Chest radiography of acute paediatric lower respiratory infections: experience versus interobserver variation

Pediatric chest radiograph interpretation in a real-life setting

Chest radiography is a frequently used imaging modality in children. However, only fair to moderate inter-observer agreement has been reported between chest radiograph interpreters. Most studies were not performed in real-world clinical settings. Our aims were to examine the agreement between emergency department pediatricians and board-certified radiologists in a pediatric real-life setting and to identify clinical risk factors for the discrepancies. Included were children aged 3 months to 18 years who underwent chest radiography in the emergency department not during the regular hours of radiologist interpretation. Every case was reviewed by an expert panel. Inter-observer agreement between emergency department pediatricians and board-certified radiologists was assessed by Cohen's kappa; risk factors for disagreement were analyzed. Among 1373 cases, the level of agreement between emergency department pediatricians and board-certified radiologists was "moderate" (k = 0.505). For radiographs performed after midnight, agreement was only "fair" (k = 0.391). The expert panel identified clinically relevant disagreements in 260 (18.9%) of the radiographs. Over-treatment of antibiotics was identified in 121 (8.9%) of the cases and under-treatment in 79 (5.8%). In a multivariable logistic regression, the following parameters were found to be significantly associated with disagreements: neurological background (p = 0.046), fever (p = 0.001), dyspnea (p = 0.014), and radiographs performed after midnight (p = 0.007).

CONCLUSIONS

Moderate agreement was found between emergency department pediatricians and board-certified radiologists in interpreting chest radiographs. Neurological background, fever, dyspnea, and radiographs performed after midnight were identified as risk factors for disagreement. Implementing these findings could facilitate the use of radiologist expertise, save time and resources, and potentially improve patient care.

WHAT IS KNOWN

• Only fair to moderate inter-observer agreement has been reported between chest radiograph interpreters. • Most studies were not performed in real-world clinical settings. Clinical risk factors for disagreements have not been reported.

WHAT IS NEW

• In this study, which included 1373 cases at the emergency department, the level of agreement between interpreters was only "moderate." • The major clinical parameters associated with interpretation discrepancies were neurological background, fever, dyspnea, and interpretations conducted during the night shift.

Nonradiology Health Care Professionals Significantly Benefit From AI Assistance in Emergency-Related Chest Radiography Interpretation

BACKGROUND

Chest radiographs (CXRs) are still of crucial importance in primary diagnostics, but their interpretation poses difficulties at times.

RESEARCH QUESTION

Can a convolutional neural network-based artificial intelligence (AI) system that interprets CXRs add value in an emergency unit setting?

STUDY DESIGN AND METHODS

A total of 563 CXRs acquired in the emergency unit of a major university hospital were retrospectively assessed twice by three board-certified radiologists, three radiology residents, and three emergency unit-experienced nonradiology residents (NRRs). They used a two-step reading process: (1) without AI support; and (2) with AI support providing additional images with AI overlays. Suspicion of four suspected pathologies (pleural effusion, pneumothorax, consolidations suspicious for pneumonia, and nodules) was reported on a five-point confidence scale. Confidence scores of the board-certified radiologists were converted into four binary reference standards of different sensitivities. Performance by radiology residents and NRRs without AI support/with AI support were statistically compared by using receiver-operating characteristics (ROCs), Youden statistics, and operating point metrics derived from fitted ROC curves.

RESULTS

NRRs could significantly improve performance, sensitivity, and accuracy with AI support in all four pathologies tested. In the most sensitive reference standard (reference standard IV), NRR consensus improved the area under the ROC curve (mean, 95% CI) in the detection of the time-critical pathology pneumothorax from 0.846 (0.785-0.907) without AI support to 0.974 (0.947-1.000) with AI support (P < .001), which represented a gain of 30% in sensitivity and 2% in accuracy (while maintaining an optimized specificity). The most pronounced effect was observed in nodule detection, with NRR with AI support improving sensitivity by 53% and accuracy by 7% (area under the ROC curve without AI support, 0.723 [0.661-0.785]; with AI support, 0.890 [0.848-0.931]; P < .001). Radiology residents had smaller, mostly nonsignificant gains in performance, sensitivity, and accuracy with AI support.

INTERPRETATION

We found that in an emergency unit setting without 24/7 radiology coverage, the presented AI solution features an excellent clinical support tool to nonradiologists, similar to a second reader, and allows for a more accurate primary diagnosis and thus earlier therapy initiation.

