index. View Large DISCUSSION Current trends in strategies to target cost-containment are specially aimed at reducing variability in cost and outcomes. A number of studies have defined the variability in outcomes following spine surgery and a handful of studies have defined the variability in cost within each individual spine-related DRG.8,9,11,12,27 None of the prior studies have demonstrated the variations in cost and outcomes at the individual surgeon level. In these analyses, utilizing prospectively collected registry data from a single-center, we demonstrate that there were significant differences in total 90-d costs and patient-specific factors among the surgeons. Despite these differences, no significant differences in improvement in the disability, pain, and quality of life outcomes 90 d after surgery were observed. Some surgeons had higher odds of performing high-cost fusion surgery. There was a significant variability in comorbidity-adjusted cost vs actual cost among the surgeons. After adjusting for preoperative comorbidities, the adjusted costs were higher than actual cost for some surgeons and lower than the actual cost for others. Our study provides valuable insights into variations in patient characteristics, outcomes, and costs among the participating surgeons at a single institution. This study can form the basis to stimulate action to improve uniformity and cost-containment for lumbar fusion surgery. Number of factors can result in variations in the cost associated with fusion surgery. It is possible that the interplay of patient-specific factors and surgery-specific factors may also induce variation in the utilization of health care services and therefore induce variability in total 90-d costs. Prior studies have demonstrated that the patients’ comorbidity burden influence 90-d costs.13-17 In this study, we have sought to take one step further, however, and simulate a workflow for surgeon-level process improvement. Figure 3 reveals which surgeons incurred average costs exceeding risk-adjusted predictions, and which surgeons were less costly than predicted. This finding highlights that there may be several exogenous factors, in addition to the patient-specific comorbidities that are tied to the way surgeons practice that influences the total 90-d costs. We can now turn our attention to a particular surgeon who was costlier than expected and investigate his or her specific cost distribution. For example, surgeons #4 and #6 had higher frequency of performing interbody graft fusion compared to other surgeons. However, after adjusting for interbody graft and other surgery-specific variables the surgeons #4 and #6 were high-cost surgeons. This suggests that there might be confounders beyond the variables included in the model. For example, surgeon #6 had higher postdischarge health care visit associated costs and surgeon #4 had higher hospital cost. Further, granular data with details on type of implant used, additional intraoperative costs, and breakdown of the cost associated with inpatient hospital stay is needed to accurately determine the practice pattern of high-cost surgeons. A habitual practice of examining surgeon-level cost variability in this manner, when coupled with evaluations of modifiable risk factors within surgeons’ high-cost patient populations, can provide the makings of a true “learning health care system” as imagined by the Institute of Medicine.28 Consistent with previous studies, in our analysis, the high cost for fusion surgery was also associated with postoperative complication and readmission, increased surgery duration, and extended length of hospital stay. Complications and readmission within 90 d global period occur at a consistent frequency, specifically when analyzing the larger data sets.20,29-32 The factors including patient age, obesity, associated comorbidities, primary diagnosis, and surgical invasiveness and complexity are associated with higher likelihood of developing complications and also influence the cost and outcomes following surgery.18,19,30,33-38 We demonstrate that complications do contribute to higher cost for fusion surgery; however, there was no difference in the complication rate among the participating surgeons. This suggests that occurrence of complication alone might not explain the variation in cost among surgeons. Clearly, measures focused on prevention of complications will be able to decrease the cost and therefore increase the cost-benefit ratio. Analogous to the previous studies, extended duration of surgery and length of hospital stay were significant drivers of the total 90-d cost. The surgery duration and length of hospital stay were significantly different among the participating surgeons. Identifying the factors associated with this variability