Search Results

You are looking at 1 - 9 of 9 items for

  • Author or Editor: Lola Chambless x
Clear All Modify Search
Restricted access

Nitin Agarwal, Michael D. White, Susan C. Pannullo and Lola B. Chambless

OBJECTIVE

Resident attrition creates a profound burden on trainees and residency programs. This study aims to analyze trends in resident attrition in neurological surgery.

METHODS

This study followed a cohort of 1275 residents who started neurosurgical residency from 2005 to 2010. Data obtained from the American Association of Neurological Surgeons (AANS) included residents who matched in neurosurgery during this time. Residents who did not finish their residency training at the program in which they started were placed into the attrition group. Residents in the attrition group were characterized by one of five outcomes: transferred neurosurgery programs; transferred to a different specialty; left clinical medicine; deceased; or unknown. A thorough internet search was conducted for residents who did not complete their training at their first neurosurgical program. Variables leading to attrition were also analyzed, including age, sex, presence of advanced degree (Ph.D.), postgraduate year (PGY), and geographical region of program.

RESULTS

Residents starting neurosurgical residency from 2005 to 2010 had an overall attrition rate of 10.98%. There was no statistically significant difference in attrition rates among the years (p = 0.337). The outcomes for residents in the attrition group were found to be as follows: 33.61% transferred neurosurgical programs, 56.30% transferred to a different medical specialty, 8.40% left clinical medicine, and 1.68% were deceased. It was observed that women had a higher attrition rate (18.50%) than men (10.35%). Most attrition (65.07%) occurred during PGY 1 or 2. The attrition group was also observed to be significantly older at the beginning of residency training, with a mean of 31.69 years of age compared to 29.31 in the nonattrition group (p < 0.001). No significant difference was observed in the attrition rates for residents with a Ph.D. (9.86%) compared to those without a Ph.D. (p = 0.472).

CONCLUSIONS

A majority of residents in the attrition group pursued training in different medical specialties, most commonly neurology, radiology, and anesthesiology. Factors associated with an increased rate of attrition were older age at the beginning of residency, female sex, and junior resident (PGY-1 to PGY-2). Resident attrition remains a significant problem within neurosurgical training, and future studies should focus on targeted interventions to identify individuals at risk to help them succeed in their medical careers.

Restricted access

Whitney E. Muhlestein, Dallin S. Akagi, Amy R. McManus and Lola B. Chambless

OBJECTIVE

Efficient allocation of resources in the healthcare system enables providers to care for more and needier patients. Identifying drivers of total charges for transsphenoidal surgery (TSS) for pituitary tumors, which are poorly understood, represents an opportunity for neurosurgeons to reduce waste and provide higher-quality care for their patients. In this study the authors used a large, national database to build machine learning (ML) ensembles that directly predict total charges in this patient population. They then interrogated the ensembles to identify variables that predict high charges.

METHODS

The authors created a training data set of 15,487 patients who underwent TSS between 2002 and 2011 and were registered in the National Inpatient Sample. Thirty-two ML algorithms were trained to predict total charges from 71 collected variables, and the most predictive algorithms combined to form an ensemble model. The model was internally and externally validated to demonstrate generalizability. Permutation importance and partial dependence analyses were performed to identify the strongest drivers of total charges. Given the overwhelming influence of length of stay (LOS), a second ensemble excluding LOS as a predictor was built to identify additional drivers of total charges.

RESULTS

An ensemble model comprising 3 gradient boosted tree classifiers best predicted total charges (root mean square logarithmic error = 0.446; 95% CI 0.439–0.453; holdout = 0.455). LOS was by far the strongest predictor of total charges, increasing total predicted charges by approximately $5000 per day.

In the absence of LOS, the strongest predictors of total charges were admission type, hospital region, race, any postoperative complication, and hospital ownership type.

CONCLUSIONS

ML ensembles predict total charges for TSS with good fidelity. The authors identified extended LOS, nonelective admission type, non-Southern hospital region, minority race, postoperative complication, and private investor hospital ownership as drivers of total charges and potential targets for cost-lowering interventions.

Full access

Heather M. Kistka, Arash Nayeri, Li Wang, Jamie Dow, Rameela Chandrasekhar and Lola B. Chambless

OBJECT

Misrepresentation of scholarly achievements is a recognized phenomenon, well documented in numerous fields, yet the accuracy of reporting remains dependent on the honor principle. Therefore, honest self-reporting is of paramount importance to maintain scientific integrity in neurosurgery. The authors had observed a trend toward increasing numbers of publications among applicants for neurosurgery residency at Vanderbilt University and undertook this study to determine whether this change was a result of increased academic productivity, inflated reporting, or both. They also aimed to identify application variables associated with inaccurate citations.

