Development of machine learning algorithms for prediction of prolonged opioid prescription after surgery for lumbar disc herniation

BACKGROUND CONTEXT

Spine surgery has been identified as a risk factor for prolonged postoperative opioid use. Preoperative prediction of opioid use could improve risk stratification, shared decision-making, and patient counseling before surgery.

PURPOSE

The primary purpose of this study was to develop algorithms for prediction of prolonged opioid prescription after surgery for lumbar disc herniation.

STUDY DESIGN/SETTING

Retrospective, case-control study at five medical centers.

PATIENT SAMPLE

Chart review was conducted for patients undergoing surgery for lumbar disc herniation between January 1, 2000 and March 1, 2018.

OUTCOME MEASURES

The primary outcome of interest was sustained opioid prescription after surgery to at least 90 to 180 days postoperatively.

METHODS

Five models (elastic-net penalized logistic regression, random forest, stochastic gradient boosting, neural network, and support vector machine) were developed to predict prolonged opioid prescription. Explanations of predictions were provided globally (averaged across all patients) and locally (for individual patients).

RESULTS

Overall, 5,413 patients were identified, with sustained postoperative opioid prescription of 416 (7.7%) at 90 to 180 days after surgery. The elastic-net penalized logistic regression model had the best discrimination (c-statistic 0.81) and good calibration and overall performance; the three most important predictors were: instrumentation, duration of preoperative opioid prescription, and comorbidity of depression. The final models were incorporated into an open access web application able to provide predictions as well as patient-specific explanations of the results generated by the algorithms. The application can be found here: https://sorg-apps.shinyapps.io/lumbardiscopioid/ CONCLUSION: Preoperative prediction of prolonged postoperative opioid prescription can help identify candidates for increased surveillance after surgery. Patient-centered explanations of predictions can enhance both shared decision-making and quality of care.

Validating the Stopping Opioids after Surgery (SOS) score for sustained postoperative prescription opioid use in spine surgical patients

BACKGROUND CONTEXT

The opioid epidemic has increased scrutiny of health-care practices and care episodes, such as surgery, that increase the risk of opioid dependence. The Stopping Opioids after Surgery (SOS) score to predict sustained prescription opioid use was previously developed within a population of patients receiving general surgery, orthopedic, and urologic procedures. Notably, the performance for this score has not been assessed in a spine surgical cohort.

PURPOSE

We sought to validate the SOS score in a series of patients undergoing cervical and lumbar spine surgery, including inpatient and outpatient cohorts.

STUDY DESIGN/SETTING

Retrospective review at two academic medical centers and three community hospitals.

OUTCOME MEASURES

Sustained prescription opioid use was defined as opioid prescription without interruption for 90 days or longer following surgery.

METHODS

The performance of the SOS score was assessed in the study population by calculating the c-statistic, receiver-operating curve, and observed rates of sustained prescription opioid use.

RESULTS

Among 7,027 patients included in this study, 2,374 (33.8%) underwent anterior cervical discectomy and fusion and 4,653 (66.2%) underwent surgery for lumbar disc herniation. The median age was 46 (interquartile range=38.0-53.5). Overall, 604 patients (8.6%) had prolonged opioid prescription. The c-statistic of the risk score was 0.764. The sensitivity of the score at the low risk cutoff of 30 was 0.72. At the high-risk cutoff of 60, the specificity was 0.99. The observed risk (95% confidence interval) of prolonged opioid prescription was 3.6% (3.1-4.2) in the low-risk group (scores <30), 17.2% (15.6-18.7) in the intermediate-risk group (scores 30-60), and 46.0% (36.2-55.9) in the high-risk group (scores >60).

CONCLUSIONS

We have validated the use of a clinically relevant bedside risk score for sustained prescription opioid use after spine surgery. The score's ease of use, combined with its exceptional performance, renders it a valuable tool for spine care providers in counseling patients and determining appropriate postdischarge management to prevent sustained opioid use.

Machine learning for prediction of sustained opioid prescription after anterior cervical discectomy and fusion

BACKGROUND CONTEXT

The severity of the opioid epidemic has increased scrutiny of opioid prescribing practices. Spine surgery is a high-risk episode for sustained postoperative opioid prescription.

