Integrating deep learning with microfluidics for biophysical classification of sickle red blood cells adhered to laminin.

Sickle cell disease, a genetic disorder affecting a sizeable global demographic, manifests in sickle red blood cells (sRBCs) with altered shape and biomechanics. sRBCs show heightened adhesive interactions with inflamed endothelium, triggering painful vascular occlusion events. Numerous studies employ microfluidic-assay-based monitoring tools to quantify characteristics of adhered sRBCs from high resolution channel images. The current image analysis workflow relies on detailed morphological characterization and cell counting by a specially trained worker. This is time and labor intensive, and prone to user bias artifacts. Here we establish a morphology based classification scheme to identify two naturally arising sRBC subpopulations-deformable and non-deformable sRBCs-utilizing novel visual markers that link to underlying cell biomechanical properties and hold promise for clinically relevant insights. We then set up a standardized, reproducible, and fully automated image analysis workflow designed to carry out this classification. This relies on a two part deep neural network architecture that works in tandem for segmentation of channel images and classification of adhered cells into subtypes. Network training utilized an extensive data set of images generated by the SCD BioChip, a microfluidic assay which injects clinical whole blood samples into protein-functionalized microchannels, mimicking physiological conditions in the microvasculature. Here we carried out the assay with the sub-endothelial protein laminin. The machine learning approach segmented the resulting channel images with 99.1±0.3% mean IoU on the validation set across 5 k-folds, classified detected sRBCs with 96.0±0.3% mean accuracy on the validation set across 5 k-folds, and matched trained personnel in overall characterization of whole channel images with R2 = 0.992, 0.987 and 0.834 for total, deformable and non-deformable sRBC counts respectively. Average analysis time per channel image was also improved by two orders of magnitude (∼ 2 minutes vs ∼ 2-3 hours) over manual characterization. Finally, the network results show an order of magnitude less variance in counts on repeat trials than humans. This kind of standardization is a prerequisite for the viability of any diagnostic technology, making our system suitable for affordable and high throughput disease monitoring.

Mass spectrometry and machine learning for the accurate diagnosis of benzylpenicillin and multidrug resistance of Staphylococcus aureus in bovine mastitis.

Staphylococcus aureus is a serious human and animal pathogen threat exhibiting extraordinary capacity for acquiring new antibiotic resistance traits in the pathogen population worldwide. The development of fast, affordable and effective diagnostic solutions capable of discriminating between antibiotic-resistant and susceptible S. aureus strains would be of huge benefit for effective disease detection and treatment. Here we develop a diagnostics solution that uses Matrix-Assisted Laser Desorption/Ionisation-Time of Flight Mass Spectrometry (MALDI-TOF) and machine learning, to identify signature profiles of antibiotic resistance to either multidrug or benzylpenicillin in S. aureus isolates. Using ten different supervised learning techniques, we have analysed a set of 82 S. aureus isolates collected from 67 cows diagnosed with bovine mastitis across 24 farms. For the multidrug phenotyping analysis, LDA, linear SVM, RBF SVM, logistic regression, naïve Bayes, MLP neural network and QDA had Cohen's kappa values over 85.00%. For the benzylpenicillin phenotyping analysis, RBF SVM, MLP neural network, naïve Bayes, logistic regression, linear SVM, QDA, LDA, and random forests had Cohen's kappa values over 85.00%. For the benzylpenicillin the diagnostic systems achieved up to (mean result ± standard deviation over 30 runs on the test set): accuracy = 97.54% ± 1.91%, sensitivity = 99.93% ± 0.25%, specificity = 95.04% ± 3.83%, and Cohen's kappa = 95.04% ± 3.83%. Moreover, the diagnostic platform complemented by a protein-protein network and 3D structural protein information framework allowed the identification of five molecular determinants underlying the susceptible and resistant profiles. Four proteins were able to classify multidrug-resistant and susceptible strains with 96.81% ± 0.43% accuracy. Five proteins, including the previous four, were able to classify benzylpenicillin resistant and susceptible strains with 97.54% ± 1.91% accuracy. Our approach may open up new avenues for the development of a fast, affordable and effective day-to-day diagnostic solution, which would offer new opportunities for targeting resistant bacteria.