Lung ultrasound: A potential tool in the diagnosis of ventilator-associated pneumonia in pediatric intensive care units

PURPOSE

Ventilator-associated pneumonia (VAP) is a common healthcare-associated infection in pediatric intensive care unit (PICU), increasing mortality, antibiotics use and duration of ventilation and hospitalization. VAP diagnosis is based on clinical and chest X-ray (CXR) signs defined by the 2018 Center for Disease Control (gold standard). However, CXR induces repetitive patients' irradiation and technical limitations. This study aimed to investigate if lung ultrasound (LUS) can substitute CXR in the VAP diagnosis.

METHODS

A monocentric and prospective study was conducted in a French tertiary care hospital. Patients under 18-year-old admitted to PICU between November 2018 and July 2020 with invasive mechanical ventilation for more than 48 h were included. The studied LUS signs were consolidations, dynamic air bronchogram, subpleural consolidations (SPC), B-lines, and pleural effusion. The diagnostic values of each sign associated with clinical signs (cCDC) were compared to the gold standard approach. LUS, chest X-ray, and clinical score were performed daily.

RESULTS

Fifty-seven patients were included. The median age was 8 [3-34] months. Nineteen (33%) children developed a VAP. In patients with VAP, B-Lines, and consolidations were highly frequent (100 and 68.8%) and, associated with cCDC, were highly sensitive (100 [79-100] % and 88 [62-98] %, respectively) and specific (95.5 [92-98] % and 98 [95-99] %, respectively). Other studied signs, including SPC, showed high specificity (>97%) but low sensibility (<50%).

CONCLUSION

LUS seems to be a powerful tool for VAP diagnosis in children with a clinical suspicion, efficiently substituting CXR, and limiting children's exposure to ionizing radiations.

Comparison of lung ultrasound and chest radiography for detecting pneumonia in children: a systematic review and meta-analysis

BACKGROUND

Lung ultrasound (LUS) is recommended as a reliable diagnostic alternative to chest X-ray (CXR) for detecting pneumonia in children.

METHODS

PubMed, Embase, and Cochrane Library databases were used to identify eligible studies from their inception until April 2023. The investigated diagnostic parameters included sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), and area under the receiver operating characteristic curves (AUC).

RESULTS

Twenty-six studies involving 3,401 children were selected for meta-analysis. The sensitivity, specificity, PLR, NLR, DOR, and AUC of LUS for detecting pneumonia in children were 0.95, 0.92, 12.31, 0.05, 108.53, and 0.98, respectively, while the sensitivity, specificity, PLR, NLR, DOR, and AUC of CXR were 0.92, 0.93, 24.63, 0.08, 488.54, and 0.99, respectively. The sensitivity of LUS was higher than that of CXR for detecting pneumonia in children (ratio: 1.03; 95% CI: 1.01-1.06; P = 0.018), whereas the DOR of LUS was significantly lower than that of CXR (ratio: 0.22; 95% CI: 0.06-0.85; P = 0.028).

CONCLUSIONS

This study found that the diagnostic performance of LUS was comparable to that of CXR for detecting pneumonia, and the sensitivity of LUS was superior to that of CXR.

Modern approaches at the management of children with community-acquired pneumonia

Experts of The Union of Pediatricians of Russia have developed current clinical guidelines for management of children with community-acquired pneumonia, which were approved by the Scientific and Practice Council of Ministry of Public Health of the Russian Federation in January 2022. Particular attention is paid to the etiological structure, modern classification, diagnostic tests and flagship approaches to antibacterial therapy of community-acquired pneumonia in children based on the principles of evidentiary medicine. 

Using Machine Learning Methods Incorporating Individual Reader Annotations to Classify Paediatric Chest Radiographs in Epidemiological Studies

