Search Results

You are looking at 1 - 10 of 40 items for

  • Author or Editor: Robert A. Hart x
  • All content x
Clear All Modify Search
Full access

Alan H. Daniels, Satoshi Kawaguchi, Alec G. Contag, Farbod Rastegar, Garrett Waagmeester, Paul A. Anderson, Melanie Arthur, and Robert A. Hart

OBJECTIVE

Beginning in 2008, the Centers for Medicare and Medicaid Service (CMS) determined that certain hospital-acquired adverse events such as surgical site infection (SSI) following spine surgery should never occur. The following year, they expanded the ruling to include deep vein thrombosis (DVT) and pulmonary embolism (PE) following total joint arthroplasty. Due to their ruling that “never events” are not the payers' responsibility, CMS insists that the costs of managing these complications be borne by hospitals and health care providers, rather than billings to health care payers for additional care required in their management. Data comparing the expected costs of such adverse events in patients undergoing spine and orthopedic surgery have not previously been reported.

METHODS

The California State Inpatient Database (CA-SID) from 2008 to 2009 was used for the analysis. All patients with primary procedure codes indicating anterior cervical discectomy and fusion (ACDF), posterior lumbar interbody fusion (PLIF), lumbar laminectomy (LL), total knee replacement (TKR), and total hip replacement (THR) were analyzed. Patients with diagnostic and/or treatment codes for DVT, PE, and SSI were separated from patients without these complication codes. Patients with more than 1 primary procedure code or more than 1 complication code were excluded. Median charges for treatment from primary surgery through 3 months postoperatively were calculated.

RESULTS

The incidence of the examined adverse events was lowest for ACDF (0.6% DVT, 0.1% PE, and 0.03% SSI) and highest for TKA (1.3% DVT, 0.3% PE, 0.6% SSI). Median inpatient charges for uncomplicated LL was $51,817, compared with $73,432 for ACDF, $143,601 for PLIF, $74,459 for THR, and $70,116 for TKR. Charges for patients with DVT ranged from $108,387 for TKR (1.5 times greater than index) to $313,536 for ACDF (4.3 times greater than index). Charges for patients with PE ranged from $127,958 for TKR (1.8 times greater than index) to $246,637 for PLIF (1.7 times greater than index). Charges for patients with SSI ranged from $168,964 for TKR (2.4 times greater than index) to $385,753 for PLIF (2.7 times greater than index).

CONCLUSIONS

Although incidence rates are low, adverse events of spinal procedures substantially increase the cost of care. Charges for patients experiencing DVT, PE, and SSI increased in this study by factors ranging from 1.8 to 4.3 times those for patients without such complications across 5 common spinal and orthopedic procedures. Cost projections by health care providers will need to incorporate expected costs of added care for patients experiencing such complications, assuming that the cost burden of such events continues to shift from payers to providers.

Restricted access

Frank J. Schwab, Ashish Patel, Christopher I. Shaffrey, Justin S. Smith, Jean-Pierre Farcy, Oheneba Boachie-Adjei, Richard A. Hostin, Robert A. Hart, Behrooz A. Akbarnia, Douglas C. Burton, Shay Bess, and Virginie Lafage

Object

Pedicle subtraction osteotomy (PSO) is a surgical procedure that is frequently performed on patients with sagittal spinopelvic malalignment. Although it allows for substantial spinopelvic realignment, suboptimal realignment outcomes have been reported in up to 33% of patients. The authors' objective in the present study was to identify differences in radiographic profiles and surgical procedures between patients achieving successful versus failed spinopelvic realignment following PSO.

Methods

This study is a multicenter retrospective consecutive PSO case series. The authors evaluated 99 cases involving patients who underwent PSO for sagittal spinopelvic malalignment. Because precise cutoffs of acceptable residual postoperative sagittal vertical axis (SVA) values have not been well defined, comparisons were focused between patient groups with a postoperative SVA that could be clearly considered either a success or a failure. Only cases in which the patients had a postoperative SVA of less than 50 mm (successful PSO realignment) or more than 100 mm (failed PSO realignment) were included in the analysis. Radiographic measures and PSO parameters were compared between successful and failed PSO realignments.