Deep-Learning System Detects Neoplasia in Patients With Barrett's Esophagus With Higher Accuracy Than Endoscopists in a Multistep Training and Validation Study With Benchmarking.

BACKGROUND & AIMS: We aimed to develop and validate a deep-learning computer-aided detection (CAD) system, suitable for use in real time in clinical practice, to improve endoscopic detection of early neoplasia in patients with Barrett's esophagus (BE). METHODS: We developed a hybrid ResNet-UNet model CAD system using 5 independent endoscopy data sets. We performed pretraining using 494,364 labeled endoscopic images collected from all intestinal segments. Then, we used 1704 unique esophageal high-resolution images of rigorously confirmed early-stage neoplasia in BE and nondysplastic BE, derived from 669 patients. System performance was assessed by using data sets 4 and 5. Data set 5 was also scored by 53 general endoscopists with a wide range of experience from 4 countries to benchmark CAD system performance. Coupled with histopathology findings, scoring of images that contained early-stage neoplasia in data sets 2-5 were delineated in detail for neoplasm position and extent by multiple experts whose evaluations served as the ground truth for segmentation. RESULTS: The CAD system classified images as containing neoplasms or nondysplastic BE with 89% accuracy, 90% sensitivity, and 88% specificity (data set 4, 80 patients and images). In data set 5 (80 patients and images) values for the CAD system vs those of the general endoscopists were 88% vs 73% accuracy, 93% vs 72% sensitivity, and 83% vs 74% specificity. The CAD system achieved higher accuracy than any of the individual 53 nonexpert endoscopists, with comparable delineation performance. CAD delineations of the area of neoplasm overlapped with those from the BE experts in all detected neoplasia in data sets 4 and 5. The CAD system identified the optimal site for biopsy of detected neoplasia in 97% and 92% of cases (data sets 4 and 5, respectively). CONCLUSIONS: We developed, validated, and benchmarked a deep-learning computer-aided system for primary detection of neoplasia in patients with BE. The system detected neoplasia with high accuracy and near-perfect delineation performance. The Netherlands National Trials Registry, Number: NTR7072.

A fully automated method for late ventricular diastole frame selection in post-dive echocardiography without ECG gating.

Venous gas emboli (VGE) are often quantified as a marker of decompression stress on echocardiograms. Bubble-counting has been proposed as an easy to learn method, but remains time-consuming, rendering large dataset analysis impractical. Computer automation of VGE counting following this method has therefore been suggested as a means to eliminate rater bias and save time. A necessary step for this automation relies on the selection of a frame during late ventricular diastole (LVD) for each cardiac cycle of the recording. Since electrocardiograms (ECG) are not always recorded in field experiments, here we propose a fully automated method for LVD frame selection based on regional intensity minimization. The algorithm is tested on 20 previously acquired echocardiography recordings (from the original bubble-counting publication), half of which were acquired at rest (Rest) and the other half after leg flexions (Flex). From the 7,140 frames analyzed, sensitivity was found to be 0.913 [95% CI: 0.875-0.940] and specificity 0.997 [95% CI: 0.996-0.998]. The method's performance is also compared to that of random chance selection and found to perform significantly better (p≺0.0001). No trend in algorithm performance was found with respect to VGE counts, and no significant difference was found between Flex and Rest (p>0.05). In conclusion, full automation of LVD frame selection for the purpose of bubble counting in post-dive echocardiography has been established with excellent accuracy, although we caution that high quality acquisitions remain paramount in retaining high reliability.

Deep learning shows the capability of high-level computer-aided diagnosis in malignant lymphoma.

