Sustainable improvement in upstream and downstream outcomes for intubated patients three years after an airway-based educational intervention in a low-resource pediatric intensive care unit.

OBJECTIVE: To determine whether implementation of an education-based intervention can sustainably improve upstream and downstream outcomes in intubated patients in a pediatric intensive care unit (PICU) in a low-resource country. DESIGN: Quality improvement study comparing airway-related morbidity in two previously studied patient cohorts pre-intervention (Epoch 1) and immediately post-intervention (Epoch 2) with a third cohort thirty-six months post-intervention (Epoch 3). SETTING: PICU of the largest public children's hospital in El Salvador. PATIENTS: 147 patients under 18 years requiring intubation and mechanical ventilation (MV) met inclusion criteria in the long-term follow-up period and were consecutively sampled without exclusion (Epoch 3) (compared to 98 previously studied patients in the short-term follow-up period (Epoch 2)). INTERVENTION: A low-cost, education-based intervention to close knowledge gaps, improve communication among PICU doctors, nurses, and respiratory therapists, and optimize patient outcomes. MEASUREMENTS AND MAIN RESULTS: The primary outcome measure was change in unplanned extubation (UE) between Epochs 2 and 3. Other outcomes included use of cuffed endotracheal tubes (ETT), rate of elective ETT change and days of MV. The 17 % decrease in UE previously reported for Epoch 2 was sustained in Epoch 3. There was a statistically significant increase in use of cuffed ETT from 35.7 % in Epoch 2-55.1 % in Epoch 3 (p = 0.003, z-score -2.99). There was also a statistically significant mean difference in rate of elective ETT change per 100 MV days from Epoch 2 to Epoch 3 of 1.7 (p = 0.007; 95 % CI 0.15-0.84). There was no change in MV days from Epoch 2 to Epoch 3 (p-value 0.764; 95 % CI -1.48-2.02). Beyond these quantifiable results, many unanticipated practice changes were observed three years after the initial intervention. CONCLUSIONS: Sustained improvement in upstream and downstream outcomes (UE, cuffed ETT use, elective ETT change) for intubated patients in a low-resource PICU were observed three years after a low-cost, low-touch, education-based intervention.

Hypoglossal Nerve Stimulation for Obstructive Sleep Apnea in a Young Child With Down Syndrome.

Obstructive sleep apnea (OSA) is common in children with Down syndrome (DS). Adenoidectomy and/or tonsillectomy are the usual first interventions employed to treat OSA in children with DS but sometimes do not achieve adequate resolution of clinical signs. Positive airway pressure treatment is often used next, but this treatment is poorly tolerated by this population. Persistent OSA can adversely affect a child's health and cognitive development. Hypoglossal nerve stimulation (HGNS), previously shown to be safe and effective in adults with OSA, has been used in children as young as 10 years old with DS and has achieved measurable neurocognitive benefits. The US Food and Drug Administration recently lowered the age for HGNS implantation to 13 years for children with DS. However, questions remain regarding treatment of refractory OSA in younger children. Here, we report the case of a 4-year-old boy with DS and treatment-refractory OSA who underwent successful HGNS implantation. The decision to proceed with HGNS implantation in such a young child involved discussions about anatomic feasibility and potential neurocognitive benefits. The device was implanted without complication and with minimal postoperative bulk. This case suggests a possible treatment option that can be discussed in the course of shared decision-making between clinicians and families of young children with DS and treatment-refractory OSA.

AI Model Versus Clinician Otoscopy in the Operative Setting for Otitis Media Diagnosis.

Prior work has demonstrated improved accuracy in otitis media diagnosis based on otoscopy using artificial intelligence (AI)-based approaches compared to clinician evaluation. However, this difference in accuracy has not been shown in a setting resembling the point-of-care. In this study, we compare the diagnostic accuracy of a machine-learning model to that of pediatricians using standard handheld otoscopes. We find that the model is more accurate than clinicians (90.6% vs 59.4%, P = .01). This is a step towards validation of AI-based diagnosis under more real-world conditions. With further validation, for example on different patient populations and in deployment, this technology could be a useful addition to the clinician's toolbox in accurately diagnosing otitis media.