METHODS

The authors retrospectively reviewed the residency applications submitted to their neurosurgery department in 2006 (n = 148) and 2012 (n = 194). The applications from 2006 were made via SF Match and those from 2012 were made using the Electronic Residency Application Service. Publications reported as “accepted” or “in press” were verified via online search of Google Scholar, PubMed, journal websites, and direct journal contact. Works were considered misrepresented if they did not exist, incorrectly listed the applicant as first author, or were incorrectly listed as peer reviewed or published in a printed journal rather than an online only or non-peer-reviewed publication. Demographic data were collected, including applicant sex, medical school ranking and country, advanced degrees, Alpha Omega Alpha membership, and USMLE Step 1 score. Zero-inflated negative binomial regression was used to identify predictors of misrepresentation.

RESULTS

Using univariate analysis, between 2006 and 2012 the percentage of applicants reporting published works increased significantly (47% vs 97%, p < 0.001). However, the percentage of applicants with misrepresentations (33% vs 45%) also increased. In 2012, applicants with a greater total of reported works (p < 0.001) and applicants from unranked US medical schools (those not ranked by US News & World Report) were more likely to have erroneous citations (p = 0.038).

CONCLUSIONS

The incidence of legitimate and misrepresented scholarly works reported by applicants to the authors’ neurosurgery residency program increased during the past 6 years. Misrepresentation is more common in applicants from unranked US medical schools and those with a greater number of reported works on their application. This trend is concerning in a profession where trustworthiness is vital. To preserve integrity in the field, programs should consider verifying citations prior to submitting their rank lists.

Free access

Andrew T. Hale, David P. Stonko, Li Wang, Megan K. Strother and Lola B. Chambless

OBJECTIVE

Prognostication and surgical planning for WHO grade I versus grade II meningioma requires thoughtful decision-making based on radiographic evidence, among other factors. Although conventional statistical models such as logistic regression are useful, machine learning (ML) algorithms are often more predictive, have higher discriminative ability, and can learn from new data. The authors used conventional statistical models and an array of ML algorithms to predict atypical meningioma based on radiologist-interpreted preoperative MRI findings. The goal of this study was to compare the performance of ML algorithms to standard statistical methods when predicting meningioma grade.

METHODS

The cohort included patients aged 18–65 years with WHO grade I (n = 94) and II (n = 34) meningioma in whom preoperative MRI was obtained between 1998 and 2010. A board-certified neuroradiologist, blinded to histological grade, interpreted all MR images for tumor volume, degree of peritumoral edema, presence of necrosis, tumor location, presence of a draining vein, and patient sex. The authors trained and validated several binary classifiers: k-nearest neighbors models, support vector machines, naïve Bayes classifiers, and artificial neural networks as well as logistic regression models to predict tumor grade. The area under the curve–receiver operating characteristic curve was used for comparison across and within model classes. All analyses were performed in MATLAB using a MacBook Pro.

RESULTS

The authors included 6 preoperative imaging and demographic variables: tumor volume, degree of peritumoral edema, presence of necrosis, tumor location, patient sex, and presence of a draining vein to construct the models. The artificial neural networks outperformed all other ML models across the true-positive versus false-positive (receiver operating characteristic) space (area under curve = 0.8895).

CONCLUSIONS

ML algorithms are powerful computational tools that can predict meningioma grade with great accuracy.

Full access

Michael C. Dewan, Gabrielle A. White-Dzuro, Philip R. Brinson, Reid C. Thompson and Lola B. Chambless

OBJECTIVE

Seizures are among the most common perioperative complications in patients undergoing craniotomy for brain tumor resection and have been associated with increased disease progression and decreased survival. Little evidence exists regarding the relationship between postoperative seizures and hospital quality measures, including length of stay (LOS), disposition, and readmission. The authors sought to address these questions by analyzing a glioma population over 15 years.

METHODS

A retrospective cohort study was used to evaluate the outcomes of patients who experienced a postoperative seizure. Patients with glioma who underwent craniotomy for resection between 1998 and 2013 were enrolled in the institutional tumor registry. Basic data, including demographics and comorbidities, were recorded in addition to hospitalization details and complications. Seizures were diagnosed by clinical examination, observation, and electroencephalography. The Student t-test and chi-square test were used to analyze differences in the means between continuous and categorical variables, respectively. Multivariate logistic and linear regression was used to compare multiple clinical variables against hospital quality metrics and survival figures, respectively.

