Artificial Intelligence Tools for the Diagnosis of Eosinophilic Esophagitis in Adults Reporting Dysphagia: Development, External Validation, and Software Creation for Point-of-Care Use.

BACKGROUND: Despite increased awareness of eosinophilic esophagitis (EoE), the diagnostic delay has remained stable over the past 3 decades. There is a need to improve the diagnostic performance and optimize resources allocation in the setting of EoE. OBJECTIVE: We developed and validated 2 point-of-care machine learning (ML) tools to predict a diagnosis of EoE before histology results during office visits. METHODS: We conducted a multicenter study in 3 European tertiary referral centers for EoE. We built predictive ML models using retrospectively extracted clinical and esophagogastroduodenoscopy (EGDS) data collected from 273 EoE and 55 non-EoE dysphagia patients. We validated the models on an independent cohort of 93 consecutive patients with dysphagia undergoing EGDS with biopsies at 2 different centers. Models' performance was assessed by area under the curve (AUC), sensitivity, specificity, and positive and negative predictive values (PPV and NPV). The models were integrated into a point-of-care software package. RESULTS: The model trained on clinical data alone showed an AUC of 0.90 and a sensitivity, specificity, PPV, and NPV of 0.90, 0.75, 0.80, and 0.87, respectively, for the diagnosis of EoE in the external validation cohort. The model trained on a combination of clinical and endoscopic data showed an AUC of 0.94, and a sensitivity, specificity, PPV, and NPV of 0.94, 0.68, 0.77, and 0.91, respectively, in the external validation cohort. CONCLUSION: Our software-integrated models (https://webapplicationing.shinyapps.io/PointOfCare-EoE/) can be used at point-of-care to improve the diagnostic workup of EoE and optimize resources allocation.

NEO-SAFE: a clinical model for patients and healthcare personnel safety in primary level hospitals.

Patient safety is a major concern in medicine. Approximately, 4 million infants die each year worldwide and 23% of these deaths are caused by perinatal asphyxia. To prevent the long-term damage of asphyxia, the resuscitation flowchart must be perfectly and promptly performed. However, high effectiveness in performing resuscitation can only be achieved and maintained if the algorithm is frequently executed. Therefore, maintaining a high level of patient care is difficult in some remote centres. The aim of this study was to evaluate the effectiveness of a new organizational model of care-network between Hub & Spoke hospitals to improve both the safety of the newborns in hospitals with a low number of births and the well-being of operators. Our project, NEO-SAFE (NEOnatal SAFety and training Elba), began in 2017 and involved the neonatal intensive care unit and the NINA Center of the Pisa University Hospital (hub) and the Hospital of Elba Island (spoke). It consisted of a continuous training program, both with 'classic' training course and 'on-job tutoring' (on side and remotely), of the health workers at spoke (i.e. nurses, midwives, and paediatricians). All four milestones of the study design were achieved. During the project, NINA Center instructors organized training courses for the staff in Portoferraio. These courses were based on learning technical and non-technical skills in a training course of increasing difficulty. Staff training needs were also monitored during the project by means of periodic questionnaires, sentinel events, and specific requests. The curve described by the rate of newborns transfer to the Pisa neonatal intensive care unit (hub) shows a monotonous decreasing trend line. On the other hand, this project allowed operators to develop greater self-confidence and greater safety in managing emergency situations, reducing stress for them and improving patient safety. The project allowed the creation of a safe, effective, low-cost, and reproducible organizational model for centres with a low number of births. Moreover, the tele-medicine approach is an important improvement in the assistance and is a window on the future.

Improving Pediatric/Neonatology Residents' Newborn Resuscitation Skills With a Digital Serious Game: DIANA.

Background: Serious games, and especially digital game based learning (DGBL) methodologies, have the potential to strengthen classic learning methodology in all medical procedures characterized by a flowchart (e.g., neonatal resuscitation algorithm). However, few studies have compared short- and long-term knowledge retention in DGBL methodologies with a control group undergoing specialist training led by experienced operators. In particular, resident doctors' learning still has limited representation in simulation-based education literature. Objective: A serious computer game DIANA (DIgital Application in Newborn Assessment) was developed, according to newborn resuscitation algorithm, to train pediatric/neonatology residents in neonatal resuscitation algorithm knowledge and implementation (from procedure knowledge to ventilation/chest compressions rate). We analyzed user learning curves after each session and compared knowledge retention against a classic theoretical teaching session. Methods: Pediatric/neonatology residents of the Azienda Ospedaliera Universitaria Pisana (AOUP) were invited to take part in the study and were split into a game group or a control group; both groups were homogeneous in terms of previous training and baseline scores. The control group attended a classic 80 min teaching session with a neonatal trainer, while game group participants played four 20 min sessions over four different days. Three written tests (pre/immediately post-training and at 28 days) were used to evaluate and compare the two groups' performances. Results: Forty-eight pediatric/neonatology residents participated in the study. While classic training by a neonatal trainer demonstrated an excellent effectiveness in short/long-term knowledge retention, DGBL methodology proved to be equivalent or better. Furthermore, after each game session, DGBL score improved for both procedure knowledge and ventilation/chest compressions rate. Conclusions: In this study, DGBL was as effective as classic specialist training for neonatal resuscitation in terms of both algorithm memorization and knowledge retention. User appreciation for the methodology and ease of administration, including remotely, support the use of DGBL methodologies for pediatric/neonatology residents education.

Effectiveness of a new sensorized videolaryngoscope for retraining on neonatal intubation in simulation environment.

BACKGROUND: In recent years, medical training has significantly increased the use of simulation for teaching and evaluation. The retraining of medical personnel in Italy is entrusted to the program of Continuous Education in Medicine, mainly based on theoretical training. The aim of this study is to assess whether the use of a new sensorized platform for the execution of the neonatal intubation procedure in simulation environment can complement theoretical retraining of experienced health professionals. METHODS: Neonatal intubation tests were performed using a commercial manikin and a modified video-laryngoscope by the addition of force and position sensors, which provide the user with feedback when the threshold is exceeded. Two categories carried out the simulation tests: anesthesiologists and pediatricians. The categories were divided into three groups each, and various configurations were tested: the first group of both specialists carried out the tests without feedback (i.e. control groups, gr. A and A1), the second groups received sound and visual feedback from the instrument (gr. B and B1) and the third ones had also the support of a physician expert in the use of the instrument (gr. C and C1). The instrumentation used by pediatricians was provided in a playful form, including a game with increasing difficulty levels. RESULTS: Both in the case with feedback only and in the case with humans support, anesthesiologists did not show a specific trend of improvement. Pediatricians, in comparison with anesthesiologists, showed a positive reaction to both the presence of feedback and that of experienced personnel. Comparing the performance of the two control groups, the two categories of experienced doctors perform similar forces. Pediatricians enjoyed the "Level Game", through which they were able to test and confront themselves, trying to improve their own performance. CONCLUSIONS: Our instrument is more effective when is playful and competitive, introducing something more than just a sound feedback, and allowing training by increasing levels. It is more effective if the users can adapt their own technique to the instrument by themselves, without any external help.