Sinus Arrest Related to Dexmedetomidine Infusion in an Infant; a Case Report and Review of Current Literature.

Background: Dexmedetomidine, an alpha 2 agonist, has emerged as a desirable sedative agent in the pediatric intensive care unit due to its minimal effect on respiratory status and reduction in delirium. Bradycardia and hypotension are common side effects, however there are emerging reports of more serious cardiovascular events, including sinus arrest and asystole. These case reports have been attributed to high vagal tone or underlying cardiac conduction dysfunction. Objectives: To describe the development of sinus arrest during sedation with dexmedetomidine in a patient without clinical features of high vagal tone, underlying cardiac conduction dysfunction, or intervening episodes of bradycardia. Case Presentation: An 11 month-old patient requiring sedation during mechanical ventilation for acute respiratory failure secondary to Adenovirus. To facilitate sedation, a dexmedetomidine infusion was initiated at .5 mcg/kg/hr and increased to maximum 1 mcg/kg/hr. Within 8 hours of initiating therapy, the patient had three episodes of sinus arrest. There was no intervening bradycardia between episodes and no further episodes occurred following discontinuation of dexmedetomidine. The patient did not have any clinical features associated with high vagal tone or underlying cardiac conduction dysfunction. Conclusions: As result of these findings, understanding risk factors for bradycardia, or more serious hemodynamic instability with dexmedetomidine infusions, is important to help identify high risk patients and weigh the associated risks and benefits of its administration.

Comparison of Bedside and Video-Based Capillary Refill Time Assessment in Children.

OBJECTIVES: Capillary refill time (CRT) to assess peripheral perfusion in children with suspected shock may be subject to poor reproducibility. Our objectives were to compare video-based and bedside CRT assessment using a standardized protocol and evaluate interrater and intrarater consistency of video-based CRT (VB-CRT) assessment. We hypothesized that measurement errors associated with raters would be low for both standardized bedside CRT and VB-CRT as well as VB-CRT across raters. METHODS: Ninety-nine children (aged 1-12 y) had 5 consecutive bedside CRT assessments by an experienced critical care clinician following a standardized protocol. Each CRT assessment was video recorded on a black background. Thirty video clips (10 with bedside CRT < 1 s, 10 with CRT 1-2 s, and 10 with CRT > 2 s) were randomly selected and presented to 10 clinicians twice in randomized order. They were instructed to push a button when they visualized release of compression and completion of a capillary refill. The correlation and absolute difference between bedside and VB-CRT were assessed. Consistency across raters and within each rater was analyzed using the intraclass correlation coefficient (ICC). A Generalizability study was performed to evaluate sources of variation. RESULTS: We found moderate agreement between bedside and VB-CRT observations (r = 0.65; P < 0.001). The VB-CRT values were shorter by 0.17 s (95% confidence interval, 0.09-0.25; P < 0.001) on average compared with bedside CRT. There was moderate agreement in VB-CRT across raters (ICC = 0.61). Consistency of repeated VB-CRT within each rater was moderate (ICC = 0.71). Generalizability study revealed the source of largest variance was from individual patient video clips (57%), followed by interaction of the VB-CRT reviewer and patient video clip (10.7%). CONCLUSIONS: Bedside and VB-CRT observations showed moderate consistency. Using video-based assessment, moderate consistency was also observed across raters and within each rater. Further investigation to standardize and automate CRT measurement is warranted.

SPRINTing to Innovation: Children's Hospital of Philadelphia's Strategic Approach to Discovering Its Untapped Innovation Potential.

PROBLEM: There is a clear and urgent need for health care innovation in the United States. Hospital employees routinely recognize pain points that affect care delivery and are in a unique position to propose innovative and practical solutions, yet leaders rarely solicit ideas for investment and development from frontline providers and staff, revealing an untapped resource with innovation potential. APPROACH: To address these deficiencies, the Children's Hospital of Philadelphia expanded its innovation infrastructure with the competition-based SPRINT program in 2015. All hospital employees are encouraged to apply with early-stage innovative ideas, and if selected, are provided with business, legal, technical, and scientific project management support to help accelerate their projects toward commercial viability. SPRINT was modeled around 4 core tenets: (1) small, dynamic, and attentive project manager-led teams; (2) low barriers to entry; (3) emphasis on outreach; and (4) fostering innovators. OUTCOMES: Over its first 4 cycles from 2015 to 2018, 271 innovative teams applied to the SPRINT program, which led to support for 30 projects (11% acceptance rate). About a quarter of the projects each year were submitted by physician-led teams (mean 23%), a third by nonphysician clinical providers (mean 33%), and almost half were submitted by employees without direct patient contact (mean 44%). Nurses have emerged as the largest applicant group. Eleven of the SPRINT-supported projects (37%) resulted in commercial endpoints. NEXT STEPS: SPRINT has proven to be an effective model for supporting institution-wide, employee-driven health care innovation, especially among frontline clinical and nonclinical personnel. Critical next steps for the program include a formal cost-benefit analysis and the earlier participation of technology transfer and intellectual property experts to improve the commercialization roadmap for many SPRINT projects.

Supervised Machine Learning Applied to Automate Flash and Prolonged Capillary Refill Detection by Pulse Oximetry.

OBJECTIVE: Develop an automated approach to detect flash (<1.0 s) or prolonged (>2.0 s) capillary refill time (CRT) that correlates with clinician judgment by applying several supervised machine learning (ML) techniques to pulse oximeter plethysmography data. MATERIALS AND METHODS: Data was collected in the Pediatric Intensive Care Unit (ICU), Cardiac ICU, Progressive Care Unit, and Operating Suites in a large academic children's hospital. Ninety-nine children and 30 adults were enrolled in testing and validation cohorts, respectively. Patients had 5 paired CRT measurements by a modified pulse oximeter device and a clinician, generating 485 waveform pairs for model training. Supervised ML models using gradient boosting (XGBoost), logistic regression (LR), and support vector machines (SVMs) were developed to detect flash (<1 s) or prolonged CRT (≥2 s) using clinician CRT assessment as the reference standard. Models were compared using Area Under the Receiver Operating Curve (AUC) and precision-recall curve (positive predictive value vs. sensitivity) analysis. The best performing model was externally validated with 90 measurement pairs from adult patients. Feature importance analysis was performed to identify key waveform characteristics. RESULTS: For flash CRT, XGBoost had a greater mean AUC (0.79, 95% CI 0.75-0.83) than logistic regression (0.77, 0.71-0.82) and SVM (0.72, 0.67-0.76) models. For prolonged CRT, XGBoost had a greater mean AUC (0.77, 0.72-0.82) than logistic regression (0.73, 0.68-0.78) and SVM (0.75, 0.70-0.79) models. Pairwise testing showed statistically significant improved performance comparing XGBoost and SVM; all other pairwise model comparisons did not reach statistical significance. XGBoost showed good external validation with AUC of 0.88. Feature importance analysis of XGBoost identified distinct key waveform characteristics for flash and prolonged CRT, respectively. CONCLUSION: Novel application of supervised ML to pulse oximeter waveforms yielded multiple effective models to identify flash and prolonged CRT, using clinician judgment as the reference standard. TWEET: Supervised machine learning applied to pulse oximeter waveform features predicts flash or prolonged capillary refill.