A pathological evaluation is one of the most important methods for the diagnosis of malignant lymphoma. A standardized diagnosis is occasionally difficult to achieve even by experienced hematopathologists. Therefore, established procedures including a computer-aided diagnosis are desired. This study aims to classify histopathological images of malignant lymphomas through deep learning, which is a computer algorithm and type of artificial intelligence (AI) technology. We prepared hematoxylin and eosin (H&E) slides of a lesion area from 388 sections, namely, 259 with diffuse large B-cell lymphoma, 89 with follicular lymphoma, and 40 with reactive lymphoid hyperplasia, and created whole slide images (WSIs) using a whole slide system. WSI was annotated in the lesion area by experienced hematopathologists. Image patches were cropped from the WSI to train and evaluate the classifiers. Image patches at magnifications of ×5, ×20, and ×40 were randomly divided into a test set and a training and evaluation set. The classifier was assessed using the test set through a cross-validation after training. The classifier achieved the highest levels of accuracy of 94.0%, 93.0%, and 92.0% for image patches with magnifications of ×5, ×20, and ×40, respectively, in comparison to diffuse large B-cell lymphoma, follicular lymphoma, and reactive lymphoid hyperplasia. Comparing the diagnostic accuracies between the proposed classifier and seven pathologists, including experienced hematopathologists, using the test set made up of image patches with magnifications of ×5, ×20, and ×40, the best accuracy demonstrated by the classifier was 97.0%, whereas the average accuracy achieved by the pathologists using WSIs was 76.0%, with the highest accuracy reaching 83.3%. In conclusion, the neural classifier can outperform pathologists in a morphological evaluation. These results suggest that the AI system can potentially support the diagnosis of malignant lymphoma.

The Impact of Artificial Intelligence in the Endoscopic Assessment of Premalignant and Malignant Esophageal Lesions: Present and Future.

In the gastroenterology field, the impact of artificial intelligence was investigated for the purposes of diagnostics, risk stratification of patients, improvement in quality of endoscopic procedures and early detection of neoplastic diseases, implementation of the best treatment strategy, and optimization of patient prognosis. Computer-assisted diagnostic systems to evaluate upper endoscopy images have recently emerged as a supporting tool in endoscopy due to the risks of misdiagnosis related to standard endoscopy and different expertise levels of endoscopists, time-consuming procedures, lack of availability of advanced procedures, increasing workloads, and development of endoscopic mass screening programs. Recent research has tended toward computerized, automatic, and real-time detection of lesions, which are approaches that offer utility in daily practice. Despite promising results, certain studies might overexaggerate the diagnostic accuracy of artificial systems, and several limitations remain to be overcome in the future. Therefore, additional multicenter randomized trials and the development of existent database platforms are needed to certify clinical implementation. This paper presents an overview of the literature and the current knowledge of the usefulness of different types of machine learning systems in the assessment of premalignant and malignant esophageal lesions via conventional and advanced endoscopic procedures. This study makes a presentation of the artificial intelligence terminology and refers also to the most prominent recent research on computer-assisted diagnosis of neoplasia on Barrett's esophagus and early esophageal squamous cell carcinoma, and prediction of invasion depth in esophageal neoplasms. Furthermore, this review highlights the main directions of future doctor-computer collaborations in which machines are expected to improve the quality of medical action and routine clinical workflow, thus reducing the burden on physicians.

Impact of pre-analytical variables on deep learning accuracy in histopathology.

AIMS: Machine learning (ML) binary classification in diagnostic histopathology is an area of intense investigation. Several assumptions, including training image quality/format and the number of training images required, appear to be similar in many studies irrespective of the paucity of supporting evidence. We empirically compared training image file type, training set size, and two common convolutional neural networks (CNNs) using transfer learning (ResNet50 and SqueezeNet). METHODS AND RESULTS: Thirty haematoxylin and eosin (H&E)-stained slides with carcinoma or normal tissue from three tissue types (breast, colon, and prostate) were photographed, generating 3000 partially overlapping images (1000 per tissue type). These lossless Portable Networks Graphics (PNGs) images were converted to lossy Joint Photographic Experts Group (JPG) images. Tissue type-specific binary classification ML models were developed by the use of all PNG or JPG images, and repeated with a subset of 500, 200, 100, 50, 30 and 10 images. Eleven models were generated for each tissue type, at each quantity of training images, for each file type, and for each CNN, resulting in 924 models. Internal accuracies and generalisation accuracies were compared. There was no meaningful significant difference in accuracies between PNG and JPG models. Models trained with more images did not invariably perform better. ResNet50 typically outperformed SqueezeNet. Models were generalisable within a tissue type but not across tissue types. CONCLUSIONS: Lossy JPG images were not inferior to lossless PNG images in our models. Large numbers of unique H&E-stained slides were not required for training optimal ML models. This reinforces the need for an evidence-based approach to best practices for histopathological ML.