Harnessing the power of electronic health records and open natural language data mining to capture meaningful patient experience during routine clinical care.

INTRODUCTION: Electronic health records (EHR) are a rich data source for both quality improvement and clinical research. Natural language processing can be harnessed to extract data from these previously difficult to access sources. OBJECTIVE: The objective of this study was to create and apply a natural language search query to extract EHR data to ask and answer quality improvement questions at a pediatric aerodigestive center. METHODS: We developed a combined natural language search query to extract clinically meaningful data along with International Statistical Classification of Diseases (ICD10) and Current Procedural Terminology (CPT) code data. This search query was applied to a single pediatric aerodigestive center to answer key clinical questions asked by families. Data were extracted from EHR data from first clinic visit, operative note, microbiology lab report, and pathology report for all new patients from 2020 to 2021. Included as three queries were: 1) if I bring my child to a pediatric aerodigestive center, how often will my child obtain a medical diagnosis without needing an intervention? 2) if my child has a diagnostic procedure, how often will a diagnosis be made? 3) if a diagnosis is made, can it be addressed during that endoscopic intervention? RESULTS: For the 711 new patients coming to the pediatric aerodigestive center from 2020 to 2021, only 26-32% required an interventional triple endoscopy (rigid/flexible bronchoscopy with esophagoduodenoscopy). Of these triple endoscopies, 75.7% resulted in a positive finding that enabled optimization of that child's care. Of the 221 patients who underwent diagnostic triple endoscopies, 40.7% underwent intervention at the same time for laryngeal cleft (injection or suture, dependent upon age). CONCLUSION: Here we created an effective model of open language search query to extract meaningful metrics of patient experience from EHR data. This model easily allows the EHR to be harnessed to create retrospective and prospective databases that can be readily queried to answer clinical questions important to patients. Such databases are widely applicable not just to pediatric aerodigestive centers but to any clinical care setting using an EHR.

Making Use of Artificial Intelligence-Generated Synthetic Tympanic Membrane Images.

This diagnostic study examines the application of generative artificial intelligence in clinical tool research and development.

Generation of synthetic tympanic membrane images: Development, human validation, and clinical implications of synthetic data.

Synthetic clinical images could augment real medical image datasets, a novel approach in otolaryngology-head and neck surgery (OHNS). Our objective was to develop a generative adversarial network (GAN) for tympanic membrane images and to validate the quality of synthetic images with human reviewers. Our model was developed using a state-of-the-art GAN architecture, StyleGAN2-ADA. The network was trained on intraoperative high-definition (HD) endoscopic images of tympanic membranes collected from pediatric patients undergoing myringotomy with possible tympanostomy tube placement. A human validation survey was administered to a cohort of OHNS and pediatrics trainees at our institution. The primary measure of model quality was the Frechet Inception Distance (FID), a metric comparing the distribution of generated images with the distribution of real images. The measures used for human reviewer validation were the sensitivity, specificity, and area under the curve (AUC) for humans' ability to discern synthetic from real images. Our dataset comprised 202 images. The best GAN was trained at 512x512 image resolution with a FID of 47.0. The progression of images through training showed stepwise "learning" of the anatomic features of a tympanic membrane. The validation survey was taken by 65 persons who reviewed 925 images. Human reviewers demonstrated a sensitivity of 66%, specificity of 73%, and AUC of 0.69 for the detection of synthetic images. In summary, we successfully developed a GAN to produce synthetic tympanic membrane images and validated this with human reviewers. These images could be used to bolster real datasets with various pathologies and develop more robust deep learning models such as those used for diagnostic predictions from otoscopic images. However, caution should be exercised with the use of synthetic data given issues regarding data diversity and performance validation. Any model trained using synthetic data will require robust external validation to ensure validity and generalizability.

Paediatric sleep apnea event prediction using nasal air pressure and machine learning.