Results

Seventy-nine patients met the inclusion criteria. Successful realignment was achieved in 61 patients (77%), while realignment failed in 18 (23%). Patients with failed realignment had larger preoperative SVA (mean 217.9 vs 106.7 mm, p < 0.01), larger pelvic tilt (mean 36.9° vs 30.7°, p < 0.01), larger pelvic incidence (mean 64.2° vs 53.7°, p < 0.01), and greater lumbar lordosis–pelvic incidence mismatch (−47.1° vs −30.9°, p < 0.01) compared with those in whom realignment was successful. Failed and successful realignments were similar regarding the vertebral level of the PSO, the median size of wedge resection 22.0° (interquartile range 16.5°−28.5°), and the numerical changes in pre- and postoperative spinopelvic parameters (p > 0.05).

Conclusions

Patients with failed PSO realignments had significantly larger preoperative spinopelvic deformity than patients in whom realignment was successful. Despite their apparent need for greater correction, the patients in the failed realignment group only received the same amount of correction as those in the successfully realigned patients. A single-level standard PSO may not achieve optimal outcome in patients with high preoperative spinopelvic sagittal malalignment. Patients with large spinopelvic deformities should receive larger osteotomies or additional corrective procedures beyond PSOs to avoid undercorrection.

Restricted access

Virginie Lafage, Neil J. Bharucha, Frank Schwab, Robert A. Hart, Douglas Burton, Oheneba Boachie-Adjei, Justin S. Smith, Richard Hostin, Christopher Shaffrey, Munish Gupta, Behrooz A. Akbarnia, and Shay Bess

Object

Sagittal spinopelvic imbalance is a major contributor to pain and disability for patients with adult spinal deformity (ASD). Preoperative planning is essential for pedicle subtraction osteotomy (PSO) candidates; however, current methods are often inaccurate because no formula to date predicts both postoperative sagittal balance and pelvic alignment. The authors of this study aimed to evaluate the accuracy of 2 novel formulas in predicting postoperative spinopelvic alignment after PSO.

Methods

This study is a multicenter retrospective consecutive PSO case series. Adults with spinal deformity (> 21 years old) who were treated with a single-level lumbar PSO for sagittal imbalance were evaluated. All patients underwent preoperative and a minimum of 6-month postoperative radiography. Two novel formulas were used to predict the postoperative spinopelvic alignment. The results predicted by the formulas were then compared with the actual postoperative radiographic values, and the formulas' ability to identify successful (sagittal vertical axis [SVA] ≤ 50 mm and pelvic tilt [PT] ≤ 25°) and unsuccessful (SVA > 50 mm or PT > 25°) outcomes was evaluated.

Results

Ninety-nine patients met inclusion criteria. The median absolute error between the predicted and actual PT was 4.1° (interquartile range 2.0°–6.4°). The median absolute error between the predicted and actual SVA was 27 mm (interquartile range 11–47 mm). Forty-one of 54 patients with a formula that predicted a successful outcome had a successful outcome as shown by radiography (positive predictive value = 0.76). Forty-four of 45 patients with a formula that predicted an unsuccessful outcome had an unsuccessful outcome as shown by radiography (negative predictive value = 0.98).

Conclusions

The spinopelvic alignment formulas were accurate when predicting unsuccessful outcomes but less reliable when predicting successful outcomes. The preoperative surgical plan should be altered if an unsuccessful result is predicted. However, even after obtaining a predicted successful outcome, surgeons should ensure that the predicted values are not too close to unsuccessful values and should identify other variables that may affect alignment. In the near future, it is anticipated that the use of these formulas will lead to better surgical planning and improved outcomes for patients with complex ASD.