might be able to precisely determine and remedy the modifiable factors and contain the escalating costs associated with fusion. In our study, surgeons #4 and #6 had higher odds of performing high-cost fusion surgery. There is a tendency for a surgeon to account for the high cost by stating that they care for sicker patients. In a retrospective study, Walid et al13 demonstrated that the comorbidities additively increase the cost associated with spine surgery. Several other authors have emphasized the importance of considering the patients’ comorbidities as a driver of cost and outcomes following spine surgery.13-20 None of the previous studies have demonstrated the surgeon level variability in cost accounted for the comorbidities. In our study after adjusting for comorbidities, surgeon #4 was found to be costlier than expected and surgeon #6 was found to be less costly than expected, based on the health state of their respective patients. Furthermore, surgeon #2 was not a high-cost surgeon overall; however, this surgeon was found to be costlier than expected, when adjusting for comorbidities. This suggests that there is no doubt that preoperative comorbidities should be accounted for when deriving the aggregate total cost. However, as mentioned above, there are other factors beyond comorbidities that influence the total 90-d cost. The value-based pay-per-performance for an episode of care (30- or 90-d) initiatives are conceptually simple and have potential to improvise the health-care quality and contain escalating costs.39-45 The ideal episode of care model should be well designed to accommodate the highs and lows in each component of total 90-d cost. Given the heterogeneity of patient-specific and surgeon-specific factors, implementation of these models in real-world practice is complex and challenging. As demonstrated in our analyses, there is significant variation in the cost and characteristics of the patients managed by each surgeon at a single institution. Without fuller accountability of these variations in each component of the cost, the episode of care model can potentially be seen as an infringement of the individual surgeons’ autonomy. Therefore, it is imperative to involve surgeons in the decision-making and to consider variability in cost at the individual surgeon level after adjusting for complexity of the patient population managed by each surgeon. Limitations There are several limitations associated with this study. This is a single-institution study, and the costs were adjusted based for Medicare allowable 2013 national payments amount. The institution costs may vary for private payers and if a different geographical multiplier is used. We used unit multiple to eliminate any geographical variation. Furthermore, we do not have granular data on different types of instrumentation, use of infuse, or type of biologics used by each surgeon. This will certainly be important as hospitals try to better understand cost saving measures to improve profit margin. This will require a partnership with surgeons, so the costs are cut but care is not compromised. Our institution, similar to other institutions will not allow detailed information regarding unique cost arrangements with companies to be published. The data on postdischarge utilization is gleaned from the electronic medical records and augmented by patient interview (to capture care outside of the facility) and is subject to recall bias. Furthermore, the indirect cost from societal perspective including patient workday losses, family work day loss, caregiver cost was not included in the cost calculation, as the goal of this study was to define variability in cost from payers’ and providers’ perspective. The confounding variables included in the analysis may not be exhaustive, therefore adding more variables might account for other variation in the 90-d costs. Finally, the number of patients operated on by each surgeon was different. Sensitivity analysis was performed. A separate model was fitted excluding patients operated on by surgeons #5 and #6. There were no differences in the effect size and model performance (AUC—0.881) between this model and all-inclusive model (AUC—0.885; Tables, Supplemental Digital Content 1 and 2). The generalized application of these data needs to consider these limitations. Nonetheless, utilizing comprehensive list of variables captured in a single-center prospective longitudinal spine registry, we present the variations in 90-d cost and outcomes for laminectomy and fusion for lumbar degenerative pathology. CONCLUSION Our study provides valuable insight into variations in 90-d costs among the surgeons performing elective lumbar laminectomy/fusions at a single institution. Specific surgeons were found to have greater odds