Sleep-disordered breathing is an important health issue for children. The objective of this study was to develop a machine learning classifier model for the identification of sleep apnea events taken exclusively from nasal air pressure measurements acquired during overnight polysomnography for paediatric patients. A secondary objective of this study was to differentiate site of obstruction exclusively from hypopnea event data using the model. Computer vision classifiers were developed via transfer learning to either normal breathing while asleep, obstructive hypopnea, obstructive apnea or central apnea. A separate model was trained to identify site of obstruction as either adeno-tonsillar or tongue base. In addition, a survey of board-certified and board-eligible sleep physicians was completed to compare clinician versus model classification performance of sleep events, and indicated very good performance of our model relative to human raters. The nasal air pressure sample database available for modelling comprised 417 normal, 266 obstructive hypopnea, 122 obstructive apnea and 131 central apnea events derived from 28 paediatric patients. The four-way classifier achieved a mean prediction accuracy of 70.0% (95% confidence interval [67.1-72.9]). Clinician raters correctly identified sleep events from nasal air pressure tracings 53.8% of the time, whereas the local model was 77.5% accurate. The site of obstruction classifier achieved a mean prediction accuracy of 75.0% (95% confidence interval [68.7-81.3]). Machine learning applied to nasal air pressure tracings is feasible and may exceed the diagnostic performance of expert clinicians. Nasal air pressure tracings of obstructive hypopneas may "encode" information regarding the site of obstruction, which may only be discernable by machine learning.

Hospital Prices for Pediatric Tympanostomy Tube Placement and Adenotonsillectomy in 2021.

OBJECTIVES: Hospital prices vary substantially for myringotomy with tympanostomy tube placement (M&T) and adenotonsillectomy (T&A). The Centers for Medicare and Medicaid Services recently implemented hospital price transparency requirements to help families make financially informed decisions about where to seek care. We sought to determine price availability and the extent of price variation for these procedures. METHODS: We performed a cross-sectional analysis of the Turquoise Health Hospital Rates Data Platform, which extracts prices for facility fees from publicly available hospital chargemasters. We determined the proportion of hospitals serving pediatric patients that published payer-specific prices for M&T and T&A. We additionally characterized the extent of variation in payer-specific prices both across and within hospitals. RESULTS: Approximately 40% (n = 909 of 2,266 hospitals) serving pediatric patients disclosed prices for M&T or T&A. Among disclosing hospitals, across-center ratios (adjusted for Medicare hospital wage indices) ranged from 11.0 (M&T; 10th percentile adjusted median price: $536.80 versus 90th percentile adjusted median price: $5,929.93) to 23.4 (revision adenoidectomy age >12 years; 10th percentile: $393.82 versus 90th percentile: $9,209.88). Median within-center price ratios for procedures ranged from 2.2 to 2.7, indicating that some private payers reimbursed the same hospital more than twice as much as other payers for the same procedure. CONCLUSION: The majority of hospitals serving pediatric patients were non-compliant with federal requirements to disclose prices for M&T and T&A. Among disclosing hospitals, there was wide variation in payer-specific prices between and within institutions. Further research is necessary to understand whether disclosure of prices will enable families to make more financially informed decisions. LEVEL OF EVIDENCE: 3 Laryngoscope, 133:948-955, 2023.

"Human vs Machine" Validation of a Deep Learning Algorithm for Pediatric Middle Ear Infection Diagnosis.

OBJECTIVE: We compared the diagnostic performance of human clinicians with that of a neural network algorithm developed using a library of tympanic membrane images derived from children taken to the operating room with the intent of performing myringotomy and possible tube placement for recurrent acute otitis media (AOM) or otitis media with effusion (OME). STUDY DESIGN: Retrospective cohort study. SETTING: Tertiary academic medical center from 2018 to 2021. METHODS: A training set of 639 images of tympanic membranes representing normal, OME, and AOM was used to train a neural network as well as a proprietary commercial image classifier from Google. Model diagnostic prediction performance in differentiating normal vs nonpurulent vs purulent effusion was scored based on classification accuracy. A web-based survey was developed to test human clinicians' diagnostic accuracy on a novel image set, and this was compared head to head against our model. RESULTS: Our model achieved a mean prediction accuracy of 80.8% (95% CI, 77.0%-84.6%). The Google model achieved a prediction accuracy of 85.4%. In a validation survey of 39 clinicians analyzing a sample of 22 endoscopic ear images, the average diagnostic accuracy was 65.0%. On the same data set, our model achieved an accuracy of 95.5%. CONCLUSION: Our model outperformed certain groups of human clinicians in assessing images of tympanic membranes for effusions in children. Reduced diagnostic error rates using machine learning models may have implications in reducing rates of misdiagnosis, potentially leading to fewer missed diagnoses, unnecessary antibiotic prescriptions, and surgical procedures.