PURPOSE

To develop machine learning algorithms for preoperative prediction of sustained opioid prescription after anterior cervical discectomy and fusion (ACDF).

STUDY DESIGN/SETTING

Retrospective, case-control study at two academic medical centers and three community hospitals.

PATIENT SAMPLE

Electronic health records were queried for adult patients undergoing ACDF for degenerative disorders between January 1, 2000 and March 1, 2018.

OUTCOME MEASURES

Sustained postoperative opioid prescription was defined as uninterrupted filing of prescription opioid extending to at least 90-180 days after surgery.

METHODS

Five machine learning models were developed to predict postoperative opioid prescription and assessed for overall performance.

RESULTS

Of 2,737 patients undergoing ACDF, 270 (9.9%) demonstrated sustained opioid prescription. Variables identified for prediction of sustained opioid prescription were male sex, multilevel surgery, myelopathy, tobacco use, insurance status (Medicaid, Medicare), duration of preoperative opioid use, and medications (antidepressants, benzodiazepines, beta-2-agonist, angiotensin-converting enzyme-inhibitors, gabapentin). The stochastic gradient boosting algorithm achieved the best performance with c-statistic=0.81 and good calibration. Global explanations of the model demonstrated that preoperative opioid duration, antidepressant use, tobacco use, and Medicaid insurance were the most important predictors of sustained postoperative opioid prescription.

CONCLUSIONS

One-tenth of patients undergoing ACDF demonstrated sustained opioid prescription following surgery. Machine learning algorithms could be used to preoperatively stratify risk these patients, possibly enabling early intervention to reduce the potential for long-term opioid use in this population.

Correction to: Spectrum of diagnostic errors in cervical spine trauma imaging and their clinical significance

The published version of this article unfortunately contained a mistake. Author given and family name Alessandrino Francesco was incorrectly interchanged. The correct presentation is given above. The original article has been corrected.

Spectrum of diagnostic errors in cervical spine trauma imaging and their clinical significance

PURPOSE

To describe and categorize diagnostic errors in cervical spine CT (CsCT) interpretation performed for trauma and to assess their clinical significance.

METHODS

All CsCTs performed for trauma with diagnostic errors that came to our attention based on clinical or imaging follow-up or quality assurance peer review from 2004 to 2017 were included. The number of CsCTs performed at our institution during the same time interval was calculated. Errors were categorized as spinal/extraspinal, involving osseous/soft tissue structures, by anatomical site and level. Images were reviewed by a radiologist and two spine surgeons. For each error, the need for surgery, immobilization, CT angiogram of the neck, and MRI was assessed; if any of these were needed, the error was considered clinically significant.

RESULTS

Of an approximate total 59,000 CsCTs, 56 reports containing diagnostic errors were included. Twelve were extraspinal, and 44 were spinal (26 fractures, 15 intervertebral disc protrusions, two subluxations, one lytic bone lesion). The most common sites of spinal fractures were vertebral body (n = 10) and transverse process (n = 8); the most common levels were C5 (n = 8) and C7 (n = 6). All (n = 26) fractures and two atlantooccipital subluxations were considered clinically significant, including three patients who would have required urgent surgical stabilization (two subluxations and one facet fracture). Two transverse processes fractures did not alter the need for surgical intervention/surgical approach, immobilization, or MRI.

CONCLUSIONS

In our study, 66% of spinal diagnostic errors on CsCT were considered clinically significant, potentially altering clinical management. Transverse process and vertebral body fractures were commonly missed.

Development of a machine learning algorithm predicting discharge placement after surgery for spondylolisthesis

PURPOSE

We aimed to develop a machine learning algorithm that can accurately predict discharge placement in patients undergoing elective surgery for degenerative spondylolisthesis.

METHODS

The National Surgical Quality Improvement Program (NSQIP) database was used to select patients that underwent surgical treatment for degenerative spondylolisthesis between 2009 and 2016. Our primary outcome measure was non-home discharge which was defined as any discharge not to home for which we grouped together all non-home discharge destinations including rehabilitation facility, skilled nursing facility, and unskilled nursing facility. We used Akaike information criterion to select the most appropriate model based on the outcomes of the stepwise backward logistic regression. Four machine learning algorithms were developed to predict discharge placement and were assessed by discrimination, calibration, and overall performance.

