Development and testing of a deep learning algorithm to detect lung consolidation among children with pneumonia using hand-held ultrasound.

BACKGROUND AND OBJECTIVES: Severe pneumonia is the leading cause of death among young children worldwide, disproportionately impacting children who lack access to advanced diagnostic imaging. Here our objectives were to develop and test the accuracy of an artificial intelligence algorithm for detecting features of pulmonary consolidation on point-of-care lung ultrasounds among hospitalized children. METHODS: This was a prospective, multicenter center study conducted at academic Emergency Department and Pediatric inpatient or intensive care units between 2018-2020. Pediatric participants from 18 months to 17 years old with suspicion of lower respiratory tract infection were enrolled. Bedside lung ultrasounds were performed using a Philips handheld Lumify C5-2 transducer and standardized protocol to collect video loops from twelve lung zones, and lung features at both the video and frame levels annotated. Data from both affected and unaffected lung fields were split at the participant level into training, tuning, and holdout sets used to train, tune hyperparameters, and test an algorithm for detection of consolidation features. Data collected from adults with lower respiratory tract disease were added to enrich the training set. Algorithm performance at the video level to detect consolidation on lung ultrasound was determined using reference standard diagnosis of positive or negative pneumonia derived from clinical data. RESULTS: Data from 107 pediatric participants yielded 117 unique exams and contributed 604 positive and 589 negative videos for consolidation that were utilized for the algorithm development process. Overall accuracy for the model for identification and localization of consolidation was 88.5%, with sensitivity 88%, specificity 89%, positive predictive value 89%, and negative predictive value 87%. CONCLUSIONS: Our algorithm demonstrated high accuracy for identification of consolidation features on pediatric chest ultrasound in children with pneumonia. Automated diagnostic support on an ultraportable point-of-care device has important implications for global health, particularly in austere settings.

Point-of-Care Ultrasound for Earlier Detection of Pediatric Pneumonia.

Case Presentation: An 8-month-old infant presented to a general emergency department with chief complaints of rhinorrhea, decreased activity, and fever. A point-of-care lung ultrasound (LUS) was performed at bedside with potential early findings of pneumonia. Based on these findings on LUS, a chest radiograph (CXR) was ordered and performed with no acute findings. He was discharged without antibiotics based on these findings; unfortunately, he returned two days later with worsening symptoms requiring chest tube placement, mechanical ventilation, and prolonged hospitalization for complicated bacterial pneumonia. Discussion: Pneumonia is a major cause of pediatric morbidity and mortality worldwide. Despite evidence supporting the utilization of LUS for the diagnosis of pediatric pneumonia, CXR remains the default imaging for clinical decision-making in most settings. In this case, earlier antibiotics and higher reliance on LUS for clinical decision-making may have prevented the morbidity associated with this hospitalization.

Management of Isolated Skull Fractures in Pediatric Patients: A Systematic Review.

OBJECTIVES: Isolated skull fractures (ISFs) in children are one of the most common emergency department injuries. Recent studies suggest these children may be safely discharged following ED evaluation with little risk of delayed neurological compromise. The aim of this study was to propose an evidence-based protocol for the management of ISF in children in an effort to reduce medically unnecessary hospital admissions. METHODS: Using PubMed and The Cochrane Library databases, a literature search using the search terms (pediatric OR child) AND skull fracture AND (isolated OR linear) was performed. Three hundred forty-three abstracts were identified and screened based on the inclusion criteria: (1) linear, nondepressed ISF; (2) no evidence of intracranial injury; (3) age 18 years or younger; and (4) data on patient outcomes and management. Data including age, Glasgow Coma Scale score on arrival, repeat imaging, admission rates, need for neurosurgical intervention, and patient outcome were collected. Two authors reviewed each study for data extraction and quality assessment. RESULTS: Fourteen articles met the eligibility criteria. Data including admission rates, outcomes, and necessity of neurosurgical intervention were analyzed. Admission rates ranged from 56.8% to 100%; however, only 8 of more than 5000 patients developed new imaging findings after admission, all of which were nonsurgical. Only 1 patient required neurosurgical intervention for a finding evident upon initial evaluation. CONCLUSIONS: Pediatric ISF patients with a presenting Glasgow Coma Scale score of 15 who are neurologically intact and tolerating feeds without concern for nonaccidental trauma or an unstable social environment can safely be discharged following ED evaluation to a responsible caregiver.

The Burden of Cost in Bronchiolitis Obliterans Syndrome: Predictions for the Next Decade.

In health economics, costs can be divided into both direct and indirect categories. Direct costs tend to consist of medical costs, which are those directly attributed to health care interventions (e.g., hospitalizations, pharmaceuticals, devices), and non-medical direct costs such as monitoring and professional caregiving. Indirect costs tend to comprise those related to lost productivity due to illness (or treatment), burden on systems outside of the healthcare domain, and other costs that can sometimes outweigh the entire sum of direct healthcare costs. The most common life-threatening complication of lung and hematopoietic stem-cell transplantation (HSCT) is bronchiolitis obliterans syndrome (BOS). BOS is currently diagnosed as a 20% decline in the forced expiratory volume in one second (FEV1) from the best (baseline) post-transplantation value, and is a major cause of morbidity and mortality amongst lung and stem cell transplant patients. BOS affects half of all lung transplant patients within the first 5 years post-transplant, rising to the majority of patients (~80%) within the first decade following transplant. We estimated both direct and indirect costs for the first 10 years following BOS diagnosis, a viewpoint that highlights a tremendous imbalance between healthcare and non-healthcare costs. The lost workforce resulting from BOS-related infirmity will cost society more than $3.7 Billion over the next decade, a figure that is more than double the estimated 10-year cost of treating BOS ($1.4B), including diagnostics, immunosuppressives, and additional complications. As such, BOS is estimated to present a burden of cost that must be evaluated in a new light to include the wider societal perspective.