PheValuator: Development and evaluation of a phenotype algorithm evaluator.

BACKGROUND: The primary approach for defining disease in observational healthcare databases is to construct phenotype algorithms (PAs), rule-based heuristics predicated on the presence, absence, and temporal logic of clinical observations. However, a complete evaluation of PAs, i.e., determining sensitivity, specificity, and positive predictive value (PPV), is rarely performed. In this study, we propose a tool (PheValuator) to efficiently estimate a complete PA evaluation. METHODS: We used 4 administrative claims datasets: OptumInsight's de-identified Clinformatics™ Datamart (Eden Prairie,MN); IBM MarketScan Multi-State Medicaid); IBM MarketScan Medicare Supplemental Beneficiaries; and IBM MarketScan Commercial Claims and Encounters from 2000 to 2017. Using PheValuator involves (1) creating a diagnostic predictive model for the phenotype, (2) applying the model to a large set of randomly selected subjects, and (3) comparing each subject's predicted probability for the phenotype to inclusion/exclusion in PAs. We used the predictions as a 'probabilistic gold standard' measure to classify positive/negative cases. We examined 4 phenotypes: myocardial infarction, cerebral infarction, chronic kidney disease, and atrial fibrillation. We examined several PAs for each phenotype including 1-time (1X) occurrence of the diagnosis code in the subject's record and 1-time occurrence of the diagnosis in an inpatient setting with the diagnosis code as the primary reason for admission (1X-IP-1stPos). RESULTS: Across phenotypes, the 1X PA showed the highest sensitivity/lowest PPV among all PAs. 1X-IP-1stPos yielded the highest PPV/lowest sensitivity. Specificity was very high across algorithms. We found similar results between algorithms across datasets. CONCLUSION: PheValuator appears to show promise as a tool to estimate PA performance characteristics.

Wide-area transepithelial sampling with computer-assisted 3-dimensional analysis (WATS) markedly improves detection of esophageal dysplasia and Barrett's esophagus: analysis from a prospective multicenter community-based study.

The 4-quadrant forceps biopsy (FB) protocol for identifying Barrett's esophagus (BE) and esophageal dysplasia (ED) suffers from poor sensitivity due to significant sampling error. We investigated the benefit of wide-area transepithelial sampling with 3-dimensional computer-assisted analysis (WATS) used adjunctively to the combination of random and targeted FB in the detection of ED, and as a secondary outcome, BE. In this multicenter prospective trial, community endoscopists at 21 sites utilized WATS as an adjunct to both targeted and random FB in patients undergoing BE screening and surveillance. Investigators alternated taking FB and WATS samples first. WATS specimens were analyzed at CDx Diagnostics (Suffern, NY) while FB samples were analyzed by each site's regular pathologists. Data were de-identified and then aggregated for analysis. Of 12,899 patients enrolled, FB identified 88 cases of ED, and WATS detected an additional 213 cases missed by FB. These 213 cases represented an absolute increase of 1.65%, raising the yield from 0.68% to 2.33%. Adding WATS to FB increased the overall detection of ED by 242% (95% CI: 191%-315%). Fewer than 61 patients needed to be tested with WATS to identify an additional case of ED. The combination of random and targeted FB identified 1,684 cases of BE, and WATS detected an additional 2,570 BE cases. The absolute incremental yield of adding WATS to FB is 19.9%, increasing the rate of detection from 13.1% to 33%. Adding WATS to FB increased the overall detection of BE by 153% (95% CI: 144-162%). The number needed to test with WATS in order to detect an additional case of BE was 5. Whether FB or WATS was done first did not impact the results. In this study, comprised of the largest series of patients evaluated with WATS, adjunctive use of the technique with targeted and random FB markedly improved the detection of both ED and BE. These results underscore the shortcomings of FB in detecting BE-associated neoplasia, which can potentially impact the management and clinical outcomes of these patients.