of performing high-cost surgeries. Adjusting for preoperative comorbidities, however, led to costs that were higher than the actual costs for certain surgeons and lower than the actual costs for others. Patient's preoperative comorbidities must therefore be accounted for when crafting value-based bundled payment models. More broadly, this study demonstrates that the designing intervention targeting “modifiable” factors tied to the way surgeons practice may increase the overall value of lumbar laminectomy and fusion. Disclosure The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. REFERENCES 1. Deyo RA, Gray DT, Kreuter W, Mirza S, Martin BI. United States trends in lumbar fusion surgery for degenerative conditions. Spine . 2005; 30( 12): 1441- 1445; discussion 1446-1447. Google Scholar CrossRef Search ADS PubMed  2. Dagenais S, Caro J, Haldeman S. A systematic review of low back pain cost of illness studies in the United States and internationally. Spine J . 2008; 8( 1): 8- 20. Google Scholar CrossRef Search ADS PubMed  3. Frymoyer JW, Cats-Baril WL. An overview of the incidences and costs of low back pain. Orthop Clin North Am . 1991; 22( 2): 263- 271. Google Scholar PubMed  4. Martin BI, Deyo RA, Mirza SK et al.   Expenditures and health status among adults with back and neck problems. JAMA . 2008; 299( 6): 656- 664. Google Scholar CrossRef Search ADS PubMed  5. Martin BI, Turner JA, Mirza SK, Lee MJ, Comstock BA, Deyo RA. Trends in health care expenditures, utilization, and health status among US adults with spine problems, 1997-2006. Spine . 2009; 34( 19): 2077- 2084. Google Scholar CrossRef Search ADS PubMed  6. Get ready: bundled payments are in your future. Hosp Case Manag . 2015; 23( 9): 109- 112. PubMed  7. Bozic KJ, Ward L, Vail TP, Maze M. Bundled payments in total joint arthroplasty: targeting opportunities for quality improvement and cost reduction. Clin Orthop Relat Res . 2014; 472( 1): 188- 193. Google Scholar CrossRef Search ADS PubMed  8. Ugiliweneza B, Kong M, Nosova K et al.   Spinal surgery: variations in health care costs and implications for episode-based bundled payments. Spine . 2014; 39( 15): 1235- 1242. Google Scholar CrossRef Search ADS PubMed  9. Birkmeyer JD, Gust C, Baser O, Dimick JB, Sutherland JM, Skinner JS. Medicare payments for common inpatient procedures: implications for episode-based payment bundling. Health Serv Res . 2010; 45( 6 pt 1): 1783- 1795. Google Scholar CrossRef Search ADS PubMed  10. Grenda TR, Pradarelli JC, Thumma JR, Dimick JB. Variation in hospital episode costs with bariatric surgery. JAMA Surg . 2015; 150( 12): 1109- 1115. Google Scholar CrossRef Search ADS PubMed  11. Schoenfeld AJ, Harris MB, Liu H, Birkmeyer JD. Variations in Medicare payments for episodes of spine surgery. Spine J . 2014; 14( 12): 2793- 2798. Google Scholar CrossRef Search ADS PubMed  12. Miller DC, Gust C, Dimick JB, Birkmeyer N, Skinner J, Birkmeyer JD. Large variations in Medicare payments for surgery highlight savings potential from bundled payment programs. Health Aff . 2011; 30( 11): 2107- 2115. Google Scholar CrossRef Search ADS   13. Walid MS, Robinson JS Jr. Economic impact of comorbidities in spine surgery. J Neurosurg Spine.  2011; 14( 3): 318- 321. Google Scholar CrossRef Search ADS PubMed  14. Walid MS, Robinson EC, Robinson JS Jr. Higher comorbidity rates in unemployed patients may significantly impact the cost of spine surgery. J Clin Neurosci . 2011; 18( 5): 640- 644. Google Scholar CrossRef Search ADS PubMed  15. Kalanithi PA, Arrigo R, Boakye M. Morbid obesity increases cost and complication rates in spinal arthrodesis. Spine . 2012; 37( 11): 982- 988. Google Scholar CrossRef Search ADS PubMed  16. Devin CJ, Chotai S, Parker SL, Tetreault L, Fehlings MG, McGirt MJ. A cost-utility analysis of lumbar decompression with and without fusion for degenerative spine disease in the elderly. Neurosurgery . 2015; 77( suppl 4): S116- S124. Google Scholar CrossRef Search ADS PubMed  17. Chotai S, Sielatycki JA, Parker SL et al.   Effect of obesity on cost per quality-adjusted life years gained following anterior cervical discectomy and fusion in elective degenerative pathology. Spine J . 2016; 16( 11): 1342- 1350. Google Scholar CrossRef Search ADS PubMed  18. Wang MY, Green BA, Shah S, Vanni S, Levi AD. Complications associated with lumbar stenosis surgery in patients older than 75 years of age. Neurosurgical Focus . 2003; 14( 2): e7. Google Scholar CrossRef Search ADS PubMed  19. Campbell PG, Yadla S, Nasser R, Malone J, Maltenfort MG, Ratliff JK. Patient comorbidity score predicting the incidence of perioperative complications: assessing the impact of comorbidities on complications in spine surgery. J Neurosurg Spine.  