Restricted access

Robert A. Hart, Christopher M. Domes, Brady Goodwin, Charles R. D'Amato, Jung U. Yoo, Ronald J. Turker, and Matthew F. Halsey

Object

The ideal surgical management of high-grade spondylolisthesis remains unclear. Concerns regarding the original Bohlman transsacral interbody fusion technique with stand-alone autologous fibular strut include late graft fracture and incomplete reduction of lumbosacral kyphosis. The authors' goal was to evaluate the radiographic and surgical outcomes of patients treated for high-grade spondylolisthesis with either transsacral S-1 screws or standard pedicle screw fixation augmenting the Bohlman posterior transsacral interbody fusion technique.

Methods

A retrospective review of patients who underwent fusion for high-grade spondylolisthesis in which a Bohlman oblique posterior interbody fusion augmented with either transsacral or standard pedicle screw fixation was performed by 4 spine surgeons was completed. Estimated blood loss, operating time, perioperative complications, and need for revision surgery were evaluated. Upright pre- and postsurgical lumbar spine radiographs were compared for slip percent and slip angle.

Results

Sixteen patients (12 female and 4 male) with an average age of 29 years (range 9–66 years) were evaluated. The average clinical follow-up was 78 months (range 5–137 months) and the average radiographic follow-up was 48 months (range 5–108 months). Ten L4–S1 and 6 L5–S1 fusions were performed. Five fibular struts and 11 titanium mesh cages were used for interbody fusion. Six patients had isolated transsacral screws placed, with 2 (33%) of the 6 requiring revision surgery for nonunion. No nonunions were observed in patients undergoing spanning pedicle screw fixation augmenting the interbody graft. Six patients experienced perioperative complications including 3 iliac crest site infections, 1 L-5 radiculopathy without motor involvement, 1 deep vein thrombosis, and 1 epidural hematoma requiring irrigation and debridement. The average estimated blood loss and operating times were 763 ml and 360 minutes, respectively. Slip percent improved from an average of 62% to 37% (n = 16; p < 0.01) and slip angle improved from an average of 18° to 8° (n = 16; p < 0.01). No patient experienced L-5 or other motor deficit postoperatively.

Conclusions

The modified Bohlman technique for treatment of high-grade spondylolisthesis has reproducible outcomes among multiple surgeons and results in significant improvements in slip percent and slip angle. Fusion rates were high (14 of 16; 88%), especially with spanning instrumentation augmenting the oblique interbody fusion. Rates of L-5 motor deficit were low in comparison with techniques involving reduction of the anterolisthesis.

Restricted access

Dana L. Cruz, Ethan W. Ayres, Matthew A. Spiegel, Louis M. Day, Robert A. Hart, Christopher P. Ames, Douglas C. Burton, Justin S. Smith, Christopher I. Shaffrey, Frank J. Schwab, Thomas J. Errico, Shay Bess, Virginie Lafage, and Themistocles S. Protopsaltis

OBJECTIVE

Neck and back pain are highly prevalent conditions that account for major disability. The Neck Disability Index (NDI) and Oswestry Disability Index (ODI) are the two most common functional status measures for neck and back pain. However, no single instrument exists to evaluate patients with concurrent neck and back pain. The recently developed Total Disability Index (TDI) combines overlapping elements from the ODI and NDI with the unique items from each. This study aimed to prospectively validate the TDI in patients with spinal deformity, back pain, and/or neck pain.

METHODS

This study is a retrospective review of prospectively collected data from a single center. The 14-item TDI, derived from ODI and NDI domains, was administered to consecutive patients presenting to a spine practice. Patients were assessed using the ODI, NDI, and EQ-5D. Validation of internal consistency, test-retest reproducibility, and validity of reconstructed NDI and ODI scores derived from TDI were assessed.