RESULTS

Nine thousand three hundred and thirty-eight patients were included. Median age was 63 (interquartile range [IQR] 54-71), and 63% (n = 5,887) were female. The non-home discharge rate was 18.6%. Our models included age, sex, diabetes, elective surgery, BMI, procedure, number of levels, ASA class, preoperative white blood cell count, and preoperative creatinine. The Bayes point machine was considered the best model based on discrimination (AUC = 0.753), calibration (slope = 1.111; intercept = - 0.002), and overall model performance (Brier score = 0.132).

CONCLUSION

This study has shown that it is possible to create a predictive machine learning algorithm with both good accuracy and calibration to predict discharge placement. Using our methodology, this type of model can be developed for many other conditions and (elective) treatments. These slides can be retrieved under Electronic Supplementary Material.

Predicting 90-Day and 1-Year Mortality in Spinal Metastatic Disease: Development and Internal Validation

BACKGROUND

Increasing prevalence of metastatic disease has been accompanied by increasing rates of surgical intervention. Current tools have poor to fair predictive performance for intermediate (90-d) and long-term (1-yr) mortality.

OBJECTIVE

To develop predictive algorithms for spinal metastatic disease at these time points and to provide patient-specific explanations of the predictions generated by these algorithms.

METHODS

Retrospective review was conducted at 2 large academic medical centers to identify patients undergoing initial operative management for spinal metastatic disease between January 2000 and December 2016. Five models (penalized logistic regression, random forest, stochastic gradient boosting, neural network, and support vector machine) were developed to predict 90-d and 1-yr mortality.

RESULTS

Overall, 732 patients were identified with 90-d and 1-yr mortality rates of 181 (25.1%) and 385 (54.3%), respectively. The stochastic gradient boosting algorithm had the best performance for 90-d mortality and 1-yr mortality. On global variable importance assessment, albumin, primary tumor histology, and performance status were the 3 most important predictors of 90-d mortality. The final models were incorporated into an open access web application able to provide predictions as well as patient-specific explanations of the results generated by the algorithms. The application can be found at https://sorg-apps.shinyapps.io/spinemetssurvival/

CONCLUSION

Preoperative estimation of 90-d and 1-yr mortality was achieved with assessment of more flexible modeling techniques such as machine learning. Integration of these models into applications and patient-centered explanations of predictions represent opportunities for incorporation into healthcare systems as decision tools in the future.

Is focused magnetic resonance imaging adequate for treatment decision making in acute traumatic thoracic and lumbar spine fractures seen on whole spine computed tomography?

PURPOSE

To assess whether a focused magnetic resonance imaging (MRI) limited to the region of known acute traumatic thoracic or lumbar fracture(s) would miss any clinically significant injuries that would change patient management.

STUDY DESIGN/SETTING

A multicenter retrospective clinical study.

PATIENT SAMPLE

Adult patients with acute traumatic thoracic and/or lumbar spine fracture(s).

OUTCOME MEASURES

Pathology identified on MRI (ligamentous disruption, epidural hematoma, and cord contusion), outside of the focused zone, an alteration in patient management, including surgical and nonsurgical, as a result of the identified pathology outside the focused zone.

METHODS

Records were reviewed for all adult trauma patients who presented to the emergency department between 2008 and 2016 with one or more fracture(s) of the thoracic and/or lumbar spine identified on computed tomography (CT) and who underwent MRI of the entire thoracic and lumbar spine within 10 days. Exclusion criteria were patients with >4 fractured levels, pathologic fractures, isolated transverse, and/or spinous process fractures, prior vertebral augmentation, and prior thoracic or lumbar spine instrumentation. Patients with neurologic deficits or cervical spine fractures were also included. MRIs were reviewed independently by one spine surgeon and one musculoskeletal fellowship-trained emergency radiologist for posterior ligamentous complex (PLC) integrity, vertebral injury, epidural hematoma, and cord contusion. The surgeon also commented on the clinical significance of the pathology identified outside the focused zone. All cases in which pathology was identified outside of the focused zone (three levels above and below the fractures) were independently reviewed by a second spine surgeon to determine whether the pathology was clinically significant and would alter the treatment plan.