2012; 16( 1): 37- 43. Google Scholar CrossRef Search ADS PubMed  20. Peter GC, Sanjay Y, Rani N, Jennifer M, Mitchell GM, John KR. Patient comorbidity score predicting the incidence of perioperative complications: assessing the impact of comorbidities on complications in spine surgery. J Neurosurg Spine . 2012; 16( 1): 37- 43. Google Scholar CrossRef Search ADS PubMed  21. Fairbank JC, Pynsent PB. The oswestry disability index. Spine . 2000; 25( 22): 2940- 2952; discussion 2952. Google Scholar CrossRef Search ADS PubMed  22. Langley GB, Sheppeard H. The visual analogue scale: its use in pain measurement. Rheumatol Int . 1985; 5( 4): 145- 148. Google Scholar CrossRef Search ADS PubMed  23. EuroQol. –a new facility for the measurement of health-related quality of life. Health Pol . 1990; 16( 3): 199- 208. CrossRef Search ADS   24. Parker SL, Fulchiero EC, Davis BJ et al.   Cost-effectiveness of multilevel hemilaminectomy for lumbar stenosis-associated radiculopathy. Spine J . 2011; 11( 8): 705- 711. Google Scholar CrossRef Search ADS PubMed  25. Adogwa O, Parker SL, Shau D et al.   Cost per quality-adjusted life year gained of revision fusion for lumbar pseudoarthrosis: defining the value of surgery. J Spinal Disord Tech . 2015; 28( 3): 101- 105. Google Scholar CrossRef Search ADS PubMed  26. Parker SL, Godil SS, Mendenhall SK, Zuckerman SL, Shau DN, McGirt MJ. Two-year comprehensive medical management of degenerative lumbar spine disease (lumbar spondylolisthesis, stenosis, or disc herniation): a value analysis of cost, pain, disability, and quality of life: clinical article. J Neurosurg Spine.  2014; 21( 2): 143- 149. Google Scholar CrossRef Search ADS PubMed  27. McGirt MJ, Sivaganesan A, Asher AL, Devin CJ. Prediction model for outcome after low-back surgery: individualized likelihood of complication, hospital readmission, return to work, and 12-month improvement in functional disability. Neurosurgical Focus . 2015; 39( 6): E13. Google Scholar CrossRef Search ADS PubMed  28. Smith M. Available at: http://www.nationalacademies.org/hmd/Activities/Quality/LearningHealthCare.aspx Accessed September 2016. 29. Carreon LY, Puno RM, Dimar JR 2nd, Glassman SD, Johnson JR. Perioperative complications of posterior lumbar decompression and arthrodesis in older adults. J Bone Joint Surg Am . 2003; 85-a( 11): 2089- 2092. Google Scholar CrossRef Search ADS PubMed  30. Cassinelli EH, Eubanks J, Vogt M, Furey C, Yoo J, Bohlman HH. Risk factors for the development of perioperative complications in elderly patients undergoing lumbar decompression and arthrodesis for spinal stenosis: an analysis of 166 patients. Spine . 2007; 32( 2): 230- 235. Google Scholar CrossRef Search ADS PubMed  31. Hoffman RM, Wheeler KJ, Deyo RA. Surgery for herniated lumbar discs: a literature synthesis. J Gen Int Med . 1993; 8( 9): 487- 496. Google Scholar CrossRef Search ADS   32. Turner JA, Loeser JD, Deyo RA, Sanders SB. Spinal cord stimulation for patients with failed back surgery syndrome or complex regional pain syndrome: a systematic review of effectiveness and complications. Pain . 2004; 108( 1-2): 137- 147. Google Scholar CrossRef Search ADS PubMed  33. Bekelis K, Desai A, Bakhoum SF, Missios S. A predictive model of complications after spine surgery: the National Surgical Quality Improvement Program (NSQIP) 2005-2010. Spine J . 2014; 14( 7): 1247- 1255. Google Scholar CrossRef Search ADS PubMed  34. Faciszewski T, Winter RB, Lonstein JE, Denis F, Johnson L. The surgical and medical perioperative complications of anterior spinal fusion surgery in the thoracic and lumbar spine in adults. A review of 1223 procedures. Spine . 1995; 20( 14): 1592- 1599. Google Scholar CrossRef Search ADS PubMed  35. Glassman SD, Alegre G, Carreon L, Dimar JR, Johnson JR. Perioperative complications of lumbar instrumentation and fusion in patients with diabetes mellitus. Spine J . 2003; 3( 6): 496- 501. Google Scholar CrossRef Search ADS PubMed  36. Guzman JZ, Iatridis JC, Skovrlj B et al.   Outcomes and complications of diabetes mellitus on patients undergoing degenerative lumbar spine surgery. Spine . 2014; 39( 19): 1596- 1604. Google Scholar CrossRef Search ADS PubMed  37. Schwender JD, Casnellie MT, Perra JH et al.   Perioperative complications in revision anterior lumbar spine surgery: incidence and risk factors. Spine . 2009; 34( 1): 87- 90. Google Scholar CrossRef Search ADS PubMed  38. Kalanithi PS, Patil CG, Boakye M. National complication rates and disposition after posterior lumbar fusion for acquired spondylolisthesis. Spine . 2009; 34( 18): 1963- 1969. Google Scholar CrossRef Search ADS PubMed  39. Cutler DM, Ghosh K. The potential for cost savings through bundled episode payments. N Engl J Med . 2012; 366( 12): 1075- 1077. Google Scholar CrossRef Search ADS PubMed  40. Delisle DR. Big things come in bundled packages: implications of bundled payment systems in health care reimbursement reform. Am J Med Qual . 