Aspiration does not mean the end of a breast-feeding relationship.

OBJECTIVE: Breastfeeding is widely recommended as optimal nutrition for infants. However, there are no known publications on the impact of prandial aspiration of breast milk fed infants with dysphagia. The goal of this study was to assess pulmonary outcomes in infants with dysphagia who were given medical clearance for intake of breast milk. METHODS: This retrospective cohort study included review of 80 infants examined between August 2016 to March 2021. Patients were evaluated by an interdisciplinary team of providers in a tertiary pediatric aerodigestive center. Patient inclusion criteria included a VFSS with documented aspiration or penetration with thin liquids. Participants met inclusion criteria if given medical clearance for intake of breast milk despite aspiration risk. Pulmonary health was monitored for three months following medical clearance for the consumption of breast milk. Pulmonary illness was defined as development of bronchiolitis, wheezing, unexplained stridor during feeding, croup, pneumonia, or persistent bacterial bronchitis requiring medical intervention. RESULTS: Forty-three males (54%) and 37 females (46%) enrolled in the study with an age range of 1 month-6 months corrected age. Mean age at initial VFSS was 3.6 months. Twenty-six out of 80 (32.5%) had a report of a mild cough but did not require intervention. Eight out of 80 (10%) received a diagnosis of a pulmonary illness. Seventy-two out of 80 (90%) did not report pulmonary illness. CONCLUSION: This pilot study reveals that the majority (90%) of this single institution, small sample size cohort of breast milk fed infants with documented oropharyngeal dysphagia remained healthy despite continued intake of breast milk. Prospective investigation is warranted to follow pulmonary health outcomes longitudinally and a head to head comparative study would be helpful to identify whether there were indeed significant changes to pulmonary health according to differential feeding regimens offered and followed.

Evaluation of Upper Airway Stimulation for Adolescents With Down Syndrome and Obstructive Sleep Apnea.

Importance: Patients with Down syndrome have a high incidence of persistent obstructive sleep apnea (OSA) and limited treatment options. Upper airway hypoglossal stimulation has been shown to be effective for adults with OSA but has not yet been evaluated for pediatric populations. Objective: To evaluate the safety and effectiveness of upper airway stimulation for adolescent patients with Down syndrome and severe OSA. Design, Setting, and Participants: This prospective single-group multicenter cohort study with 1-year follow-up was conducted between April 1, 2015, and July 31, 2021, among a referred sample of 42 consecutive adolescent patients with Down syndrome and persistent severe OSA after adenotonsillectomy. Intervention: Upper airway stimulation. Main Outcomes and Measures: The prespecified primary outcomes were safety and the change in apnea-hypopnea index (AHI) from baseline to 12 months postoperatively. Polysomnographic and quality of life outcomes were assessed at 1, 2, 6, and 12 months postoperatively. Results: Among the 42 patients (28 male patients [66.7%]; mean [SD] age, 15.1 [3.0] years), there was a mean (SD) decrease in AHI of 12.9 (13.2) events/h (95% CI, -17.0 to -8.7 events/h). With the use of a therapy response definition of a 50% decrease in AHI, the 12-month response rate was 65.9% (27 of 41), and 73.2% of patients (30 of 41) had a 12-month AHI of less than 10 events/h. The most common complication was temporary tongue or oral discomfort, which occurred in 5 patients (11.9%). The reoperation rate was 4.8% (n = 2). The mean (SD) improvement in the OSA-18 total score was 34.8 (20.3) (95% CI, -42.1 to -27.5), and the mean (SD) improvement in the Epworth Sleepiness Scale score was 5.1 (6.9) (95% CI, -7.4 to -2.8). The mean (SD) duration of nightly therapy was 9.0 (1.8) hours, with 40 patients (95.2%) using the device at least 4 hours a night. Conclusions and Relevance: Upper airway stimulation was able to be safely performed for 42 adolescents who had Down syndrome and persistent severe OSA after adenotonsillectomy with positive airway pressure intolerance. There was an acceptable adverse event profile with high rates of therapy response and quality of life improvement. Trial Registration: ClinicalTrials.gov Identifier: NCT02344108.