RESULTS

In total, 126 patients with 216 fractures identified on CT were included, with a median age of 49 years. There were 81 males (64%). Sixty-two (49%) patients had isolated thoracolumbar junction injuries and 36 (29%) had injuries limited to a single fractured level. Forty-seven (37%) patients were managed operatively. PLC injury was identified by both readers in 36 (29%) patients with a percent agreement of 96% and κ coefficient of 0.91 (95% CI 0.87-0.95). Both readers independently agreed that there was no pathology identified on the complete thoracic and lumbar spine MRIs outside the focused zone in 107 (85%) patients. Injury outside the focused zone was identified by at least one reader in 19 (15%) patients. None of the readers identified PLC injury, cord edema, or noncontiguous epidural hematoma outside the focused zone. Percent agreement for outside pathology between the two readers was 92% with a κ coefficient of 0.60 (95% CI 0.48-0.72). The two spine surgeons independently agreed that none of the identified pathology outside of the focused zone altered management.

CONCLUSIONS

A focused MRI protocol of three levels above and below known thoracolumbar spine fractures would have missed radiological abnormality in 15% of patients. However, the pathology, such as vertebral body edema not appreciated on CT, was not clinically significant and did not alter patient care. Based on these findings, the investigators conclude that a focused protocol would decrease the imaging time while providing the information of the injured segment with minimal risk of missing any clinically significant injuries.

Injuries to the Rigid Spine: What the Spine Surgeon Wants to Know

The biomechanical stability of the spine is altered in patients with a rigid spine, rendering it vulnerable to fracture even from relatively minor impact. The rigid spine entities are ankylosing spondylitis (AS), diffuse idiopathic skeletal hyperostosis, degenerative spondylosis, and a surgically fused spine. The most common mechanism of injury resulting in fracture is hyperextension, which often leads to unstable injury in patients with a rigid spine per the recent AOSpine classification system. Due to the increased risk of spinal fractures in this population, performing a spine CT is the first step when a patient with a rigid spine presents with new back pain or suspected spinal trauma. In addition, there should be a low threshold for performing a non-contrast-enhanced spine MRI in patients with a rigid spine, especially those with AS who may have an occult fracture, epidural hematoma, or spinal cord injury. Unfortunately, owing to insufficient imaging and an unfamiliarity with fracture patterns in the setting of a rigid spine, fracture diagnosis is often delayed, leading to significant morbidity and even death. The radiologist's role is to recognize the imaging features of a rigid spine, identify any fractures at CT and MRI, and fully characterize the extent of injury. Reasons for surgical intervention include neurologic deficit or concern for deterioration, an unstable fracture, or the presence of an epidural hematoma. By understanding the imaging features of various rigid spine conditions and vigilantly examining images for occult fractures, the radiologist can avoid a missed or delayed diagnosis of an injured rigid spine. ^©RSNA, 2019.

Development of machine learning algorithms for prediction of discharge disposition after elective inpatient surgery for lumbar degenerative disc disorders

OBJECTIVE

If not anticipated and prearranged, hospital stay can be prolonged while the patient awaits placement in a rehabilitation unit or skilled nursing facility following elective spine surgery. Preoperative prediction of the likelihood of postoperative discharge to any setting other than home (i.e., nonroutine discharge) after elective inpatient spine surgery would be helpful in terms of decreasing hospital length of stay. The purpose of this study was to use machine learning algorithms to develop an open-access web application for preoperative prediction of nonroutine discharges in surgery for elective inpatient lumbar degenerative disc disorders.

METHODS

The American College of Surgeons National Surgical Quality Improvement Program was queried to identify patients who underwent elective inpatient spine surgery for lumbar disc herniation or lumbar disc degeneration between 2011 and 2016. Four machine learning algorithms were developed to predict nonroutine discharge and the best algorithm was incorporated into an open-access web application.