Introduction: Epidemiological studies that involve interpretation of chest radiographs (CXRs) suffer from inter-reader and intra-reader variability. Inter-reader and intra-reader variability hinder comparison of results from different studies or centres, which negatively affects efforts to track the burden of chest diseases or evaluate the efficacy of interventions such as vaccines. This study explores machine learning models that could standardize interpretation of CXR across studies and the utility of incorporating individual reader annotations when training models using CXR data sets annotated by multiple readers. Methods: Convolutional neural networks were used to classify CXRs from seven low to middle-income countries into five categories according to the World Health Organization's standardized methodology for interpreting paediatric CXRs. We compared models trained to predict the final/aggregate classification with models trained to predict how each reader would classify an image and then aggregate predictions for all readers using unweighted mean. Results: Incorporating individual reader's annotations during model training improved classification accuracy by 3.4% (multi-class accuracy 61% vs 59%). Model accuracy was higher for children above 12 months of age (68% vs 58%). The accuracy of the models in different countries ranged between 45% and 71%. Conclusions: Machine learning models can annotate CXRs in epidemiological studies reducing inter-reader and intra-reader variability. In addition, incorporating individual reader annotations can improve the performance of machine learning models trained using CXRs annotated by multiple readers.

Clinically focused multi-cohort benchmarking as a tool for external validation of artificial intelligence algorithm performance in basic chest radiography analysis

Artificial intelligence (AI) algorithms evaluating [supine] chest radiographs ([S]CXRs) have remarkably increased in number recently. Since training and validation are often performed on subsets of the same overall dataset, external validation is mandatory to reproduce results and reveal potential training errors. We applied a multicohort benchmarking to the publicly accessible (S)CXR analyzing AI algorithm CheXNet, comprising three clinically relevant study cohorts which differ in patient positioning ([S]CXRs), the applied reference standards (CT-/[S]CXR-based) and the possibility to also compare algorithm classification with different medical experts’ reading performance. The study cohorts include [1] a cohort, characterized by 563 CXRs acquired in the emergency unit that were evaluated by 9 readers (radiologists and non-radiologists) in terms of 4 common pathologies, [2] a collection of 6,248 SCXRs annotated by radiologists in terms of pneumothorax presence, its size and presence of inserted thoracic tube material which allowed for subgroup and confounding bias analysis and [3] a cohort consisting of 166 patients with SCXRs that were evaluated by radiologists for underlying causes of basal lung opacities, all of those cases having been correlated to a timely acquired computed tomography scan (SCXR and CT within < 90 min). CheXNet non-significantly exceeded the radiology resident (RR) consensus in the detection of suspicious lung nodules (cohort [1], AUC AI/RR: 0.851/0.839, p = 0.793) and the radiological readers in the detection of basal pneumonia (cohort [3], AUC AI/reader consensus: 0.825/0.782, p = 0.390) and basal pleural effusion (cohort [3], AUC AI/reader consensus: 0.762/0.710, p = 0.336) in SCXR, partly with AUC values higher than originally published (“Nodule”: 0.780, “Infiltration”: 0.735, “Effusion”: 0.864). The classifier “Infiltration” turned out to be very dependent on patient positioning (best in CXR, worst in SCXR). The pneumothorax SCXR cohort [2] revealed poor algorithm performance in CXRs without inserted thoracic material and in the detection of small pneumothoraces, which can be explained by a known systematic confounding error in the algorithm training process. The benefit of clinically relevant external validation is demonstrated by the differences in algorithm performance as compared to the original publication. Our multi-cohort benchmarking finally enables the consideration of confounders, different reference standards and patient positioning as well as the AI performance comparison with differentially qualified medical readers.

Artificial Intelligence in Chest Radiography Reporting Accuracy: Added Clinical Value in the Emergency Unit Setting Without 24/7 Radiology Coverage

OBJECTIVES

Chest radiographs (CXRs) are commonly performed in emergency units (EUs), but the interpretation requires radiology experience. We developed an artificial intelligence (AI) system (precommercial) that aims to mimic board-certified radiologists' (BCRs') performance and can therefore support non-radiology residents (NRRs) in clinical settings lacking 24/7 radiology coverage. We validated by quantifying the clinical value of our AI system for radiology residents (RRs) and EU-experienced NRRs in a clinically representative EU setting.

MATERIALS AND METHODS

A total of 563 EU CXRs were retrospectively assessed by 3 BCRs, 3 RRs, and 3 EU-experienced NRRs. Suspected pathologies (pleural effusion, pneumothorax, consolidations suspicious for pneumonia, lung lesions) were reported on a 5-step confidence scale (sum of 20,268 reported pathology suspicions [563 images × 9 readers × 4 pathologies]) separately by every involved reader. Board-certified radiologists' confidence scores were converted into 4 binary reference standards (RFSs) of different sensitivities. The RRs' and NRRs' performances were statistically compared with our AI system (trained on nonpublic data from different clinical sites) based on receiver operating characteristics (ROCs) and operating point metrics approximated to the maximum sum of sensitivity and specificity (Youden statistics).