2013; 28( 4): 339- 344. Google Scholar CrossRef Search ADS PubMed  41. Iorio R. Strategies and tactics for successful implementation of bundled payments: bundled payment for care improvement at a large, urban, academic medical center. J Arthroplasty . 2015; 30( 3): 349- 350. Google Scholar CrossRef Search ADS PubMed  42. Kazberouk A, McGuire K, Landon BE. A survey of innovative reimbursement models in spine care. Spine . 2016; 41( 4): 344- 352. Google Scholar CrossRef Search ADS PubMed  43. Mechanic RE. Opportunities and challenges for episode-based payment. N Engl J Med . 2011; 365( 9): 777- 779. Google Scholar CrossRef Search ADS PubMed  44. Rossi VJ, Ahn J, Bohl DD, Tabaraee E, Singh K. Economic factors in the future delivery of spinal healthcare. World J Orthop . 2015; 6( 5): 409- 412. Google Scholar CrossRef Search ADS PubMed  45. Sood N, Huckfeldt PJ, Escarce JJ, Grabowski DC, Newhouse JP. Medicare's bundled payment pilot for acute and postacute care: analysis and recommendations on where to begin. Health Aff . 2011; 30( 9): 1708- 1717. Google Scholar CrossRef Search ADS   Supplemental digital content is available for this article at www.neurosurgery-online.com. COMMENTS In this manuscript, the authors have presented a cost-utilization and outcome analysis of inter-surgeon variability for elective lumbar decompression and fusion. The authors have outlined the variations in patient profile, cost, and outcomes for each surgeon performing laminectomy and fusion for degenerative spine disease at a single center. The authors found that when adjusted for preoperative comorbidities, the costs for certain surgeons were found to be higher than the comorbidity-adjusted 90-day cost. As a per surgeon utilization study, the manuscript represents an important addition to the literature. Single center studies such as this are uniquely situated to assess per surgeon costs as they control for institutional variables. With that said, a multi-center study would provide insight into cost factors that remain out of the surgeons control. In one of the supplemental tables, the authors provide an analysis that suggests that the higher cost surgeons fused more levels and had longer Operating Room time, indicating that they were doing more involved cases than the low-cost surgeons. How much of the cost difference does this explain? In addition to adjusting for preoperative morbidities, the authors should consider adjusting for intraoperative factors (type of fusion surgery, length of construct, etc). Moreover, 3 of 7 surgeons did 33 or fewer cases between 2011 and 2015, with 1 surgeon doing only 9 cases in that period. The other surgeons in the study did well over 100 cases each. Although in some analyses 2 of the 3 lower-volume surgeons were excluded, it is unclear the relationship this had to the overall cost driver analysis. In future analyses, the authors may consider delving into the reasons that explain the cost variations. For instance, what aspects of patient care or operative factors (implantable devices, etc) drive the costs for each surgeon? It would also be interesting to see how the publication of the article has impacted the cost decisions of the surgeons identified as higher cost relative to their colleagues. Overall, this article highlights an important topic: surgeon variability and its relationship to costs. Mohamad Bydon Rochester, Minnesota The authors examined a longitudinal registry from a single institution for 90-day costs and outcomes for 752 patients undergoing laminectomy and fusion surgery. They found that while there were no differences in the complication rate, 90-day readmission rate, or measurements of clinical outcome, 2 of the 7 surgeons incurred significantly higher costs. The subject relates to critical analysis of the cost for the care that spinal surgeons deliver—a matter universally recognized as an area we as a society need to better understand to permit the best care that we can afford. The study is limited in providing any practical information about the value of any particular surgeon's approach to operative management of degenerative spine disease; as the authors note that they lacked any "granular data" to assess the various surgeons with regard to use of particular spinal implants, biologics, or costs associated with the inpatient stay. Still, the study is useful in providing needed data to validate the importance of considering the preoperative morbidities in creating bundled payment models. John Kenneth Houten Bronx, New York Copyright © 2017 by the Congress of Neurological Surgeons

Journal

NeurosurgeryOxford University Press

Published: Apr 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off