RESULTS

The rate of nonroutine discharge for 26,364 patients who underwent elective inpatient surgery for lumbar degenerative disc disorders was 9.28%. Predictive factors selected by random forest algorithms were age, sex, body mass index, fusion, level, functional status, extent and severity of comorbid disease (American Society of Anesthesiologists classification), diabetes, and preoperative hematocrit level. On evaluation in the testing set (n = 5273), the neural network had a c-statistic of 0.823, calibration slope of 0.935, calibration intercept of 0.026, and Brier score of 0.0713. On decision curve analysis, the algorithm showed greater net benefit for changing management over all threshold probabilities than changing management on the basis of the American Society of Anesthesiologists classification alone or for all patients or for no patients. The model can be found here: https://sorg-apps.shinyapps.io/discdisposition/.

CONCLUSIONS

Machine learning algorithms show promising results on internal validation for preoperative prediction of nonroutine discharges. If found to be externally valid, widespread use of these algorithms via the open-access web application by healthcare professionals may help preoperative risk stratification of patients undergoing elective surgery for lumbar degenerative disc disorders.

Variation in costs among surgeons for lumbar spinal stenosis

BACKGROUND CONTEXT

Lumbar spinal stenosis is a common condition in the elderly for which costs vary substantially by region. Comparing differences between surgeons from a single institution, thereby omitting regional variation, could aid in identifying factors associated with higher costs and individual drivers of costs. The use of decision aids (DAs) has been suggested as one of the possible tools for diminishing costs and cost variation.

PURPOSE

(1) To determine factors associated with higher costs for treatment of spinal stenosis in the first year after diagnosis in a single institution; (2) to find individual drivers of costs for providers with higher costs; and (3) to determine if the use of DAs can decrease costs and cost variability.

STUDY DESIGN

Retrospective cohort study.

PATIENT SAMPLE

A total of 10,858 patients in 18 different practices diagnosed with lumbar spinal stenosis between January 2003 and July 2015 in three associated hospitals of a single institution.

OUTCOME MEASURES

Mean cost for a patient per provider in US dollars within 1 year after diagnosis of lumbar spinal stenosis.

METHODS

We collected all diagnostic testing, office visits, injections, surgery, and occupational or physical therapy related to lumbar spinal stenosis within 1 year after initial diagnosis. We used multivariable linear regression to determine independent predictors for costs. Providers were grouped in tiers based on mean total costs per patient to find drivers of costs. To assess the DAs effect on costs and cost variability, we matched DA patients one-to-one with non-DA patients.

RESULTS

Male gender (β 0.10, 95% confidence interval [CI] 0.05-0.15, p<.001), seeing an additional provider (β 0.77, 95% CI 0.69-0.86, p<.001), and having an additional spine diagnosis (β 0.79, 95% CI 0.74-0.84, p<.001) were associated with higher costs. Providers in the high cost tier had more office visits (p<.001), more imaging procedures (p<.001), less occupational or physical therapy (p=.002), and less surgery (p=.001) compared with the middle tier. Eighty-two patients (0.76%) received a DA as part of their care; there was no statistically significant difference between the DA group and the matched group in costs (p=.975).

CONCLUSIONS

Male gender, seeing an additional provider, and having an additional spine diagnosis were independently associated with higher costs. The main targets for cost reduction we found are imaging procedures and number of office visits. Decision aids were not found to affect cost.

Geographic variations in clinical presentation and outcomes of decompressive surgery in patients with symptomatic degenerative cervical myelopathy: analysis of a prospective, international multicenter cohort study of 757 patients

BACKGROUND CONTEXT

Degenerative cervical myelopathy (DCM) is a progressive degenerative spine disease and the most common cause of spinal cord impairment in adults worldwide. Few studies have reported on regional variations in demographics, clinical presentation, disease causation, and surgical effectiveness.

PURPOSE

The objective of this study was to evaluate differences in demographics, causative pathology, management strategies, surgical outcomes, length of hospital stay, and complications across four geographic regions.

STUDY DESIGN/SETTING

This is a multicenter international prospective cohort study.