RESULTS

The NRRs lose diagnostic accuracy to RRs with increasingly sensitive BCRs' RFSs for all considered pathologies. Based on our external validation data set, the AI system/NRRs' consensus mimicked the most sensitive BCRs' RFSs with areas under ROC of 0.940/0.837 (pneumothorax), 0.953/0.823 (pleural effusion), and 0.883/0.747 (lung lesions), which were comparable to experienced RRs and significantly overcomes EU-experienced NRRs' diagnostic performance. For consolidation detection, the AI system performed on the NRRs' consensus level (and overcomes each individual NRR) with an area under ROC of 0.847 referenced to the BCRs' most sensitive RFS.

CONCLUSIONS

Our AI system matched RRs' performance, meanwhile significantly outperformed NRRs' diagnostic accuracy for most of considered CXR pathologies (pneumothorax, pleural effusion, and lung lesions) and therefore might serve as clinical decision support for NRRs.

Using Machine Learning Methods Incorporating Individual Reader Annotations to Classify Paediatric Chest Radiographs in Epidemiological Studies

Introduction: Epidemiological studies that involve interpretation of chest radiographs (CXRs) suffer from inter-reader and intra-reader variability. Inter-reader and intra-reader variability hinder comparison of results from different studies or centres, which negatively affects efforts to track the burden of chest diseases or evaluate the efficacy of interventions such as vaccines. This study explores machine learning models that could standardize interpretation of CXR across studies and the utility of incorporating individual reader annotations when training models using CXR data sets annotated by multiple readers. Methods: Convolutional neural networks were used to classify CXRs from seven low to middle-income countries into five categories according to the World Health Organization's standardized methodology for interpreting paediatric CXRs. We compared models trained to predict the final/aggregate classification with models trained to predict how each reader would classify an image and then aggregate predictions for all readers using unweighted mean. Results: Incorporating individual reader's annotations during model training improved classification accuracy by 3.4% (multi-class accuracy 61% vs 59%). Model accuracy was higher for children above 12 months of age (68% vs 58%). The accuracy of the models in different countries ranged between 45% and 71%. Conclusions: Machine learning models can annotate CXRs in epidemiological studies reducing inter-reader and intra-reader variability. In addition, incorporating individual reader annotations can improve the performance of machine learning models trained using CXRs annotated by multiple readers.

Interobserver agreement in interpretation of chest radiographs for pediatric community acquired pneumonia: Findings of the pedCAPNETZ-cohort

Although chest radiograph (CXR) is commonly used in diagnosing pediatric community acquired pneumonia (pCAP), limited data on interobserver agreement among radiologists exist. PedCAPNETZ is a prospective, observational, and multicenter study on pCAP. N = 233 CXR from patients with clinical diagnosis of pCAP were retrieved and n = 12 CXR without pathological findings were added. All CXR were interpreted by a radiologist at the site of recruitment and by two external, blinded pediatric radiologists. To evaluate interobserver agreement, the reporting of presence or absence of pCAP in CXR was analyzed, and prevalence and bias-adjusted kappa (PABAK) statistical testing was applied. Overall, n = 190 (82%) of CXR were confirmed as pCAP by two external pediatric radiologists. Compared with patients with pCAP negative CXR, patients with CXR-confirmed pCAP displayed higher C-reactive protein levels and a longer duration of symptoms before enrollment (p < .007). Further parameters, that is, age, respiratory rate, and oxygen saturation showed no significant difference. The interobserver agreement between the onsite radiologists and each of the two independent pediatric radiologists for the presence of pCAP was poor to fair (69%; PABAK = 0.39% and 76%; PABAK = 0.53, respectively). The concordance between the external radiologists was fair (81%; PABAK = 0.62). With regard to typical CXR findings for pCAP, chance corrected interrater agreement was highest for pleural effusions, infiltrates, and consolidations and lowest for interstitial patterns and peribronchial thickening. Our data show a poor interobserver agreement in the CXR-based diagnosis of pCAP and emphasized the need for harmonized interpretation standards.

Image-based deep learning in diagnosing the etiology of pneumonia on pediatric chest X-rays

PURPOSE

Comparing the efficacy of a deep-learning model in classifying the etiology of pneumonia on pediatric chest X-rays (CXRs) with that of human readers.