Estimating the Burden of Cost in Chronic Graft-versus-Host Disease: A Human Capital Approach.

With advances in organ matching and preventing acute graft-versus-host-disease (aGvHD), chronic graft-versus-host disease (cGvHD) following allogeneic hematopoietic stem cell transplantation (HSCT) has become a focus of transplant-related morbidity and mortality. Given that cGvHD often presents years following a transplant, our objective was to estimate its burden of cost resulting from allogeneic HSCT based on published estimates of incidence, morbidity, the value of lost work time and survivorship. Our choice of a ten-year time horizon is novel to the field of rare disease and was determined to be meaningful after consultations with present co-authors, including five physicians, one of whom is a transplant surgeon. A total of 44 450 cGvHD patients in the United States were estimated to require treatment over the next decade (from 2015 to 2025). This estimate is based on the last 5 years of trends reported in the transplant registries. What is not reported in any registry is that these patients will accrue a total of 605 631 years of lost wages, a collective lost productivity that will cost society over $27 Billion in the decade ahead: more than five times ($27B vs. $5.2B) the estimated ten-year cost of treating the condition.

Opioid-related Policies in New England Emergency Departments.

OBJECTIVES: The opioid abuse and overdose epidemic in the United States has led to the need for new practice policies to guide clinicians. We describe implementation of opioid-related policies in emergency departments (EDs) in New England to gauge progress and determine where further work is needed. METHODS: This study analyzed data from the 2015 National Emergency Department Inventory-New England survey. The survey queried directors of every ED (n = 195) in the six New England states to determine the implementation of five specific policies related to opioid management. ED characteristics (e.g., annual visits, location, and admission rates) were also obtained and a multivariable analysis was conducted to identify ED characteristics independently associated with the number of opioid-related policies implemented. RESULTS: Overall, 169 EDs (87%) responded, with a >80% response rate in each state. Implementation of opioid-related policies varied as follows: 1) use of a screening tool for patients with suspected prescription opioid abuse potential (n = 30, 18%), 2) access state prescription drug monitoring program (PDMP) before prescribing opioids (n = 132, 78%), 3) notify the primary opioid prescriber when prescribing opioids for ED patients with chronic pain (n = 69, 41%), 4) refer patients with opioid abuse to recovery resources (n = 117, 70%), and 5) prescribe naloxone to patients at risk of opioid overdose after ED discharge (n = 19, 12%). EDs located in metropolitan areas and with at least one attending physician on duty 24/7 were less likely to implement opioid policies (incident rate ratio [IRR] = 0.65, 95% confidence interval [CI] = 0.48-0.89; and IRR = 0.78, 95% CI = 0.6-1.0, respectively) while EDs with ≥15% hospitalization rate that used electronic computerized medication ordering and those in Rhode Island were more likely to implement opioid policies (IRR = 1.23, 95% CI = 1.03-1.48; IRR = 1.95, 95% CI = 1.19-3.22; and IRR = 1.30, 95% CI = 1.08-1.56, respectively). CONCLUSIONS: The implementation of opioid-related policies varies among New England EDs. The presence of policies recommending use of screening tools and prescribing naloxone for at-risk patients was low, whereas those regarding utilization of the PDMP and referral of patients with opioid abuse to recovery resources were more common. These data provide important benchmarks for future evaluations and recommendations.

Prevention of treatment-related fluid overload reduces estimated effective cost of prothrombin complex concentrate in patients requiring rapid vitamin K antagonist reversal.

INTRODUCTION: Fresh frozen plasma (FFP) is a frequently used human blood product to reverse the effects of vitamin K antagonists. While FFP is relatively economical, its large fluid volume can lead to hospitalization complications, therefore increasing the overall cost of use. MATERIALS & METHODS: A recently published article by Sarode et al., in Circulation, described the rate of volume overload associated with FFP use for reversal of vitamin K antagonists. This condition, described as transfusion-associated circulatory overload, has a defined rate of intensive care admission, which also has a well-reported average cost. The additional monetary value of intensive care unit admission and caring for fluid overload is then compared to the cost of another product, four-factor prothrombin complex concentrates, which does not, as per the Sarode paper, result in fluid overload. RESULTS: The increased costs attributed to FFP-associated fluid overload for vitamin K antagonist reversal partly defrays the increased upfront cost of four-factor prothrombin complex concentrates. DISCUSSION: FFP is commonly used to acutely reverse the effects of vitamin K antagonists. However, its use requires significant time for infusion, may lead to fluid overload, and is not fully effective in compete anticoagulation reversal. One alternative therapy for anticoagulant reversal is use of prothrombin complex concentrates, which are rapidly infused, are not associated with fluid overload, and are effective in complete reversal of coagulation measurements. This should be considered for patients with acute bleeding emergencies.