PATIENT SAMPLE

This study includes a total of 757 symptomatic patients with DCM undergoing surgical decompression of the cervical spine.

OUTCOME MEASURES

The outcome measures are the Neck Disability Index (NDI), the Short Form 36 version 2 (SF-36v2), the modified Japanese Orthopaedic Association (mJOA) scale, and the Nurick grade.

MATERIALS AND METHODS

The baseline characteristics, disease causation, surgical approaches, and outcomes at 12 and 24 months were compared among four regions: Europe, Asia Pacific, Latin America, and North America.

RESULTS

Patients from Europe and North America were, on average, older than those from Latin America and Asia Pacific (p=.0055). Patients from Latin America had a significantly longer duration of symptoms than those from the other three regions (p<.0001). The most frequent causes of myelopathy were spondylosis and disc herniation. Ossification of the posterior longitudinal ligament was most prevalent in Asia Pacific (35.33%) and in Europe (31.75%), and hypertrophy of the ligamentum flavum was most prevalent in Latin America (61.25%). Surgical approaches varied by region; the majority of cases in Europe (71.43%), Asia Pacific (60.67%), and North America (59.10%) were managed anteriorly, whereas the posterior approach was more common in Latin America (66.25%). At the 24-month follow-up, patients from North America and Asia Pacific exhibited greater improvements in mJOA and Nurick scores than those from Europe and Latin America. Patients from Asia Pacific and Latin America demonstrated the most improvement on the NDI and SF-36v2 PCS. The longest duration of hospital stay was in Asia Pacific (14.16 days), and the highest rate of complications (34.9%) was reported in Europe.

CONCLUSIONS

Regional differences in demographics, causation, and surgical approaches are significant for patients with DCM. Despite these variations, surgical decompression for DCM appears effective in all regions. Observed differences in the extent of postoperative improvements among the regions should encourage the standardization of care across centers and the development of international guidelines for the management of DCM.

Critical analysis of trends in lumbar fusion for degenerative disorders revisited: influence of technique on fusion rate and clinical outcomes

PURPOSE

Lumbar fusion for degenerative disorders is among the most common spine surgical procedures performed. The purpose of this study was to analyze fusion, complications, and clinical success for lumbar fusion performed with various surgical techniques as reported in the literature from 2000 to 2015 and compare with previous critical analysis of outcomes from 1980 to 2000.

METHODS

A systematic review of the literature to identify all studies of adult lumbar fusion for degenerative disorders published between January 1, 2000, and August 31, 2015, was performed adhering to PRISMA guidelines. Studies were included if they enabled analysis of outcomes of individual fusion techniques.

RESULTS

Data from 8599 patients extracted from 160 studies were recorded. Posterior and transforaminal lumbar interbody fusion (PLIF and TLIF) had significantly higher fusion rates compared to instrumented posterolateral fusion (PLF) (OR 3.20 and 2.46, respectively). Clinical success rate was statistically higher with MIS versus non-MIS fusion (OR 2.44). While methodological quality was higher in studies from 2000 to 2015 than prior decades, the outcomes of comparable procedures were about the same.

CONCLUSIONS

Lumbar fusions for degenerative disorders from 2000 to 2015 demonstrate a trend toward more interbody fusions and MIS techniques than prior decades. Clinical success with MIS appears more likely than with non-MIS fusions, despite equivalent fusion and complication rates. While these data are intriguing, they should be interpreted cautiously considering the level of heterogeneity of the studies available. Further, high-quality comparative studies are warranted to better understand the relative benefits of more complex interbody and MIS fusions for these conditions. These slides can be retrieved under Electronic Supplementary Material.

Association of Annular Defect Width After Lumbar Discectomy With Risk of Symptom Recurrence and Reoperation: Systematic Review and Meta-analysis of Comparative Studies

STUDY DESIGN

Systematic review and meta-analysis of comparative studies.

OBJECTIVE

To characterize the association of annular defect width after lumbar discectomy with the risk of symptom recurrence and reoperation.

SUMMARY OF BACKGROUND DATA

Large annular defect width after lumbar discectomy has been reported to increase risk of symptom recurrence. However, this association has not been evaluated in a systematic manner.