Relationship between the polyp detection rate and the post-colonoscopy colorectal cancer rate.

AIM: the adenoma detection rate is the quality indicator of colonoscopy that is most closely related to the development of interval colorectal cancer or post-colonoscopy colorectal cancer. However, the recording of this indicator in different units of gastrointestinal endoscopy is obstructed due to the large consumption of resources required for its calculation. Several alternatives have been proposed, such as the polyp detection rate. The objective of this study was to evaluate the relationship between the polyp detection rate and its influence on post-colonoscopy colorectal cancer rate. PATIENTS AND METHODS: in this study, 12,482 colonoscopies conducted by 14 endoscopists were analyzed. The polyp detection rate was calculated for each endoscopist. Endoscopists were grouped into quartiles (Q1, Q2, Q3, and Q4), from lowest to highest polyp detection rate, in order to evaluate whether there were any differences in the development of post-colonoscopy colorectal cancer. RESULTS: the lowest polyp detection rate was 20.66% and the highest was 52.16%, with a median of 32.78 and a standard deviation of ± 8.54. A strong and positive association between polyp endoscopy diagnosis and adenoma histopathology result was observed and a linear regression was performed. A significantly higher post-colonoscopy colorectal cancer rate was observed in the group of endoscopists with a lower polyp detection rate (p < 0.02). CONCLUSION: polyp detection rate is a valuable quality indicator of colonoscopy and its calculation is much simpler than that of the adenoma detection rate. In our study, the prevalence of post-colonoscopy colorectal cancer was inversely and significantly related to the endoscopists' polyp detection rate.

A systematic review of the psychometric properties, usability and clinical impacts of mobile mood-monitoring applications in young people.

BACKGROUND: Mobile mood-monitoring applications are increasingly used by mental health providers, widely advocated within research, and a potentially effective method to engage young people. However, little is known about their efficacy and usability in young populations. METHOD: A systematic review addressing three research questions focused on young people: (1) what are the psychometric properties of mobile mood-monitoring applications; (2) what is their usability; and (3) what are their positive and negative clinical impacts? Findings were synthesised narratively, study quality assessed and compared with evidence from adult studies. RESULTS: We reviewed 25 articles. Studies on the psychometric properties of mobile mood-monitoring applications were sparse, but indicate questionable to excellent internal consistency, moderate concurrent validity and good usability. Participation rates ranged from 30% to 99% across studies, and appeared to be affected by methodological factors (e.g. payments) and individual characteristics (e.g. IQ score). Mobile mood-monitoring applications are positively perceived by youth, may reduce depressive symptoms by increasing emotional awareness, and could aid in the detection of mental health and substance use problems. There was very limited evidence on potential negative impacts. CONCLUSIONS: Evidence for the use of mood-monitoring applications in youth is promising but limited due to a lack of high-quality studies. Future work should explicate the effects of mobile mood-monitoring applications on effective self-regulation, clinical outcomes across disorders and young people's engagement with mental health services. Potential negative impacts in this population should also be investigated, as the adult literature suggests that application use could potentially increase negativity and depression symptoms.

Simultaneous inference for factorial multireader diagnostic trials.

We study inference methods for the analysis of multireader diagnostic trials. In these studies, data are usually collected in terms of a factorial design involving the factors Modality and Reader. Furthermore, repeated measures appear in a natural way since the same patient is observed under different modalities by several readers and the repeated measures may have a quite involved dependency structure. The hypotheses are formulated in terms of the areas under the ROC curves. Currently, only global testing procedures exist for the analysis of such data. We derive rank-based multiple contrast test procedures and simultaneous confidence intervals which take the correlation between the test statistics into account. The procedures allow for testing arbitrary multiple hypotheses. Extensive simulation studies show that the new approaches control the nominal type 1 error rate very satisfactorily. A real data set illustrates the application of the proposed methods.

A Device Histogram-Based Simple Predictor of Mortality Risk in ICD and CRT-D Patients: The Heart Rate Score.