RESULTS

A total of 252 patients (mean age 55 years, 56% female) completed initial assessments (back pain, n = 115; neck pain, n = 52; back and neck pain, n = 55; spinal deformity, n = 55; and no pain/deformity, n = 29). Of these patients, 155 completed retests within 14 days. Patients represented a wide range of disability (mean ODI score: 36.3 ± 21.6; NDI score: 30.8 ± 21.8; and TDI score: 34.1 ± 20.0). TDI demonstrated excellent internal consistency (Cronbach’s alpha = 0.922) and test-retest reliability (intraclass correlation coefficient = 0.96). Differences between actual and reconstructed scores were not clinically significant. Subanalyses demonstrated TDI’s ability to quantify the degree of disability due to back or neck pain in patients complaining of pain in both regions.

CONCLUSIONS

The TDI is a valid and reliable disability measure in patients with back and/or neck pain and can capture each spine region’s contribution to total disability. The TDI could be a valuable method for total spine assessment in a clinical setting, and its completion is less time consuming than that for both the ODI and NDI.

Free access

Justin K. Scheer, Jessica A. Tang, Justin S. Smith, Eric Klineberg, Robert A. Hart, Gregory M. Mundis Jr., Douglas C. Burton, Richard Hostin, Michael F. O'Brien, Shay Bess, Khaled M. Kebaish, Vedat Deviren, Virginie Lafage, Frank Schwab, Christopher I. Shaffrey, Christopher P. Ames, and the International Spine Study Group

Object

Complications and reoperation for surgery to correct adult spinal deformity are not infrequent, and many studies have analyzed the rates and factors that influence the likelihood of reoperation. However, there is a need for more comprehensive analyses of reoperation in adult spinal deformity surgery from a global standpoint, particularly focusing on the 1st year following operation and considering radiographic parameters and the effects of reoperation on health-related quality of life (HRQOL). This study attempts to determine the prevalence of reoperation following surgery for adult spinal deformity, assess the indications for these reoperations, evaluate for a relation between specific radiographic parameters and the need for reoperation, and determine the potential impact of reoperation on HRQOL measures.

Methods

A retrospective review was conducted of a prospective, multicenter, adult spinal deformity database collected through the International Spine Study Group. Data collected included age, body mass index, sex, date of surgery, information regarding complications, reoperation dates, length of stay, and operation time. The radiographic parameters assessed were total number of levels instrumented, total number of interbody fusions, C-7 sagittal vertical axis, uppermost instrumented vertebra (UIV) location, and presence of 3-column osteotomies. The HRQOL assessment included Oswestry Disability Index (ODI), 36-Item Short Form Health Survey physical component and mental component summary, and SRS-22 scores. Smoking history, Charlson Comorbidity Index scores, and American Society of Anesthesiologists Physical Status classification grades were also collected and assessed for correlation with risk of early reoperation. Various statistical tests were performed for evaluation of specific factors listed above, and the level of significance was set at p < 0.05.

Results

Fifty-nine (17%) of a total of 352 patients required reoperation. Forty-four (12.5%) of the reoperations occurred within 1 year after the initial surgery, including 17 reoperations (5%) within 30 days.

Two hundred sixty-eight patients had a minimum of 1 year of follow-up. Fifty-three (20%) of these patients had a 3-column osteotomy, and 10 (19%) of these 53 required reoperation within 1 year of the initial procedure. However, 3-column osteotomy was not predictive of reoperation within 1 year, p = 0.5476). There were no significant differences between groups with regard to the distribution of UIV, and UIV did not have a significant effect on reoperation rates. Patients needing reoperation within 1 year had worse ODI and SRS-22 scores measured at 1-year follow-up than patients not requiring operation.

Conclusions

Analysis of data from a large multicenter adult spinal deformity database shows an overall 17% reoperation rate, with a 19% reoperation rate for patients treated with 3-column osteotomy and a 16% reoperation rate for patients not treated with 3-column osteotomy. The most common indications for reoperation included instrumentation complications and radiographic failure. Reoperation significantly affected HRQOL outcomes at 1-year follow-up. The need for reoperation may be minimized by carefully considering spinal alignment, termination of fixation, and type of surgical procedure (presence of osteotomy). Precautions should be taken to avoid malposition or instrumentation (rod) failure.