METHODS

We built a clinical-pediatric CXR set containing 4035 patients to exploit a deep-learning model called Resnet-50 for differentiating viral from bacterial pneumonia. The dataset was split into training (80%) and validation (20%). Model performance was assessed by receiver operating characteristic curve and area under the curve (AUC) on the first test set of 400 CXRs collected from different studies. For the second test set composed of 100 independent examinations obtained from the daily clinical practice at our institution, the kappa coefficient was selected to measure the interrater agreement in a pairwise fashion for the reference standard, all reviewers, and the model. Gradient-weighted class activation mapping was used to visualize the significant areas contributing to the model prediction.

RESULTS

On the first test set, the best-performing classifier achieved an AUC of 0.919 (p < .001), with a sensitivity of 79.0% and specificity of 88.9%. On the second test set, the classifier achieved performance similar to that of human experts, which resulted in a sensitivity of 74.3% and specificity of 90.8%, positive and negative likelihood ratios of 8.1 and 0.3, respectively. Contingence tables and kappa values further revealed that expert reviewers and model reached substantial agreements on differentiating the etiology of pediatric pneumonia.

CONCLUSIONS

This study demonstrated that the model performed similarly as human reviewers and recognized the regions of pathology on CXRs.

National Current Care Guidelines for paediatric lower respiratory tract infections reduced the use of chest radiographs but local variations were observed

AIM

Our aim was to evaluate the impact of the 2014 Finnish Current Care Guidelines for paediatric lower respiratory tract infections (LRTIs), particularly on taking of chest radiographs.

METHODS

This study used official national data and regional (Pirkanmaa) data on children aged 0-16 years who underwent chest radiographs in 2011 and 2015. We also collected data for LRTI diagnoses from local registers, including prescribed antibiotics and taking of chest radiographs. The local cohort comprised children aged 0-15 who presented to the primary care emergency room or to the hospital emergency department (Tampere university hospital) in November-December 2012-2015.

RESULTS

Chest radiographs for Finnish children aged 0-16 fell from 2011 to 2015: by 15.9% nationally and by 16.9% in Pirkanmaa. When asylum seekers with chest radiographs for tuberculosis screening were excluded, the estimated national reduction was 29.9%. In the local cohort, chest radiographs increased from 82 to 139 (69.5%) between 2012/2013 and 2014/2015 as the occurrence of community-acquired pneumonia (CAP) increased. However, the proportion of patients with CAP who had chest radiograph taken tended to decrease from 84.6% to 71.3% (p = 0.078).

CONCLUSION

Decreases in national and regional chest imaging trends were observed after the 2014 guidance for children`s LRTI was introduced.

Interpretation of pediatric chest radiographs by non-radiologist clinicians in Botswana using World Health Organization criteria for endpoint pneumonia

BACKGROUND

In low- and middle-income countries, chest radiographs are most frequently interpreted by non-radiologist clinicians.

OBJECTIVE

We examined the reliability of chest radiograph interpretations performed by non-radiologist clinicians in Botswana and conducted an educational intervention aimed at improving chest radiograph interpretation accuracy among non-radiologist clinicians.

MATERIALS AND METHODS

We recruited non-radiologist clinicians at a referral hospital in Gaborone, Botswana, to interpret de-identified chest radiographs for children with clinical pneumonia. We compared their interpretations with those of two board-certified pediatric radiologists in the United States. We evaluated associations between level of medical training and the accuracy of chest radiograph findings between groups, using logistic regression and kappa statistics. We then developed an in-person training intervention led by a pediatric radiologist. We asked participants to interpret 20 radiographs before and immediately after the intervention, and we compared their responses to those of the facilitating radiologist. For both objectives, our primary outcome was the identification of primary endpoint pneumonia, defined by the World Health Organization as presence of endpoint consolidation or endpoint effusion.

RESULTS

Twenty-two clinicians interpreted chest radiographs in the primary objective; there were no significant associations between level of training and correct identification of endpoint pneumonia; concordance between respondents and radiologists was moderate (κ=0.43). After the training intervention, participants improved agreement with the facilitating radiologist for endpoint pneumonia from fair to moderate (κ=0.34 to κ=0.49).

CONCLUSION

Non-radiologist clinicians in Botswana do not consistently identify key chest radiographic findings of pneumonia. A targeted training intervention might improve non-radiologist clinicians' ability to interpret chest radiographs.