BACKGROUND: We hypothesized that survival in implantable cardioverter defibrillator (ICD) and cardiac resynchronization therapy defibrillator (CRT-D) patients is predicted by baseline Heart Rate Score. METHODS: Heart Rate Score is determined from the atrial paced and sensed histogram of a DDD ICD or CRT-D, and defined as percent of beats in the histogram in the tallest 10 beats/min range bin. It was calculated at initial remote monitoring for patients enrolled in LATITUDE® without persistent atrial fibrillation, and with pulse generators implanted in 2006-2011. Univariate, multivariate, and Kaplan-Meier analyses determined the impact of Heart Rate Score on survival. RESULTS: Of 57,893 ICDs and 67,929 CRT-Ds followed for 2.4 ± 1.5 years, each 10% increase in Heart Rate Score was associated with decreased survival (CRT-D hazard ratio [HR] 1.07 95%, confidence interval 1.06-1.07, P < 0.0001; ICD HR 1.05, 95% confidence interval 1.04-1.06, P < 0.0001). Multivariate analysis showed survival decreased with increasing age, atrial fibrillation, presence of a shock in first-year follow-up, and increasing programmed lower pacing rate in ICD and CRT-D patients. Increased percent right ventricular pacing predicted mortality in ICD patients, while male gender and lower percent left ventricular pacing predicted mortality in CRT patients. Heart Rate Score predicted survival independent of those variables. Heart Rate Score correlates with heart rate variability (standard deviation of average R-R intervals [SDANN]) when both are obtainable, but SDANN was only present in 6% of patients with Heart Rate Score >70%. CONCLUSION: A simple device histogram measure, Heart Rate Score, predicts survival in ICD and CRT-D patients independent of the available variables, and even when SDANN is unavailable.

ECG-derived spatial QRS-T angle is strongly associated with hypertrophic cardiomyopathy.

INTRODUCTION: ECG-derived vectorcardiography (VCG) has diagnostic and prognostic value in various diseases. Hypertrophic cardiomyopathy (HCM), a genetic disease with unexplained left ventricular hypertrophy, is one of the most common causes of sudden cardiac death (SCD) in young persons. Genotype positive status is associated with increased risk of systolic dysfunction, heart failure, and (SCD). Herein, we aimed to determine the diagnostic utility of derived VCG parameters in a large cohort of genotyped HCM patients. METHODS: Between 1997 and 2007, genetic testing was performed on 1053 unrelated patients with HCM. Of these, 967 had 12-lead ECGs suitable for computerized derivation of VCG parameters, including the spatial mean and peaks QRS-T angles, spatial ventricular gradient (SVG), spatial QRS, QT, and Tpeak-Tend (TpTe) intervals. ECGs were also evaluated using Seattle ECG criteria. Differences between HCM patients and healthy controls as well as between genotype positive versus genotype negative HCM patients were assessed. RESULTS: Spatial peaks (129.3±26.4 vs.30.5±24.2 degrees) and spatial mean QRS-T angles (121.8±38.6 vs. 47.3±27.6 degrees) were significantly higher in patients with HCM than in controls (P<0.001). The spatial peaks and mean QRS-T angles identified 94% and 84% of HCM patients, respectively, while Seattle criteria identified 70.7% of patients (P<0.001). Genotype positive patients had higher spatial mean QRS-T angles, spatial TpTe (P<0.001 respectively), spatial peaks QRS-T angles (P=0.017) and lower SVG (P<0.001) than genotype negative patients. CONCLUSIONS: ECG-derived spatial QRS-T angles can differentiate patients with HCM from controls and could provide a better tool than traditional Seattle criteria. Clinical usefulness of VCG to differentiate genotype-negative from genotype-positive patients has yet to be established.

Assessment of continuous acoustic respiratory rate monitoring as an addition to a pulse oximetry-based patient surveillance system.