METHODS

A systematic literature search of MEDLINE and EMBASE was performed to identify comparative studies of large versus small annular defects following lumbar discectomy that reported symptom recurrence or reoperation rates. Main outcomes were reported with pooled odds ratios (OR) and 95% confidence intervals (CIs). Sensitivity analyses were performed to assess the robustness of the meta-analysis findings.

RESULTS

After screening 696 records, we included data from 7 comparative studies involving 1653 lumbar discectomy patients, of whom 499 (30%) had large annular defects and 1154 (70%) had small annular defects. Methodological quality of studies was good overall. The median follow-up period was 2.9 years. The risk of symptom recurrence (OR = 2.5, 95% CI = 1.3-4.5, P = 0.004) and reoperation (OR = 2.3, 95% CI = 1.5-3.7, P < 0.001) was higher in patients with large versus small annular defects. Publication bias was not evident. The associations between annular defect width and risk of symptom recurrence and reoperation remained statistically significant in all sensitivity analyses.

CONCLUSION

Annular defect width after lumbar discectomy is an under-reported modifier of patient outcome. Risk for symptom recurrence and reoperation is higher in patients with large versus small annular defects following lumbar discectomy.

LEVEL OF EVIDENCE

2.

Establishing benchmarks for the volume-outcome relationship for common lumbar spine surgical procedures

BACKGROUND CONTEXT

The importance of surgeon volume as a quality measure has been defined for a number of surgical specialties. Meaningful procedural volume benchmarks have not been established, however, particularly with respect to lumbar spine surgery.

PURPOSE

We aimed to establish surgeon volume benchmarks for the performance of four common lumbar spine surgical procedures (discectomy, decompression, lumbar interbody fusion, and lumbar posterolateral fusion).

STUDY DESIGN

A retrospective review of data in the Florida Statewide Inpatient Dataset (2011-2014) was carried out.

PATIENT SAMPLE

Patients who underwent one of the four lumbar spine surgical procedures under study comprised the study sample.

OUTCOME MEASURE

The development of a complication or hospital readmission within 90 days of the surgical procedure was the surgical outcome.

METHODS

For each specific procedure, individual surgeon volume was separately plotted against the number of complications and readmissions in a spline analysis that adjusted for co-variates. Spline cut-points were used to create a categorical variable of procedure volume for each individual procedure. Log-binomial regression analysis was then separately performed using the categorical volume-outcome metric for each individual procedure and for the outcomes of 90-day complications and 90-day readmissions.

RESULTS

In all, 187,185 spine surgical procedures met inclusion criteria, performed by 5,514 different surgeons at 178 hospitals. Spline analysis determined that the procedure volume cut-point was 25 for decompressions, 40 for discectomy, 43 for interbody fusion, and 35 for posterolateral fusions. For surgeons who failed to meet the volume metric, there was a 63% increase in the risk of complications following decompressions, a 56% increase in the risk of complications following discectomy, a 15% increase in the risk of complications following lumbar interbody fusions, and a 47% increase in the risk of complications following posterolateral fusions. Findings were similar for readmission measures.

CONCLUSIONS

The results of this work allow us to identify meaningful volume-based benchmarks for the performance of common lumbar spine surgical procedures including decompression, discectomy, and fusion-based procedures. Based on our determinations, readily achievable goals for individual surgeons would approximate an average of four discectomy and lumbar interbody fusion procedures per month, three posterolateral lumbar fusions per month, and at least one decompression surgery every other week.

Is There Variation in Procedural Utilization for Lumbar Spine Disorders Between a Fee-for-Service and Salaried Healthcare System?

BACKGROUND

Whether compensation for professional services drives the use of those services is an important question that has not been answered in a robust manner. Specifically, there is a growing concern that spine care practitioners may preferentially choose more costly or invasive procedures in a fee-for-service system, irrespective of the underlying lumbar disorder being treated.

QUESTIONS/PURPOSES

(1) Were proportions of interbody fusions higher in the fee-for-service setting as opposed to the salaried Department of Defense setting? (2) Were the odds of interbody fusion increased in a fee-for-service setting after controlling for indications for surgery?