[The chest x-ray in acute bronchiolitis: Technical quality, findings, and an assessment of its reliability]

INTRODUCTION

Despite the recommendations of the current Clinical Practice Guidelines, the chest x-ray continues to be a widely used diagnostic test in the assessment of infants with acute bronchiolitis (AB). However, there have not been many studies that have assessed its reproducibility in these patients. In the present study, an evaluation is made on the radiographs, describing their quality, their radiological findings, and provides new evidence on the agreement between observers.

METHOD

Out of a total of 281 infants admitted due to acute bronchiolitis, 140 chest x-rays were performed. Twelve doctors from different specialities evaluated the presence or absence of 10 radiological signs previously agreed by consensus. The level of agreement between 2 observers, and in groups of 3 or more, were analysed using the Cohen and Fleiss kappa index, respectively.

RESULTS

Only 8.5% of the radiographs showed evidence of a complicated AB. The between-observer agreement in groups of 3 or more was medium, and with little variability (kappa: 0.20-0.40). However, between 2 observers, each observer against radiologist, the variability was wider, (kappa: -0.20-0.60). This level of agreement was associated with factors including, the sign to evaluate, the medical specialty, and level of professional experience.

CONCLUSION

The low levels of agreement between observers and the wide variability, makes the chest x-ray an unreliable diagnostic tool, and is not recommended for the assessment of infants with AB.

Lung ultrasound in diagnosing pneumonia in childhood: a systematic review and meta-analysis

PURPOSE

Pneumonia is the third leading cause of death in children under 5 years of age worldwide. In pediatrics, both the accuracy and safety of diagnostic tools are important. Lung ultrasound (LUS) could be a safe diagnostic tool for this reason. We searched in the literature for diagnostic studies about LUS to predict pneumonia in pediatric patients using systematic review and meta-analysis.

METHODS

The Medline, CINAHL, Cochrane Library, Embase, SPORTDiscus, ScienceDirect, and Web of Science databases from inception to September 2017 were searched. All studies that evaluated the diagnostic accuracy of LUS in determining the presence of pneumonia in patients under 18 years of age were included.

RESULTS

1042 articles were found by systematic search. 76 articles were assessed for eligibility. Seventeen studies were included in the systematic review. We included 2612 pooled cases. The age of the pooled sample population ranged from 0 to about 21 years old. Summary sensitivity, specificity, and AUC were 0.94 (IQR: 0.89-0.97), 0.93 (IQR: 0.86-0.98), and 0.98 (IQR: 0.94-0.99), respectively. No agreement on reference standard was detected: nine studies used chest X-rays, while four studies considered the clinical diagnosis. Only one study used computed tomography.

CONCLUSIONS

LUS seems to be a promise tool for diagnosing pneumonia in children. However, the high heterogeneity found across the individual studies, and the absence of a reliable reference standard, make the finding questionable. More methodologically rigorous studies are needed.

Standardized Interpretation of Chest Radiographs in Cases of Pediatric Pneumonia From the PERCH Study

Background.

Chest radiographs (CXRs) are a valuable diagnostic tool in epidemiologic studies of pneumonia. The World Health Organization (WHO) methodology for the interpretation of pediatric CXRs has not been evaluated beyond its intended application as an endpoint measure for bacterial vaccine trials.

Methods.

The Pneumonia Etiology Research for Child Health (PERCH) study enrolled children aged 1–59 months hospitalized with WHO-defined severe and very severe pneumonia from 7 low- and middle-income countries. An interpretation process categorized each CXR into 1 of 5 conclusions: consolidation, other infiltrate, both consolidation and other infiltrate, normal, or uninterpretable. Two members of a 14-person reading panel, who had undertaken training and standardization in CXR interpretation, interpreted each CXR. Two members of an arbitration panel provided additional independent reviews of CXRs with discordant interpretations at the primary reading, blinded to previous reports. Further discordance was resolved with consensus discussion.

Results.

A total of 4172 CXRs were obtained from 4232 cases. Observed agreement for detecting consolidation (with or without other infiltrate) between primary readers was 78% (κ = 0.50) and between arbitrators was 84% (κ = 0.61); agreement for primary readers and arbitrators across 5 conclusion categories was 43.5% (κ = 0.25) and 48.5% (κ = 0.32), respectively. Disagreement was most frequent between conclusions of other infiltrate and normal for both the reading panel and the arbitration panel (32% and 30% of discordant CXRs, respectively).