Full access

Themistocles S. Protopsaltis, Justin K. Scheer, Jamie S. Terran, Justin S. Smith, D. Kojo Hamilton, Han Jo Kim, Greg M. Mundis Jr., Robert A. Hart, Ian M. McCarthy, Eric Klineberg, Virginie Lafage, Shay Bess, Frank Schwab, Christopher I. Shaffrey, Christopher P. Ames, and International Spine Study Group

OBJECT

Regional cervical sagittal alignment (C2–7 sagittal vertical axis [SVA]) has been shown to correlate with health-related quality of life (HRQOL). The study objective was to examine the relationship between cervical and thoracolumbar alignment parameters with HRQOL among patients with operative and nonoperative adult thoracolumbar deformity.

METHODS

This is a multicenter prospective data collection of consecutive patients with adult thoracolumbar spinal deformity. Clinical measures of disability included the Oswestry Disability Index (ODI), Scoliosis Research Society-22 Patient Questionnaire (SRS-22), and 36-Item Short-Form Health Survey (SF-36). Cervical radiographic parameters were correlated with global sagittal parameters within the nonoperative and operative cohorts. A partial correlation analysis was performed controlling for C-7 SVA. The operative group was subanalyzed by the magnitude of global deformity (C-7 SVA ≥ 5 cm vs < 5 cm).

RESULTS

A total of 318 patients were included (186 operative and 132 nonoperative). The mean age was 55.4 ± 14.9 years. Operative patients had significantly worse baseline HRQOL and significantly larger C-7 SVA, pelvic tilt (PT), mismatch between pelvic incidence and lumbar lordosis (PI-LL), and C2-7 SVA. The operative patients with baseline C-7 SVA ≥ 5 cm had significantly larger C2-7 lordosis (CL), C2-7 SVA, C-7 SVA, PI-LL, and PT than patients with a normal C-7 SVA. For all patients, baseline C2-7 SVA and CL significantly correlated with baseline ODI, Physical Component Summary (PCS), SRS Activity domain, and SRS Appearance domain. Baseline C2-7 SVA also correlated with SRS Pain and SRS Total. For the operative patients with baseline C-7 SVA ≥ 5 cm, the 2-year C2-7 SVA significantly correlated with 2-year Mental Component Summary, SRS Mental, SRS Satisfaction, and decreases in ODI. Decreases in C2-7 SVA at 2 years significantly correlated with lower ODI at 2 years. Using partial correlations while controlling for C-7 SVA, the C2-7 SVA correlated significantly with baseline ODI (r = 0.211, p = 0.002), PCS (r = −0.178, p = 0.009), and SRS Activity (r = −0.145, p = 0.034) for the entire cohort. In the subset of operative patients with larger thoracolumbar deformities, the change in C2-7 SVA correlated with change in ODI (r = −0.311, p = 0.03).

CONCLUSIONS

Changes in cervical lordosis correlate to HRQOL improvements in thoracolumbar deformity patients at 2-year follow-up. Regional cervical sagittal parameters such as CL and C2–7 SVA are correlated with clinical measures of regional disability and health status in patients with adult thoracolumbar scoliosis. This effect may be direct or a reciprocal effect of the underlying global deformities on regional cervical alignment. However, the partial correlation analysis, controlling for the magnitude of the thoracolumbar deformity, suggests that there is a direct effect of cervical alignment on health measures. Improvements in regional cervical alignment postoperatively correlated positively with improved HRQOL.