Hybrid laryngotracheal reconstruction vs single and double stage: Indications and outcomes.

OBJECTIVES: To describe outcomes from laryngotracheal reconstruction and decannulation rates for patients undergoing single stage, double stage and hybrid staged procedures at a single tertiary care institution and evaluate if the 1.5LTR is a viable reconstructive option for patients with subglottic stenosis. STUDY DESIGN: Retrospective chart review. SETTING: Tertiary care otolaryngology specialty hospital and internationally. SUBJECTS: All patients who underwent LTR by a single pediatric otolaryngology surgeon from 2008 to 2018. METHODS: Charts were assessed for age, gender, etiology, type of reconstruction, comorbidities, length of stay, tracheostomy status and socioeconomic status. Analysis was performed using Microsoft Excel and multivariate logistic regression models. RESULTS: 96 patients underwent laryngotracheal reconstruction at MEEI. Internationally, 36 patients underwent laryngotracheal reconstruction with the primary surgeon. Overall decannulation rates for ssLTR, dsLTR, and 1.5LTR were 95.6%, 77.8%, and 91.2% respectively. Our Operation Specific Decannulation Rates (one open airway procedure only) for ssLTR, dsLTR, and 1.5LTR were 87.5%, 33%, and 88% respectively. Adjusted odds of decannulation were not significantly different between males and females, white and non-white patients, or socioeconomic status. Neurological comorbidity was statistically significant for a decreased rate of decannulation (p = 0.0216). CONCLUSION: The 1.5LTR is a viable option for airway reconstruction with strengths derived from both the ssLTR and dsLTR. At our institution we have seen decannulation rates and operation specific decannulation rates with the 1.5LTR approaching our ssLTR. It has replaced the bulk of our dsLTRs, which we reserve for patients that have significant neurological deficits and cannot tolerate extended sedation. LEVEL OF EVIDENCE: 4.

Long-term stability of hypoglossal nerve stimulation for the treatment of obstructive sleep apnea in children with Down syndrome.

IMPORTANCE: Obstructive sleep apnea (OSA) occurs in 55-97% of people with Down syndrome (DS). Even after adenotonsillectomy, residual OSA often persists into adulthood due, in part, to tongue base collapse. Implantable hypoglossal nerve stimulators are being investigated in children and young adults with DS and persistent, moderate to severe OSA. However, the long-term necessity for such an intervention-especially as patients mature and voltage adjustment becomes warranted-has not been previously reported in the pediatric DS population. OBJECTIVE: To assess the long-term need for implantable hypoglossal nerve stimulators and the necessity for voltage adjustment in children and young adults with Down syndrome. DESIGN: This is a case series from an ongoing clinical trial assessing safety and efficacy of hypoglossal nerve stimulation among 42 children and young adults with DS and persistent OSA, despite adenotonsillectomy and trialed positive airway pressure (PAP) therapy. We focus here on the first 4 participants who have undergone implantation by age 13 and have completed at least 44 months of follow-up. PARTICIPANTS: 4 participants (2 male, 2 female; ages 10-13 years) with DS and persistent, severe OSA (AHI > 10 events/h) underwent hypoglossal nerve stimulator implantation and were followed for 44-58 months. SETTING: Participants completed in-lab sleep studies at baseline (before implantation), 1 year postoperatively, and 44-58 months postoperatively. During their most recent follow-up, 2 participants completed split-night sleep studies in which assessment was done with the device both on and off. INTERVENTIONS: Hypoglossal nerve stimulator implantation. MAIN OUTCOMES AND MEASURES: Stability in titrated and untitrated OSA as measured by the apnea-hypopnea index (AHI); growth measures including BMI; and quality of life as measured by the OSA-18 questionnaire. RESULTS: Compared to baseline, all 4 participants maintained reductions of at least 50% in AHI over the course of follow-up. At recent follow-up, two participants had persistent, moderate OSA despite stimulation therapy. The other two participants achieved 100% reductions in AHI with stimulation therapy; when they underwent split-night sleep studies, the severe OSA persisted with the device turned off. Improvement in OSA-18 quality of life scores was observed in three of the four participants. CONCLUSION: and Relevance: Hypoglossal nerve stimulation continues to effectively control OSA in children with DS as they mature, while their underlying untitrated OSA appears to persist into adulthood. TRIAL REGISTRATION: clinicaltrials.gov Identifier: NCT2344108.