Conclusions.

Agreement was similar to that of previous evaluations using the WHO methodology for detecting consolidation, but poor for other infiltrates despite attempts at a rigorous standardization process.

Interventions to improve antimicrobial prescribing of doctors in training (IMPACT): a realist review

Background

Interventions to improve the antimicrobial prescribing practices of doctors have been implemented widely to curtail the emergence and spread of antimicrobial resistance, but have been met with varying levels of success.

Objectives

This study aimed to generate an in-depth understanding of how antimicrobial prescribing interventions ‘work’ (or do not work) for doctors in training by taking into account the wider context in which prescribing decisions are enacted.

Design

The review followed a realist approach to evidence synthesis, which uses an interpretive, theory-driven analysis of qualitative, quantitative and mixed-methods data from relevant studies.

Setting

Primary and secondary care.

Participants

Not applicable.

Interventions

Studies related to antimicrobial prescribing for doctors in training.

Main outcome measures

Not applicable.

Data sources

EMBASE (via Ovid), MEDLINE (via Ovid), MEDLINE In-Process & Other Non-Indexed Citations (via Ovid), PsycINFO (via Ovid), Web of Science core collection limited to Science Citation Index Expanded (SCIE) and Conference Proceedings Citation Index – Science (CPCI-S) (via Thomson Reuters), Cochrane Central Register of Controlled Trials (CENTRAL), Cochrane Database of Systematic Reviews, the Health Technology Assessment (HTA) database (all via The Cochrane Library), Applied Social Sciences Index and Abstracts (ASSIA) (via ProQuest), Google Scholar (Google Inc., Mountain View, CA, USA) and expert recommendations.

Review methods

Clearly bounded searches of electronic databases were supplemented by citation tracking and grey literature. Following quality standards for realist reviews, the retrieved articles were systematically screened and iteratively analysed to develop theoretically driven explanations. A programme theory was produced with input from a stakeholder group consisting of practitioners and patient representatives.

Results

A total of 131 articles were included. The overarching programme theory developed from the analysis of these articles explains how and why doctors in training decide to passively comply with or actively follow (1) seniors’ prescribing habits, (2) the way seniors take into account prescribing aids and seek the views of other health professionals and (3) the way seniors negotiate patient expectations. The programme theory also explains what drives willingness or reluctance to ask questions about antimicrobial prescribing or to challenge the decisions made by seniors. The review outlines how these outcomes result from complex inter-relationships between the contexts of practice doctors in training are embedded in (hierarchical relationships, powerful prescribing norms, unclear roles and responsibilities, implicit expectations about knowledge levels and application in practice) and the mechanisms triggered in these contexts (fear of criticism and individual responsibility, reputation management, position in the clinical team and appearing competent). Drawing on these findings, we set out explicit recommendations for optimal tailoring, design and implementation of antimicrobial prescribing interventions targeted at doctors in training.

Limitations

Most articles included in the review discussed hospital-based, rather than primary, care. In cases when few data were available to fully capture all the nuances between context, mechanisms and outcomes, we have been explicit about the strength of our arguments.

Conclusions

This review contributes to our understanding of how antimicrobial prescribing interventions for doctors in training can be better embedded in the hierarchical and interprofessional dynamics of different health-care settings.

Future work

More work is required to understand how interprofessional support for doctors in training can contribute to appropriate prescribing in the context of hierarchical dynamics.

Study registration

This study is registered as PROSPERO CRD42015017802.

Funding

The National Institute for Health Research Health Services and Delivery Research programme.

Lung ultrasound for the diagnosis of community-acquired pneumonia in children

Ultrasound (US) has been proposed as an alternative first-line imaging modality to diagnose community-acquired pneumonia in children. Lung US has the potential benefits over chest radiography of being radiation free, subject to fewer regulatory requirements, relatively lower cost and with immediate bedside availability of results. However, the uptake of lung US into clinical practice has been slow and it is not yet included in clinical guidelines for community-acquired pneumonia in children. The aim of this review is to give an overview of the equipment and techniques used to perform lung US in children with suspected pneumonia and the interpretation of relevant sonographic findings. We also summarise the current evidence of diagnostic accuracy and reliability of lung US compared to alternative imaging modalities in children and critically consider the strengths and limitations of lung US for use in children presenting with suspected community-acquired pneumonia.

