Summary: A model to classify the difficulty of videolaryngoscopic tracheal intubation has yet to be established. The videolaryngoscopic intubation and difficult airway classification (VIDIAC) study aimed to develop one based on variables associated with difficult videolaryngoscopic tracheal intubation. We studied 374 videolaryngoscopic tracheal intubations in 320 adults scheduled for ear, nose and throat or oral and maxillofacial surgery, for whom airway management was expected to be difficult. The primary outcome was whether an anaesthetist issued a ‘difficult airway alert’ after videolaryngoscopy. An alert was issued after 183 (49%) intubations. Random forest and lasso regression analysis selected six intubation‐related variables associated with issuing an alert: impaired epiglottic movement; increased lifting force; direct epiglottic lifting; vocal cords clearly visible; vocal cords not visible; and enlarged arytenoids. Internal validation was performed by a 10‐fold cross‐validation, repeated 20 times. The mean (SD or 95%CI) area under the receiver operating characteristic curve was 0.92 (0.05) for the cross validated coefficient model and 0.92 (0.89–0.95) for a simplified unitary score (VIDIAC score with component values of −1 or 1 only). The calibration belt for the coefficient model was consistent with observed alert probabilities, from 0% to 100%, while the unitary VIDIAC score overestimated probabilities < 20% and underestimated probabilities > 70%. Discrimination of the VIDIAC score for patients more or less likely to be issued an alert was better than discrimination by the Cormack–Lehane classification, with mean (95%CI) areas under the receiver operating characteristic curve of 0.92 (0.89–0.95) vs. 0.75 (0.70–0.80), respectively, p < 0.001. Our model and score can be used to calculate the probabilities of difficult airway alerts after videolaryngoscopy.

Collaborative robotic needle insertions have the potential to improve placement accuracy and safety, e.g., during epidural anesthesia. Epidural anesthesia provides effective regional pain management but can lead to serious complications, such as nerve injury or cerebrospinal fluid leakage. Robotic assistance might prevent inadvertent puncture by providing haptic feedback to the physician. Haptic feedback can be realized on the basis of force measurements at the needle. However, contact should be avoided for delicate structures. We propose a proximity-based method to provide feedback prior to contact. We measure the distance to boundary layers, visualize the proximity for the operator and further feedback it as a haptic resistance. We compare our approach to haptic feedback based on needle forces and visual feedback without haptics. Participants are asked to realize needle insertions with each of the three feedback modes. We use phantoms that mimic the structures punctured during epidural anesthesia. We show that visual feedback improves needle placement, but only proximity-based haptic feedback reduces accidental puncture. The puncture rate is 62% for force-based haptic feedback, 60% for visual feedback and 6% for proximity-based haptic feedback. Final needle placement inside the epidural space is achieved in 38%, 70% and 96% for force-based haptic, visual and proximity-based haptic feedback, respectively. Our results suggest that proximity-based haptic feedback could improve needle placement safety in the context of epidural anesthesia.

Background: This study aimed to determine whether prehospital qSOFA (quick sequential organ failure assessment) assessment was associated with a shortened ‘time to antibiotics’ and ‘time to intravenous fluid resuscitation’ compared with standard assessment. Methods: This retrospective study included patients who were referred to our Emergency Department between 2014 and 2018 by emergency medical services, in whom sepsis was diagnosed during hospitalization. Two multivariable regression models were fitted, with and without qSOFA parameters, for ‘time to antibiotics’ (primary endpoint) and ‘time to intravenous fluid resuscitation’. Results: In total, 702 patients were included. Multiple linear regression analysis showed that antibiotics and intravenous fluids were initiated earlier if infections were suspected and emergency medical services involved emergency physicians. A heart rate above 90/min was associated with a shortened time to antibiotics. If qSOFA parameters were added to the models, a respiratory rate ≥ 22/min and altered mentation were independent predictors for earlier antibiotics. A systolic blood pressure ≤ 100 mmHg and altered mentation were independent predictors for earlier fluids. When qSOFA parameters were added, the explained variability of the model increased by 24% and 38%, respectively (adjusted R² 0.106 versus 0.131 for antibiotics and 0.117 versus 0.162 for fluids). Conclusion: Prehospital assessment of qSOFA parameters was associated with a shortened time to a targeted sepsis therapy.

Missing data present a major challenge for deep learning, and various imputation techniques exist. However, imputation quality generally decreases as missing data rates increase. In time-series data from electronic health records, missing rates for individual parameters up to 99% can be observed, posing a problem for mere imputation. In addition, underlying patterns between missing and observed data can contain valuable information that can be learned by deep learning models. In this work, we propose a strategy called missing adjusted normalization and nullity encoding representation strategy (MANNERS), which can be applied to training pipelines for representation learning. MANNERS encodes missingness, masks loss evaluation at missing data points, and applies rebalancing such that the signal from variables with high missing rates is not lost. We evaluated MANNERS on reconstruction and downstream classification, regression, and synthetic data generation tasks. We showed performance improvements in the presence of very high missing rates compared with state-of-the-art imputation-only techniques.