Full access

Joshua Bakhsheshian, Justin K. Scheer, Jeffrey L. Gum, Richard Hostin, Virginie Lafage, Shay Bess, Themistocles S. Protopsaltis, Douglas C. Burton, Malla Kate Keefe, Robert A. Hart, Gregory M. Mundis Jr., Christopher I. Shaffrey, Frank Schwab, Justin S. Smith, Christopher P. Ames, and The International Spine Study Group

OBJECTIVE

Mental disease burden can have a significant impact on levels of disability and health-related quality of life (HRQOL) measures. Therefore, the authors investigated the significance of mental health status in adults with spinal deformity and poor physical function.

METHODS

A retrospective analysis of a prospective multicenter database of 365 adult spinal deformity (ASD) patients who had undergone surgical treatment was performed. Health-related QOL variables were examined preoperatively and at the 2-year postoperative follow-up. Patients were grouped by their 36-Item Short Form Health Survey mental component summary (MCS) and physical component summary (PCS) scores. Both groups had PCS scores ≤ 25th percentile for matched norms; however, the low mental health (LMH) group consisted of patients with an MCS score ≤ 25th percentile, and the high mental health (HMH) group included patients with an MCS score ≥ 75th percentile.

RESULTS

Of the 264 patients (72.3%) with a 2-year follow-up, 104 (28.5%) met the inclusion criteria for LMH and 40 patients (11.0%) met those for HMH. The LMH group had a significantly higher overall rate of comorbidities, specifically leg weakness, depression, hypertension, and self-reported neurological and psychiatric disease processes, and were more likely to be unemployed as compared with the HMH group (p < 0.05 for all). The 2 groups had similar 2-year postoperative improvements in HRQOL (p > 0.05) except for the greater improvements in the MCS and the Scoliosis Research Society-22r questionnaire (SRS-22r) mental domain (p < 0.05) in the LMH group and greater improvements in PCS and SRS-22r satisfaction and back pain domains (p < 0.05) in the HMH group. The LMH group had a higher rate of reaching a minimal clinically important difference (MCID) on the SRS-22r mental domain (p < 0.01), and the HMH group had a higher rate of reaching an MCID on the PCS and SRS-22r activity domain (p < 0.05). On multivariable logistic regression, having LMH was a significant independent predictor of failure to reach an MCID on the PCS (p < 0.05). At the 2-year postoperative follow-up, 14 LMH patients (15.1%) were categorized as HMH. Two LMH patients (2.2%), and 3 HMH patients (7.7%) transitioned to a PCS score ≥ 75th percentile for age- and sex-matched US norms (p < 0.01).

CONCLUSIONS

While patients with poor mental and physical health, according to their MCS and PCS scores, have higher medical comorbidity and unemployment rates, they still demonstrate significant improvements in HRQOL measurements postoperatively. Both LMH and HMH patient groups demonstrated similar improvements in most HRQOL domains, except that the LMH patients had difficulties in obtaining improvements in the PCS domain.

Full access

Alex Soroceanu, Douglas C. Burton, Bassel Georges Diebo, Justin S. Smith, Richard Hostin, Christopher I. Shaffrey, Oheneba Boachie-Adjei, Gregory M. Mundis Jr., Christopher Ames, Thomas J. Errico, Shay Bess, Munish C. Gupta, Robert A. Hart, Frank J. Schwab, Virginie Lafage, and International Spine Study Group

OBJECT

Adult spinal deformity (ASD) surgery is known for its high complication rate. This study examined the impact of obesity on complication rates, infection, and patient-reported outcomes in patients undergoing surgery for ASD.

METHODS

This study was a retrospective review of a multicenter prospective database of patients with ASD who were treated surgically. Patients with available 2-year follow-up data were included. Obesity was defined as having a body mass index (BMI) ≥ 30 kg/m2. Data collected included complications (total, minor, major, implant-related, radiographic, infection, revision surgery, and neurological injury), estimated blood loss (EBL), operating room (OR) time, length of stay (LOS), and patient-reported questionnaires (Oswestry Disability Index [ODI], Short Form-36 [SF-36], and Scoliosis Research Society [SRS]) at baseline and at 6 weeks, 1 year, and 2 years postoperatively. The impact of obesity was studied using multivariate modeling, accounting for confounders.