Inter-observer and intra-observer reliability of mechanical axis alignment before and after total knee arthroplasty using long leg radiographs

BACKGROUND

Long leg radiographs (LLRs) are commonly performed for assessment of mechanical alignment and operative planning in patients undergoing total knee arthroplasty (TKA). The aim of this study was to determine the inter- and intra-observer reliability of alignment measured by observers of different levels of experience.

METHODS

Forty patients on the waiting list for a TKA had pre- and post-operative standardised LLRs. We analysed the measurements of mechanical axis alignment between an orthopaedic surgeon, a senior orthopaedic registrar, a junior orthopaedic registrar, and a medical student. Reviewers performed blinded measurements on the same computer screen. These measurements were repeated three months later to assess intra-observer reliability. Furthermore high-resolution screens were compared with standard hospital computer screens to investigate whether monitoring quality influenced the accuracy of measurements of alignment.

RESULTS

Inter-observer reliability was high for pre-operative LLRs with an intra-class correlation (ICC) of >0.9 at all experience levels. Post-operative ICC was lowest between the surgeon and the medical student at 0.7. Intra-observer reliability was high at all experience levels. Larger deformities appeared to have exaggerated measurements for both pre- and post-operative images. There appeared to be no effect of the monitor size and quality on the accuracy of measurement.

CONCLUSIONS

Long leg radiographs can be used to measure mechanical axis alignment with strong reliability across different levels of experience. This information is important for the evaluation of knee alignment measurements in current clinical practice, to assess severity of deformity and to accompany pre-operative planning and post-operative evaluation.

LEVEL OF EVIDENCE

III.

Lung ultrasound in the diagnosis of pneumonia in children: proposal for a new diagnostic algorithm

Background. Despite guideline recommendations, chest radiography (CR) for the diagnosis of community-acquired pneumonia (CAP) in children is commonly used also in mild and/or uncomplicated cases. The aim of this study is to assess the reliability of lung ultrasonography (LUS) as an alternative test in these cases and suggest a new diagnostic algorithm. Methods. We reviewed the medical records of all patients admitted to the pediatric ward from February 1, 2013 to December 31, 2014 with respiratory signs and symptoms. We selected only cases with mild/uncomplicated clinical course and in which CR and LUS were performed within 24 h of each other. The LUS was not part of the required exams recorded in medical records but performed independently. The discharge diagnosis, made only on the basis of history and physical examination, laboratory and instrumental tests, including CR (without LUS), was used as a reference test to compare CR and LUS findings. Results. Of 52 selected medical records CAP diagnosis was confirmed in 29 (55.7%). CR was positive in 25 cases, whereas LUS detected pneumonia in 28 cases. Four patients with negative CR were positive in ultrasound findings. Instead, one patient with negative LUS was positive in radiographic findings. The LUS sensitivity was 96.5% (95% CI [82.2%-99.9%]), specificity of 95.6% (95% CI [78.0%-99.9%]), positive likelihood ratio of 22.2 (95% CI [3.2-151.2]), and negative likelihood ratio of 0.04 (95% CI [0.01-0.25]) for diagnosing pneumonia. Conclusion. LUS can be considered as a valid alternative diagnostic tool of CAP in children and its use must be promoted as a first approach in accordance with our new diagnostic algorithm.

The radiological diagnosis of pneumonia in children

Despite the importance of paediatric pneumonia as a cause of short and long-term morbidity and mortality worldwide, a reliable gold standard for its diagnosis remains elusive. The utility of clinical, microbiological and radiological diagnostic approaches varies widely within and between populations and is heavily dependent on the expertise and resources available in various settings. Here we review the role of radiology in the diagnosis of paediatric pneumonia. Chest radiographs (CXRs) are the most widely employed test, however, they are not indicated in ambulatory settings, cannot distinguish between viral and bacterial infections and have a limited role in the ongoing management of disease. A standardised definition of alveolar pneumonia on a CXR exists for epidemiological studies targeting bacterial pneumonias but it should not be extrapolated to clinical settings. Radiography, computed tomography and to a lesser extent ultrasonography and magnetic resonance imaging play an important role in complicated pneumonias but there are limitations that preclude their use as routine diagnostic tools. Large population-based studies are needed in different populations to address many of the knowledge gaps in the radiological diagnosis of pneumonia in children, however, the feasibility of such studies is an important barrier.