RESULTS

In total, 342 patients with glioma underwent craniotomy for first-time resection. The mean age was 51.0 ± 17.3 years, 192 (56.1%) patients were male, and the median survival time for all grades was 15.4 months (range 6.2–24.0 months). High-grade glioma (Grade III or IV) was seen in 71.9% of patients. Perioperative antiepileptic drugs were administered to 88% of patients. Eighteen (5.3%) patients experienced a seizure within 14 days postoperatively, and 9 (50%) of these patients experienced first-time seizures. The mean time to the first postoperative seizure was 4.3 days (range 0–13 days). There was no significant association between tumor grade and the rate of perioperative seizure (Grade I, 0%; II, 7.0%; III, 6.1%; IV, 5.2%; p = 0.665). A single ictal episode occurred in 11 patients, while 3 patients experienced 2 seizures and 4 patients developed 3 or more seizures. Compared with their seizure-free counterparts, patients who experienced a perioperative seizure had an increased average hospital (6.8 vs 3.6 days, p = 0.032) and ICU LOS (5.4 vs 2.3 days; p < 0.041). Seventy-five percent of seizure-free patients were discharged home in comparison with 55.6% of seizure patients (p = 0.068). Patients with a postoperative seizure were significantly more likely to visit the emergency department within 90 days (44.4% vs 19.0%; OR 3.41 [95% CI 1.29–9.02], p = 0.009) and more likely to be readmitted within 90 days (50.0% vs 18.4%; OR 4.45 [95% CI 1.69–11.70], p = 0.001). In addition, seizure-free patients had a longer median overall survival (15.6 months [interquartile range 6.6–24.4 months] vs 3.0 months [interquartile range 1.0–25.0 months]; p = 0.013).

CONCLUSIONS

Patients with perioperative seizures following glioma resection required longer hospital and ICU LOS, were readmitted at higher rates than seizure-free patients, and experienced shorter overall survival. Biological and clinical factors that predispose to the development of seizures after glioma surgery portend a worse outcome. Efforts to identify these factors and reduce the risk of postoperative seizure should remain a priority among neurosurgical oncologists.

Restricted access

Hilary Highfield Nickols, Lola B. Chambless, Robert P. Carson, Cheryl M. Coffin, Matthew M. Pearson and Ty W. Abel

Intramedullary spinal cord teratomas are rare entities in infants. Management of these lesions is primarily surgical, with outcome dependent on rapid surgical decompression and complete gross-total tumor resection. The lesions are typically of the mature type, with immature teratomas displaying unique pathological features. The authors report a case of an extensive intramedullary immature teratoma in an infant with resolution of quadriplegia following grosstotal radical resection. At the 1-year follow-up, there was radiographic evidence of tumor, and surgical reexploration yielded portions of immature teratoma and extensive gliosis.

Restricted access

Jonathan A. Forbes, Lola B. Chambless, Jason G. Smith, Curtis A. Wushensky, Richard L. Lebow, JoAnn Alvarez and Matthew M. Pearson

Object

The question of whether to obtain routine or selective preoperative imaging of the neuraxis in pediatric patients with cerebellar neoplasms remains a controversial topic. Staging of the neuraxis is generally considered beneficial in patients with neoplasms associated with an elevated risk of leptomeningeal dissemination (LD). When these studies are obtained preoperatively, there is a decrease in the number of false-positive images related to debris in the immediate postoperative period. Additionally, knowledge of the extent of spread has the potential to affect the risk/benefit analysis of aggressive resection. Although the majority of pediatric neurosurgeons surveyed choose to obtain selective preoperative imaging of the neuraxis in cases of cerebellar neoplasms “with findings suggestive of high-grade pathology,” an evidence-based protocol in the literature is lacking. The goal of this study was to assess radiological characteristics of tumors with an elevated risk of LD and identify a method to help guide preoperative imaging of the neuraxis.

Methods

The authors first reviewed the literature to gain an appreciation of the risk of LD of pediatric cerebellar neoplasms based on underlying histopathology and/or grade. Available evidence indicates preoperative imaging of the neuraxis in patients with Grade I tumors to be of questionable utility. In contrast, evidence suggested that preoperative imaging of the neuraxis in patients with Grades II–IV neoplasms was clinically warranted.

The authors then evaluated an extensive base of neuroradiological literature to identify possible MR imaging and/or CT findings with the potential to differentiate Grade I from higher-grade neoplasms in pediatric patients. They analyzed the preoperative radiological findings in 50 pediatric patients who had undergone craniotomy for resection of cerebellar neoplasms at Vanderbilt Children's Hospital since 2003 with reference to 7 chosen radiological criteria. Logistic regression models were fit using radiological features to determine the best predictors of Grades II–IV tumors. Receiver operating characteristic methods were used to identify diagnostic properties of the best predictors.

Results

The relative T2 signal intensity (RT2SI), an indirect measure of the water content of the solid component of the tumor, was best able to identify neoplasms with an elevated risk of LD. An RT2SI value of 0.71 was selected by the authors as the best operating point on the curve. Of the 31 neoplasms retrospectively designated as hypointense T2-weighted lesions (RT2SI ≤ 0.71), 30 (97%) were Grade II or higher. All medulloblastomas, ependymomas, and high-grade (Grades III and IV) neoplasms were hypointense T2-weighted lesions. Of the 19 T2-weighted hyperintense neoplasms (RT2SI > 0.71), 16 (84%) were Grade I and 3 were Grade II.

Conclusions

Measurement of the RT2SI can help predict Grade II–IV tumors at an elevated risk of leptomeningeal spread and guide staging of the neuraxis. Pediatric patients with cerebellar neoplasms found to have an RT2SI of less than or equal to 0.71 are recommended for neuraxis imaging prior to surgery.