Severe pain management in the emergency department: patient pathway as a new factor associated with IV morphine prescription.

Background: Across the world, 25-29% of the population suffer from pain. Pain is the most frequent reason for an emergency department (ED) visit. This symptom is involved in approximately 70% of all ED visits. The effective management of acute pain with adequate analgesia remains a challenge, especially for severe pain. Intravenous (IV) morphine protocols are currently indicated. These protocols are based on patient-reported scores, most often after an immediate evaluation of pain intensity at triage. However, they are not systematically prescribed. This aspect could be explained by the fact that physicians individualize opioid pain management for each patient and each care pathway to determine the best benefit-risk balance. Few data are available regarding bedside organizational factors involved in this phenomenon. Objective: This study aimed to analyze the organizational factors associated with no IV morphine prescription in a standardized context of opioid management in a tertiary-care ED. Methods: A 3-month prospective study with a case-control design was conducted in a French university hospital ED. This study focused on factors associated with protocol avoidance despite a visual analog scale (VAS) ≥60 or a numeric rating scale (NRS) ≥6 at triage. Pain components, physician characteristics, patient epidemiologic characteristics, and care pathways were considered. Qualitative variables (percentages) were compared using Fisher's exact test or the chi-squared tests. Student's t-test was used to compare continuous variables. The results were expressed as means with their standard deviation (SD). Factors associated with morphine avoidance were identified by logistic regression. Results: A total of 204 patients were included in this study. A total of 46 cases (IV morphine) and 158 controls (IV morphine avoidance) were compared (3:1 ratio). Pain patterns and patient's epidemiologic characteristics were not associated with an IV morphine prescription. Regarding NRS intervals, the results suggest a practice disconnected from the patient's initial self-report. IV morphine avoidance was significantly associated with care pathways. A significant difference between the IV morphine group and the IV morphine avoidance group was observed for "self-referral" [adjusted odds ratio (aOR): 5.11, 95% CIs: 2.32-12.18, p < 0.0001] and patients' trajectories (Fisher's exact test; p < 0.0001), suggesting IV morphine avoidance in ambulatory pathways. In addition, "junior physician grade" was associated with IV morphine avoidance (aOR: 2.35, 95% CIs: 1.09-5.25, p = 0.03), but physician gender was not. Conclusion: This bedside case-control study highlights that IV morphine avoidance in the ED could be associated with ambulatory pathways. It confirms the decreased choice of "NRS-only" IV morphine protocols for all patients, including non-trauma patterns. Modern pain education should propose new tools for pain evaluation that integrate the heterogeneity of ED pathways.

Bedside Evaluation of Early VAS/NRS Based Protocols for Intravenous Morphine in the Emergency Department: Reasons for Poor Follow-Up and Targeted Practices.

Intravenous (IV) morphine protocols based on patient-reported scores, immediately at triage, are recommended for severe pain in Emergency Departments. However, a low follow-up is observed. Scarce data are available regarding bedside organization and pain etiologies to explain this phenomenon. The objective was the real-time observation of motivations and operational barriers leading to morphine avoidance. In a single French hospital, 164 adults with severe pain at triage were included in a cross-sectional study of the prevalence of IV morphine titration; caregivers were interviewed by real-time questionnaires on "real" reasons for protocol avoidance or failure. IV morphine prevalence was 6.1%, prescription avoidance was mainly linked to "Pain reassessment" (61.0%) and/or "alternative treatment prioritization" (49.3%). To further evaluate the organizational impact on prescription decisions, a parallel assessment of "simulated" prescription conditions was simultaneously performed for 98/164 patients; there were 18 titration decisions (18.3%). Treatment prioritization was a decision driver in the same proportion, while non-eligibility for morphine was more frequently cited (40.6% p = 0.001), with higher concerns about pain etiologies. Anticipation of organizational constraints cannot be excluded. In conclusion, IV morphine prescription is rarely based on first pain scores. Triage assessment is used for screening by bedside physicians, who prefer targeted practices to automatic protocols.

Hybrid computing using a neural network with dynamic external memory.

Artificial neural networks are remarkably adept at sensory processing, sequence learning and reinforcement learning, but are limited in their ability to represent variables and data structures and to store data over long timescales, owing to the lack of an external memory. Here we introduce a machine learning model called a differentiable neural computer (DNC), which consists of a neural network that can read from and write to an external memory matrix, analogous to the random-access memory in a conventional computer. Like a conventional computer, it can use its memory to represent and manipulate complex data structures, but, like a neural network, it can learn to do so from data. When trained with supervised learning, we demonstrate that a DNC can successfully answer synthetic questions designed to emulate reasoning and inference problems in natural language. We show that it can learn tasks such as finding the shortest path between specified points and inferring the missing links in randomly generated graphs, and then generalize these tasks to specific graphs such as transport networks and family trees. When trained with reinforcement learning, a DNC can complete a moving blocks puzzle in which changing goals are specified by sequences of symbols. Taken together, our results demonstrate that DNCs have the capacity to solve complex, structured tasks that are inaccessible to neural networks without external read-write memory.