A low-cost educational intervention to reduce unplanned extubation in low-resourced pediatric intensive care units.

INTRODUCTION: Unplanned extubation (UE) is orders of magnitude worse in low-income Pediatric Intensive Care Units (PICUs) than their high-income counterparts. Furthermore, a significant percent (20 %) of UEs result in a destabilizing event or cardiac collapse that negatively contributes to morbidity and mortality. As the principles of safe airway management are universal, we hypothesize that a multi-disciplinary educational intervention bundle which included provision of low-cost cuffed endotracheal tubes (ETT) and ETT tape will decrease the rate of unplanned extubation (UE) in a low-resourced PICU. METHODS: This is a pre-post interventional study powered to evaluate UE of intubated pediatric patients in an El Salvadorian PICU after a multi-disciplinary educational effort and provision of low-cost disposable materials. A multidisciplinary (otolaryngologists, intensivists, anesthesiologists, respiratory therapists, and nurses) educational curriculum involving hands on training, online video modules readily available via bedside QR codes, and pre- and post-testing was administered. The cost of the intervention materials was $1.32 per child. PICU mortality was evaluated as an exploratory outcome. RESULTS: Nine-hundred and fifty-seven (859 pre-intervention and 98 post-intervention) patients met inclusion criteria. Patients with one or more UEs decreased significantly from 29.4 % to 17.3 % post-intervention (p = 0.01; CI: 0.28-0.88) with an odds ratio of 0.51. The use of a cuffed ETT increased from 12 % to 36 % (p < 0.001; CI: 0.17-0.44; OR:3.74) and cuffed ETT use was associated with a reduction in UE with an odds ratio of 0.40 (p < 0.001; CI: 0.24-0.66). Finally, there was a 4.3 % decrease in pediatric mortality from 26.7 % to 22.4 % that equates to a number needed to treat to prevent a single child mortality of 23. Therefore, the ICER per mortality prevented is $30.7 and the ICER per Disability Adjusted Life Year (DALY) is $0.44. CONCLUSION: This multi-faceted intervention bundle is an accessible, scalable, cost-effective means to reduce UE and has implications in reducing global pediatric mortality.

Opioid prescribing practices in academic otolaryngology: A single institutional survey.

OBJECTIVES: Opioids are highly addictive medications and otolaryngologists have a responsibility to practice opioid stewardship. We investigated postoperative opioid prescribing patterns among resident and attending physicians as an educational platform to underscore the importance of conscientious opioid prescribing. METHODS: This quality improvement study was designed as a cross-sectional electronic survey. Residents and attending clinical faculty members at a single academic institution were queried from February through April 2020. An electronic survey was distributed to capture postoperative opioid prescribing patterns after common procedures. At the conclusion of the study, results were sent to all faculty and residents. RESULTS: A total of 29 attending otolaryngologists and 22 residents completed the survey. Resident physicians prescribed on average fewer postoperative opioid pills than attendings. Among attendings, the largest number of opioids were prescribed following tonsillectomy (dose varied by patient age), neck dissection (12.6 pills), brow lift (13.3 pills), facelift (13.3 pills), and open reduction of facial trauma (10.7 pills). For residents, surgeries with the most postoperatively prescribed opioids were for tonsillectomy (varied by patient age), neck dissection (13.4 pills), open reduction of facial trauma (10.5 pills), parotidectomy (10.0 pills), and thyroid/parathyroidectomy (9.0 pills). The largest volume of postoperative opioids for both groups was prescribed following tonsillectomy. Attendings prescribed significantly more opioids after facelift and brow lift than did residents (p = 0.01 and p = 0.003, respectively). CONCLUSION: There was good concordance between resident and attending prescribers. Improvement in opioid prescribing and pain management should be an essential component of otolaryngology residency education and attending continuing medical education. LEVEL OF EVIDENCE: 4.