RESULTS

Of 241 patients who satisfied inclusion criteria, 175 patients were nonobese and 66 were obese. Regression models showed that obese patients had a higher overall incidence of major complications (IRR 1.54, p = 0.02) and wound infections (odds ratio 4.88, p = 0.02). Obesity did not increase the number of minor complications (p = 0.62), radiographic complications (p = 0.62), neurological complications (p = 0.861), or need for revision surgery (p = 0.846). Obesity was not significantly correlated with OR time (p = 0.23), LOS (p = 0.9), or EBL (p = 0.98). Both groups experienced significant improvement overtime, as measured on the ODI (p = 0.0001), SF-36 (p = 0.0001), and SRS (p = 0.0001) questionnaires. However, the overall magnitude of improvement was less for obese patients (ODI, p = 0.0035; SF-36, p = 0.0012; SRS, p = 0.022). Obese patients also had a lower rate of improvement over time (SRS, p = 0.0085; ODI, p = 0.0001; SF-36, p = 0.0001).

CONCLUSIONS

This study revealed that obese patients have an increased risk of complications following ASD correction. Despite these increased complications, obese patients do benefit from surgical intervention; however, their improvement in health-related quality of life (HRQL) is less than that of nonobese patients.

Free access

Taemin Oh, Justin K. Scheer, Justin S. Smith, Richard Hostin, Chessie Robinson, Jeffrey L. Gum, Frank Schwab, Robert A. Hart, Virginie Lafage, Douglas C. Burton, Shay Bess, Themistocles Protopsaltis, Eric O. Klineberg, Christopher I. Shaffrey, Christopher P. Ames, and the International Spine Study Group

OBJECTIVE

Patients with adult spinal deformity (ASD) experience significant quality of life improvements after surgery. Treatment, however, is expensive and complication rates are high. Predictive analytics has the potential to use many variables to make accurate predictions in large data sets. A validated minimum clinically important difference (MCID) model has the potential to assist in patient selection, thereby improving outcomes and, potentially, cost-effectiveness.

METHODS

The present study was a retrospective analysis of a multiinstitutional database of patients with ASD. Inclusion criteria were as follows: age ≥ 18 years, radiographic evidence of ASD, 2-year follow-up, and preoperative Oswestry Disability Index (ODI) > 15. Forty-six variables were used for model training: demographic data, radiographic parameters, surgical variables, and results on the health-related quality of life questionnaire. Patients were grouped as reaching a 2-year ODI MCID (+MCID) or not (−MCID). An ensemble of 5 different bootstrapped decision trees was constructed using the C5.0 algorithm. Internal validation was performed via 70:30 data split for training/testing. Model accuracy and area under the curve (AUC) were calculated. The mean quality-adjusted life years (QALYs) and QALYs gained at 2 years were calculated and discounted at 3.5% per year. The QALYs were compared between patients in the +MCID and –MCID groups.

RESULTS

A total of 234 patients met inclusion criteria (+MCID 129, −MCID 105). Sixty-nine patients (29.5%) were included for model testing. Predicted versus actual results were 50 versus 40 for +MCID and 19 versus 29 for −MCID (i.e., 10 patients were misclassified). Model accuracy was 85.5%, with 0.96 AUC. Predicted results showed that patients in the +MCID group had significantly greater 2-year mean QALYs (p = 0.0057) and QALYs gained (p = 0.0002).

CONCLUSIONS

A successful model with 85.5% accuracy and 0.96 AUC was constructed to predict which patients would reach ODI MCID. The patients in the +MCID group had significantly higher mean 2-year QALYs and QALYs gained. This study provides proof of concept for using predictive modeling techniques to optimize patient selection in complex spine surgery.