NSQIP as a Predictor of Length of Stay in Patients Undergoing Free Flap Reconstruction

Predictive Models for Length of Stay and Discharge Disposition in Elective Spine Surgery: Development, Validation, and Comparison to the ACS NSQIP Risk Calculator

STUDY DESIGN

A retrospective study at a single academic institution.

OBJECTIVE

The purpose of this study is to utilize machine learning to predict hospital length of stay (LOS) and discharge disposition following adult elective spine surgery, and to compare performance metrics of machine learning models to the American College of Surgeon's National Surgical Quality Improvement Program's (ACS NSQIP) prediction calculator.

SUMMARY OF BACKGROUND DATA

A total of 3678 adult patients undergoing elective spine surgery between 2014 and 2019, acquired from the electronic health record.

METHODS

Patients were divided into three stratified cohorts: cervical degenerative, lumbar degenerative, and adult spinal deformity groups. Predictive variables included demographics, body mass index, surgical region, surgical invasiveness, surgical approach, and comorbidities. Regression, classification trees, and least absolute shrinkage and selection operator (LASSO) were used to build predictive models. Validation of the models was conducted on 16% of patients (N=587), using area under the receiver operator curve (AUROC), sensitivity, specificity, and correlation. Patient data were manually entered into the ACS NSQIP online risk calculator to compare performance. Outcome variables were discharge disposition (home vs. rehabilitation) and LOS (days).

RESULTS

Of 3678 patients analyzed, 51.4% were male (n=1890) and 48.6% were female (n=1788). The average LOS was 3.66 days. In all, 78% were discharged home and 22% discharged to rehabilitation. Compared with NSQIP (Pearson R2 =0.16), the predictions of poisson regression ( R2 =0.29) and LASSO ( R2 =0.29) models were significantly more correlated with observed LOS ( P =0.025 and 0.004, respectively). Of the models generated to predict discharge location, logistic regression yielded an AUROC of 0.79, which was statistically equivalent to the AUROC of 0.75 for NSQIP ( P =0.135).

CONCLUSION

The predictive models developed in this study can enable accurate preoperative estimation of LOS and risk of rehabilitation discharge for adult patients undergoing elective spine surgery. The demonstrated models exhibited better performance than NSQIP for prediction of LOS and equivalent performance to NSQIP for prediction of discharge location.

Evaluation of a novel database for quality assurance at a head and neck service in New Zealand: an audit of free flap head and neck reconstruction

BACKGROUND

Clinical audit is a critical quality improvement exercise, yet efficient audit tools are lacking. The main objective of this study was to evaluate a recently deployed database in facilitating the process of clinical audit, and the secondary objective was to evaluate the outcomes of free flap reconstruction of the head and neck at our centre.

METHODS

A head and neck cancer-specific database was customized to suit the needs of our head and neck multidisciplinary team. Data has been entered prospectively into this database since March of 2018. An audit of free flap reconstruction of the head and neck over a 12-month period was performed using the database and analysed as a case study to examine its efficacy as a clinical audit tool. Additionally, the outcomes of free flap reconstruction at our centre were compared to those reported in the international literature.

RESULTS

The database allows flexible and specific queries, analysis and export of data, and can provide immediate results. However, issues with data quality and completeness were identified. In this audit, the overall 30-day complication rate and 30-day mortality in patients undergoing free flap reconstruction of the head and neck were 58% and 3%, respectively.

CONCLUSION

The database is fit for its intended purpose as an audit tool. Outcomes of free flap reconstruction of the head and neck at our centre are comparable to those of institutions overseas.

Early discharge after free-tissue transfer does not increase adverse events

INTRODUCTION

Demonstrate that carefully selected free flap patients may be discharged early after surgery without increasing the rates of postoperative complications or readmissions.

METHODS

Based on a published article in Laryngoscope 2016 of 51 free-tissue transfers, a retrospective chart review was performed on an expanded cohort who underwent free-tissue transfer for head and neck reconstruction between February 2010 and May 2018 and discharged by postoperative day 3.