METHODS

Patients surgically treated for lumbar disc herniation, spinal stenosis, and spondylolisthesis (2006-2014) were identified. Patients were divided into two groups based on whether the surgery was performed in the fee-for-service setting (beneficiaries receive care at a civilian facility with expenses covered by TRICARE insurance) or at a Department of Defense facility (direct care). There were 28,344 patients in the entire study, 21,290 treated in fee-for-service and 7054 treated in Department of Defense facilities. Differences in the rates of fusion-based procedures, discectomy, and decompression between both healthcare settings were assessed using multinomial logistic regression to adjust for differences in case-mix and surgical indication.

RESULTS

TRICARE beneficiaries treated for lumbar spinal disorders in the fee-for-service setting had higher odds of receiving interbody fusions (fee-for-service: 7267 of 21,290 [34%], direct care: 1539 of 7054 [22%], odds ratio [OR]: 1.25 [95% confidence interval 1.20-1.30], p < 0.001). Purchased care patients were more likely to receive interbody fusions for a diagnosis of disc herniation (adjusted OR 2.61 [2.36-2.89], p < 0.001) and for spinal stenosis (adjusted OR 1.39 [1.15-1.69], p < 0.001); however, there was no difference for patients with spondylolisthesis (adjusted OR 0.99 [0.84-1.16], p = 0.86).

CONCLUSIONS

The preferential use of interbody fusion procedures was higher in the fee-for-service setting irrespective of the underlying diagnosis. These results speak to the existence of provider inducement within the field of spine surgery. This reality portends poor performance for surgical practices and hospitals in Accountable Care Organizations and bundled payment programs in which provider inducement is allowed to persist.

LEVEL OF EVIDENCE

Level III, economic and decision analysis.

The prevalence of spinal epidural lipomatosis on magnetic resonance imaging

BACKGROUND

Spinal epidural lipomatosis (SEL) refers to an excessive accumulation of fat within the epidural space. It can be idiopathic or secondary, resulting in significant morbidity. The prevalence of SEL, including idiopathic and secondary SEL, and its respective risk factors are poorly defined.

PURPOSE

We sought to: (1) assess the prevalence of SEL among patients who underwent a dedicated magnetic resonance imaging (MRI) scan of the spine-including incidental SEL (ie, SEL without any spine-related symptoms), SEL with spine-related symptoms, and symptomatic SEL (ie, with symptoms specific for SEL); and (2) assess factors associated with overall SEL and subgroups. In addition, we assessed differences between SEL subgroups.

METHODS

We reviewed the records of 28,902 patients, aged 18 years and older with a spine MRI (2004 to 2015) at two tertiary care centers. We identified SEL cases by searching radiology reports for SEL, including synonyms and misspellings. Prevalence numbers were calculated as a percentage of the total number of patients. We used multivariate logistic regression analysis to identify factors associated with overall SEL and subgroups.

RESULTS

The prevalence of overall SEL was 2.5% (731 of 28,902): incidental SEL, 0.6% (168 of 28,902); SEL with symptoms, 1.8% (526 of 28,902); and symptomatic SEL, 0.1% (37 of 28,902). Factors associated with overall SEL in multivariate analysis were the following: older age (odds ratio [OR]: 1.01, 95% confidence interval [CI]: 1.01-1.02, p<.001), higher modified Charlson comorbidity index (OR: 1.10, 95% CI: 1.07-1.13, p<.001), male sex (OR: 2.01, 95% CI: 1.71-2.37, p<.001), BMI>30 (OR: 2.59, 95% CI: 1.97-3.41, p<.001), Black/African American race (OR: 1.66, 95% CI: 1.24-2.23, p=.001), systemic corticosteroid use (OR: 2.59, 95% CI: 1.69-3.99, p<.001), and epidural corticosteroid injections (OR: 3.48, 95% CI: 2.82-4.30, p<.001).

CONCLUSIONS

We found that about 1 in 40 patients undergoing a spine MRI had SEL; 23% of whom with no symptoms, 72% with spine-related symptoms, and 5% with symptoms specific for SEL. Our data help identify patients who might warrant an increased index of suspicion for SEL.