Technology advances make it possible to consider continuous acoustic respiratory rate monitoring as an integral component of physiologic surveillance systems. This study explores technical and logistical aspects of augmenting pulse oximetry-based patient surveillance systems with continuous respiratory rate monitoring and offers some insight into the impact on patient deterioration detection that may result. Acoustic respiratory rate sensors were introduced to a general care pulse oximetry-based surveillance system with respiratory rate alarms deactivated. Simulation was used after 4324 patient days to determine appropriate alarm thresholds for respiratory rate, which were then activated. Data were collected for an additional 4382 patient days. Physiologic parameters, alarm data, sensor utilization and patient/staff feedback were collected throughout the study and analyzed. No notable technical or workflow issues were observed. Sensor utilization was 57 %, with patient refusal leading reasons for nonuse (22.7 %). With respiratory rate alarm thresholds set to 6 and 40 breaths/min., the majority of nurse pager clinical notifications were triggered by low oxygen saturation values (43 %), followed by low respiratory rate values (21 %) and low pulse rate values (13 %). Mean respiratory rate collected was 16.6 ± 3.8 breaths/min. The vast majority (82 %) of low oxygen saturation states coincided with normal respiration rates of 12-20 breaths/min. Continuous respiratory rate monitoring can be successfully added to a pulse oximetry-based surveillance system without significant technical, logistical or workflow issues and is moderately well-tolerated by patients. Respiratory rate sensor alarms did not significantly impact overall system alarm burden. Respiratory rate and oxygen saturation distributions suggest adding continuous respiratory rate monitoring to a pulse oximetry-based surveillance system may not significantly improve patient deterioration detection.

Accuracy and Efficiency of Recording Pediatric Early Warning Scores Using an Electronic Physiological Surveillance System Compared With Traditional Paper-Based Documentation.

Pediatric Early Warning Scores are advocated to assist health professionals to identify early signs of serious illness or deterioration in hospitalized children. Scores are derived from the weighting applied to recorded vital signs and clinical observations reflecting deviation from a predetermined "norm." Higher aggregate scores trigger an escalation in care aimed at preventing critical deterioration. Process errors made while recording these data, including plotting or calculation errors, have the potential to impede the reliability of the score. To test this hypothesis, we conducted a controlled study of documentation using five clinical vignettes. We measured the accuracy of vital sign recording, score calculation, and time taken to complete documentation using a handheld electronic physiological surveillance system, VitalPAC Pediatric, compared with traditional paper-based charts. We explored the user acceptability of both methods using a Web-based survey. Twenty-three staff participated in the controlled study. The electronic physiological surveillance system improved the accuracy of vital sign recording, 98.5% versus 85.6%, P < .02, Pediatric Early Warning Score calculation, 94.6% versus 55.7%, P < .02, and saved time, 68 versus 98 seconds, compared with paper-based documentation, P < .002. Twenty-nine staff completed the Web-based survey. They perceived that the electronic physiological surveillance system offered safety benefits by reducing human error while providing instant visibility of recorded data to the entire clinical team.

Quantitative Romberg's test in acute carbon monoxide poisoning treated by hyperbaric oxygen.

OBJECTIVE: The aim of this study was to evaluate whether monitoring of acute carbon monoxide-poisoned (COP) patients by means of quantitative Romberg's test (QR-test) during a hyperbaric oxygen (HBO2) therapy regimen could be a useful supplement in the evaluation of neurological status. METHODS: We conducted a retrospective study (2000-2014) in which we evaluated data containing quantitative sway measurements of acute COP patients (n = 58) treated in an HBO2 regimen. Each patient was tested using QR-test before and after each HBO2 treatment. Data were analyzed using linear mixed models (LMM). In each LMM, sway prior to HBO2 therapy was set as the fixed effect and change in sway after HBO2 therapy was set as the response variable. Patient, treatment number, weight and age were set as random effects for all LMMs. RESULTS: From the LMMs we found that larger values of sway prior to HBO2 produced a negative change in sway. We found no correlation between CO level and sway (P=0.1028; P=0.8764; P=0.4749; P=0.5883). Results showed that loss of visual input caused a significant increase in mean sway (P=0.028) and sway velocity (P⟨0.0001). CONCLUSIONS: The Quantitative Romberg's test is a fast, useful supplement to neurological evaluation and a potential valuable tool for monitoring postural stability during the course of treatment in acute COP patients.

Computerized vital signs analysis and late onset infections in extremely low gestational age infants.

