Matthew C. Davis, Elizabeth N. Kuhn, Bonita S. Agee, Robert A. Oster and James M. Markert
Many neurosurgical training programs have moved from a 24-hour resident call system to a night float system, but the impact on outcomes is unclear. Here, the authors compare length of stay (LOS) for neurosurgical patients admitted before and after initiation of a night float system at a tertiary care training hospital.
The neurosurgical residency at the University of Alabama at Birmingham transitioned from 24-hour call to a night float resident coverage system in July 2013. In this cohort study, all patients admitted to the neurosurgical service for 1 year before and 1 year after this transition were compared with respect to hospital and ICU LOSs, adjusted for potential confounders.
A total of 4619 patients were included. In the initial bivariate analysis, night float was associated with increased ICU LOS (p = 0.032) and no change in overall LOS (p = 0.65). However, coincident with the transition to a night float system was an increased frequency of resident service transitions, which were highly associated with hospital LOS (p < 0.01) and ICU LOS (p < 0.01). After adjusting for resident service transitions, initiation of the night float system was associated with decreased hospital LOS (p = 0.047) and no change in ICU LOS (p = 0.35).
This study suggests that a dedicated night float resident may improve night-to-night continuity of care and decrease hospital LOS, but caution must be exercised when initiation of night float results in increased resident service transitions.
Elizabeth N. Kuhn, Matthew C. Davis, Bonita S. Agee, Robert A. Oster and James M. Markert
Handoffs and services changes are potentially modifiable sources of medical error and delays in transition of care. This cohort study assessed the relationship between resident service handoffs and length of stay for neurosurgical patients.
All patients admitted to the University of Alabama at Birmingham neurosurgical service between July 1, 2012, and July 1, 2014, were retrospectively identified. A service handoff was defined as any point when a resident handed off coverage of a service for longer than 1 weekend. A conditional probability distribution was constructed to adjust length of stay for the increasing probability of a random handoff. The Student t-test and ANCOVA were used to assess relationships between resident service handoffs and length of hospital stay, adjusted for potential confounders.
A total of 3038 patients met eligibility criteria and were included in the statistical analyses. Adjusted length of hospital stay (5.32 vs 3.53 adjusted days) and length of ICU stay (4.38 vs 2.96 adjusted days) were both longer for patients who experienced a service handoff, with no difference in mortality. In the ANCOVA model, resident service handoff remained predictive of both length of hospital stay (p < 0.001) and length of ICU stay (p < 0.001).
Occurrence of a resident service handoff is an independent predictor of length of hospital and ICU stay in neurosurgical patients. This finding is novel in the neurosurgical literature. Future research might identify mechanisms for improving continuity of care and mitigating the effect of resident handoffs on patient outcomes.
Ross L. Dawkins, Joseph H. Miller, Omar I. Ramadan, Michael C. Lysek, Elizabeth N. Kuhn, Brandon G. Rocque, Michael J. Conklin, R. Shane Tubbs, Beverly C. Walters, Bonita S. Agee and Curtis J. Rozzelle
There are many classification systems for injuries of the thoracolumbar spine. The recent Thoracolumbar Injury Classification and Severity Score (TLICS) has been shown to be a reliable tool for adult patients. The aim of this study was to assess the reliability of the TLICS system in pediatric patients. The validity of the TLICS system is assessed in a companion paper.
The medical records of pediatric patients with acute, traumatic thoracolumbar fractures at a single Level 1 trauma center were retrospectively reviewed. A TLICS was calculated for each patient using CT and MRI, along with the neurological examination recorded in the patient’s medical record. TLICSs were compared with the type of treatment received. Five raters scored all patients separately to assess interrater reliability.
TLICS calculations were completed for 81 patients. The mean patient age was 10.9 years. Girls represented 51.8% of the study population, and 80% of the study patients were white. The most common mechanisms of injury were motor vehicle accidents (60.5%), falls (17.3%), and all-terrain vehicle accidents (8.6%). The mean TLICS was 3.7 ± 2.8. Surgery was the treatment of choice for 33.3% of patients. The agreement between the TLICS-suggested treatment and the actual treatment received was statistically significant (p < 0.0001). The interrater reliability of the TLICS system ranged from moderate to very good, with a Fleiss’ generalized kappa (κ) value of 0.69 for the TLICS treatment suggestion among all patients; however, interrater reliability decreased when MRI was used to contribute to the TLICS. The κ value decreased from 0.73 to 0.57 for patients with CT only vs patients with CT/MRI or MRI only, respectively (p < 0.0001). Furthermore, the agreement between suggested treatment and actual treatment was worse when MRI was used as part of injury assessment.
The TLICS system demonstrates good interrater reliability among physicians assessing thoracolumbar fracture treatment in pediatric patients. Physicians should be cautious when using MRI to aid in the surgical decision-making process.
Joseph H. Miller, Clarence Gill, Elizabeth N. Kuhn, Brandon G. Rocque, Joshua Y. Menendez, Jilian A. O'Neill, Bonita S. Agee, Steven T. Brown, Marshall Crowther, R. Drew Davis, Drew Ferguson and James M. Johnston
Pediatric sports-related concussions are a growing public health concern. The factors that determine injury severity and time to recovery following these concussions are poorly understood. Previous studies suggest that initial symptom severity and diagnosis of attention deficit hyperactivity disorder (ADHD) are predictors of prolonged recovery (> 28 days) after pediatric sports-related concussions. Further analysis of baseline patient characteristics may allow for a more accurate prediction of which patients are at risk for delayed recovery after a sports-related concussion.
The authors performed a single-center retrospective case-control study involving patients cared for at the multidisciplinary Concussion Clinic at Children's of Alabama between August 2011 and January 2013. Patient demographic data, medical history, sport concussion assessment tool 2 (SCAT2) and symptom severity scores, injury characteristics, and patient balance assessments were analyzed for each outcome group. The control group consisted of patients whose symptoms resolved within 28 days. The case group included patients whose symptoms persisted for more than 28 days. The presence or absence of the SCAT2 assessment had a modifying effect on the risk for delayed recovery; therefore, stratum-specific analyses were conducted for patients with recorded SCAT2 scores and for patients without SCAT2 scores. Unadjusted ORs and adjusted ORs (aORs) for an association of delayed recovery outcome with specific risk factors were calculated with logistic regression analysis.
A total of 294 patients met the inclusion criteria of the study. The case and control groups did not statistically significantly differ in age (p = 0.7). For the patients who had received SCAT2 assessments, a previous history of concussion (aOR 3.67, 95% CI 1.51–8.95), presenting SCAT2 score < 80 (aOR 5.58, 95% CI 2.61–11.93), and female sex (aOR 3.48, 95% CI 1.43–8.49) were all associated with a higher risk for postconcussive symptoms lasting more than 28 days. For patients without SCAT2 scores, female sex and reporting a history of ADHD significantly increased the odds of prolonged recovery (aOR 4.41, 95% CI 1.93–10.07 and aOR 3.87, 95% CI 1.13–13.24, respectively). Concussions resulting from playing a nonhelmet sport were also associated with a higher risk for prolonged symptoms in patients with and without SCAT2 scores (OR 2.59, 95% CI 1.28–5.26 and OR 2.17, 95% CI 0.99–7.73, respectively). Amnesia, balance abnormalities, and a history of migraines were not associated with symptoms lasting longer than 28 days.
This case-control study suggests candidate risk factors for predicting prolonged recovery following sports-related concussion. Large prospective cohort studies of youth athletes examined and treated with standardized protocols will be needed to definitively establish these associations and confirm which children are at highest risk for delayed recovery.