RESULTS

101 patients who underwent 104 free flaps with average age of 56 (3-84) years old were reviewed. Free flap indications included orbital and maxillary defects (n = 22), palatal defects (n = 16), nasal and septal defects (n = 16), cranioplasty and scalp defects (n = 16), mandibular defects due to osteoradionecrosis (n = 14), facial contouring and parotid defects (n = 12), and complex postsurgical and radiotherapy wounds or fistula closure (n = 8). Free flaps performed were anterolateral thigh (n = 97), radial forearm (n = 2), serratus (n = 2), latissimus (n = 1), fibula (n = 1) and supraclavicular (n = 1). The recipient vessels used via minimal access approaches were facial (n = 43), superficial temporal (n = 29), angular (n = 20) and others. There were 3 flap failures (2.9%) recognized in follow-up. No flap failures or perioperative complications were associated with early discharge. There were only 2 patients readmitted and 1 watched in observation within 30 days postoperatively.

CONCLUSION

An updated review of our institutional experience with more than double the cohort size substantiates previous conclusions that early discharge after free-tissue transfer is a safe option in select patients. Moreover, earlier discharge is a critical management choice that reduces cost and decreases hospital-related adverse events.

Predictive value of the ACS NSQIP calculator for head and neck reconstruction free tissue transfer

BACKGROUND

Predictive models to forecast the likelihood of specific outcomes after surgical intervention allow informed shared decision-making by surgeons and patients. Previous studies have suggested that existing general surgical risk calculators poorly forecast head and neck surgical outcomes. However, no large study has addressed this question while subdividing subjects by surgery performed.

OBJECTIVES

To determine the accuracy of the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) Surgical Risk Calculator in estimating length of hospital stay and risk of postoperative complications after free tissue transfer surgery.

STUDY DESIGN

A retrospective chart review of patients at one institution was performed using Current Procedural Terminology codes for anterolateral thigh (ALT) flap, fibula free flap (FFF), and radial forearm free flap (RFFF) reconstruction. Output data from the ACS NSQIP surgical risk calculator were compared with the observed rates in our patients.

METHODS

Incidences of cardiac complications, pneumonia, venous thromboembolism, return to the operating room, and discharge to skilled nursing facility (SNF) were compared to predicted incidences. Length of stay was also compared to the predicted length of stay.

RESULTS

Three hundred thirty-six free flap reconstructions with 197 ALT flaps, 85 RFFFs, and 54 FFFFs were included. Brier scores were calculated using ACS NSQIP forecast and actual incidences. No Brier score was <0.01 for the entire sample or any subgroup, which indicates that the NSQIP risk calculator does not accurately forecast outcomes after free tissue reconstruction.

CONCLUSION

The ACS NSQIP failed to accurately forecast postoperative outcomes after head and neck free flap reconstruction for the entire sample or subgroup analyses.

LEVEL OF EVIDENCE

4 Laryngoscope, 130:679-684, 2020.

Impact of the insurance type of head and neck cancer patients on their hospitalization utilization patterns

BACKGROUND

Head and neck cancer (HNC) patients with Medicaid, Medicare, or no insurance show poor outcomes in comparison with privately insured patients. It was hypothesized that nonprivate insurance coverage biases the selection of the treatment site to favor hospitals that are not associated with optimum treatment outcomes. This study assessed the relation between the insurance type of HNC patients and the hospital type for inpatient care.

METHODS

Adult HNC patients were identified from the Nationwide Inpatient Sample (2012 and 2013). The primary exposure was the insurance provider type. The outcome was the hospital type, which was classified by the hospital's ownership and its location and teaching status. Multivariate multinomial logistic regression models were constructed to control for the patient's age, sex, race, income, mortality risk, and geographic location. The analysis was weighted and was adjusted for multiple comparisons.

RESULTS

In all, 37,466 HNC patients representing 187,330 patients nationally were identified. After adjustments for age, sex, race, income, and mortality risk, in comparison with privately insured patients, Medicaid, Medicare, and uninsured patients demonstrated 1.14 to 2.29 increased odds of undergoing treatment at rural, urban nonteaching, private investor-owned, or government (nonfederal) hospitals (P < .05). This trend remained apparent even after adjustments for the geographic location.

CONCLUSIONS

Uninsured patients or patients insured by government programs predominantly underwent care for HNC at hospital types most often associated with inferior survival outcomes. This finding could explain some proportion of insurance-related disparities in HNC outcomes. Further studies are warranted to determine whether interventions to promote equitable access to optimal hospital settings for patients, regardless of their insurance type, might improve outcomes among nonprivate insurance holders. Cancer 2018;124:760-8. © 2017 American Cancer Society.