AIM: Current clinical and laboratory diagnostics for neonatal infection are inadequate. An infant's systemic inflammatory response may be identified earlier than clinical suspicion by a computerized algorithm (RALIS) incorporating multiple vital signs (VS). We tested the ability of RALIS to detect late onset infection (LOI) earlier than clinically suspected. METHODS: We conducted a retrospective review of infants enrolled in a birth cohort study at Prentice Women's Hospital. VS data (heart rate, respirations, temperature, desaturation, bradycardia) were extracted from electronic records of 73 premature infants (born ≤28 weeks' gestation; survived first month). RALIS generated a continuous output for the first 28 days of life. A score ≥5 for 6 h triggered an alert. The time of RALIS alert to time of clinical suspicion of infection (time culture sent) was measured for each episode of suspected and/or confirmed LOI. RESULTS: Among the 73 infants followed with RALIS, there were 34 episodes of culture-positive LOI, seven culture-negative but treated episodes, and 13 false-positive culture (untreated) episodes. Twenty-five infants had no culture-positive or treated sepsis events during the observation period. There was a positive linear association between alert and culture (ß=0.88, P<0.001). Mean absolute time difference between alert and culture was 59.4 h before culture. Sensitivity and specificity of RALIS for LOI were 0.82 and 0.44. CONCLUSION: The RALIS algorithm is a sensitive indicator for early detection of infection in preterm infants. Further modifications to improve the specificity of the algorithm are needed prior to application of VS modeling to patient antibiotic treatment decisions.

Accuracy of the pacemaker-mediated tachycardia algorithm in Boston Scientific devices.

INTRODUCTION: The incidence of pacemaker-mediated tachycardia (PMT) varies as a function of patient characteristics, device programming and algorithm specificities. We investigated the efficacy of the Boston Scientific algorithm by reviewing PMT episodes in a large device population. METHODS: In this multicenter study, we included 328 patients implanted with a Boston Scientific device: 157 non-dependent patients with RYTHMIQ™ activated (RYTHMIQ group), 76 patients with permanent AV-conduction disorder (AV-block group) and 95 Cardiac Resynchronization Therapy patients (CRT group). For each patient, we reviewed the last 10 remote monitoring-transmitted EGMs diagnosed as PMT. RESULTS: We analyzed 784 PMT episodes across 118 patients. In the RYTHMIQ group, the diagnosis of PMT was correct in most episodes (80%) of which 69% was directly related to the prolongation of the AV-delay associated with the RYTHMIQ algorithm. The usual triggers for PMT were also observed (PVC 16%, PAC 9%). The remainder of the episodes (20%) in RYTHMIQ patients and most episodes of AV-block (66%) and CRT patients (74%) were incorrectly diagnosed as PMT during sinus tachycardia at the maximal tracking rate. The inappropriate intervention of the algorithm during exercise causes non-conducted P-waves, loss of CRT (sustained in six patients) and may have been pro-arrhythmogenic in one patient (induction of ventricular tachycardia). CONCLUSION: Algorithms to minimize ventricular pacing can occasionally have unintended consequences such as PMT. The PMT algorithm in Boston Scientific devices is associated with a high rate of incorrect PMT diagnosis during exercise resulting in inappropriate therapy with non-conducted P-waves, loss of CRT and limited risk of pro-arrhythmic events.

The evaluation of an open source online training system for teaching 12 lead electrocardiographic interpretation.

INTRODUCTION: The aim of this study is to present and evaluate the integration of a low resource JavaScript based ECG training interface (CrowdLabel) and a standardised curriculum for self-guided tuition in ECG interpretation. METHODS: Participants practiced interpreting ECGs weekly using the CrowdLabel interface to assist with the learning of the traditional didactic taught course material during a 6 week training period. To determine competency students were tested during week 7. RESULTS: A total of 245 unique ECG cases were submitted by each student. Accuracy scores during the training period ranged from 0-59.5% (median = 33.3%). Conversely accuracy scores during the test ranged from 30 - 70% (median = 37.5%) (p < 0.05). There was no correlation between students who interpreted high numbers of ECGs during the training period and their marks obtained. CONCLUSIONS: CrowdLabel is shown to be a readily accessible dedicated learning platform to support ECG interpretation competency.