Machine Learning for Accurate Intraoperative Pediatric Middle Ear Effusion Diagnosis.

OBJECTIVES: Misdiagnosis of acute and chronic otitis media in children can result in significant consequences from either undertreatment or overtreatment. Our objective was to develop and train an artificial intelligence algorithm to accurately predict the presence of middle ear effusion in pediatric patients presenting to the operating room for myringotomy and tube placement. METHODS: We trained a neural network to classify images as " normal" (no effusion) or "abnormal" (effusion present) using tympanic membrane images from children taken to the operating room with the intent of performing myringotomy and possible tube placement for recurrent acute otitis media or otitis media with effusion. Model performance was tested on held-out cases and fivefold cross-validation. RESULTS: The mean training time for the neural network model was 76.0 (SD ± 0.01) seconds. Our model approach achieved a mean image classification accuracy of 83.8% (95% confidence interval [CI]: 82.7-84.8). In support of this classification accuracy, the model produced an area under the receiver operating characteristic curve performance of 0.93 (95% CI: 0.91-0.94) and F1-score of 0.80 (95% CI: 0.77-0.82). CONCLUSIONS: Artificial intelligence-assisted diagnosis of acute or chronic otitis media in children may generate value for patients, families, and the health care system by improving point-of-care diagnostic accuracy. With a small training data set composed of intraoperative images obtained at time of tympanostomy tube insertion, our neural network was accurate in predicting the presence of a middle ear effusion in pediatric ear cases. This diagnostic accuracy performance is considerably higher than human-expert otoscopy-based diagnostic performance reported in previous studies.

Association of perioperative ibuprofen exposure with post-tonsillectomy bleeding requiring operative management.

INTRODUCTION: Pediatric tonsillectomy is one of the most common surgical procedures performed in the United States. The safety of ibuprofen use after surgery is debated given concern for increased bleeding. The primary objective of this study was to compare the rate of post-tonsillectomy hemorrhage requiring operative management in patients who received ibuprofen perioperatively vs. patients who did not. METHODS: Retrospective cohort study of patients 0-18 years old who underwent tonsillectomy with or without adenoidectomy (T&A) with recorded inpatient medication administration data at a single tertiary care institution from 1/2005-1/2019. The association between perioperative medication administration and return to operating room (OR) for control was evaluated using multivariable logistic regression adjusted for patient demographics and operative indication. Secondary outcomes evaluated included the time to operative bleed when it occurred. RESULTS: A total of 4098 patients with a median age of 6 years old (IQR 4-10) underwent T&A over the study period. The overall rate of post-tonsillectomy hemorrhage requiring OR was 3.37% (n = 138/4098). After adjustment for confounders, the odds of bleeding requiring OR did not differ significantly between the ibuprofen (OR 1.16, 95% CI (0.76, 1.74), 3.55%, n = 41/1,156, p = 0.47) and non-ibuprofen groups (3.30%, n = 97/2942). The median time to bleeding requiring OR was postoperative day 6.5 (IQR6-8) in the ibuprofen group and day 6 (IQR 3-8) in the non-ibuprofen group. CONCLUSIONS: No difference in post-tonsillectomy hemorrhage requiring OR was observed between patients receiving perioperative ibuprofen versus those patients not receiving this medication. Additional research is required to definitively determine a safe dose and interval for ibuprofen administration following tonsillectomy.