Estimating the pertussis burden in adolescents and adults in the United States between 2007 and 2019.

We developed a machine learning algorithm to identify undiagnosed pertussis episodes in adolescent and adult patients with reported acute respiratory disease (ARD) using clinician notes in an electronic healthcare record (EHR) database. Here, we utilized the algorithm to better estimate the overall pertussis incidence within the Optum Humedica clinical repository from 1 January 2007 through 31 December 2019. The incidence of diagnosed pertussis episodes was 1-5 per 100,000 annually, consistent with data registered by the US Centers for Disease Control and Prevention (CDC) over the same time period. Among 18,573,496 ARD episodes assessed, 1,053,946 were identified (i.e. algorithm-identified) as likely undiagnosed pertussis episodes. Accounting for these undiagnosed pertussis episodes increased the estimated pertussis incidence by 110-fold on average (34-474 per 100,000 annually). Risk factors for pertussis episodes (diagnosed and algorithm-identified) included asthma (Odds ratio [OR] 2.14; 2.12-2.16), immunodeficiency (OR 1.85; 1.78-1.91), chronic obstructive pulmonary disease (OR 1.63; 1.61-1.65), obesity (OR 1.44; 1.43-1.45), Crohn's disease (OR 1.39; 1.33-1.45), diabetes type 1 (OR 1.21; 1.17-1.24) and type 2 (OR 1.12; 1.1-1.13). Of note, all these risk factors, except Crohn's disease, increased the likelihood of severe pertussis. In conclusion, the incidence of pertussis in the adolescent and adult population in the USA is likely substantial, but considerably under-recognized, highlighting the need for improved clinical awareness of the disease and for improved control strategies in this population. These results will help better inform public health vaccination and booster programs, particularly in those with underlying comorbidities.

Development of an algorithm for finding pertussis episodes in a population-based electronic health record database.

While tetanus-diphtheria-acellular pertussis (Tdap) vaccines for adolescents and adults were licensed in 2005 and immunization strategies proposed, the burden of pertussis in this population remains under-recognized mainly due to atypical disease presentation, undermining efforts to optimize protection through vaccination. We developed a machine learning algorithm to identify undiagnosed/misdiagnosed pertussis episodes in patients diagnosed with acute respiratory disease (ARD) using signs, diseases and symptoms from clinician notes and demographic information within electronic health-care records (Optum Humedica repository [2007-2019]). We used two patient cohorts aged ≥11 years to develop the model: a positive pertussis cohort (4,515 episodes in 4,316 patients) and a negative pertussis (ARD) cohort (4,573,445 episodes and patients), defined using ICD 9/10 codes. To improve contrast between positive pertussis and negative pertussis (ARD) episodes, only episodes with ≥7 symptoms were selected. LightGBM was used as the machine learning model for pertussis episode identification. Model validity was determined using laboratory-confirmed pertussis positive and negative cohorts. Model explainability was obtained using the Shapley additive explanations method. The predictive performance was as follows: area under the precision-recall curve, 0.24 (SD, 7 × 10-3); recall, 0.72 (SD, 4 × 10-3); precision, 0.012 (SD, 1 × 10-3); and specificity, 0.94 (SD, 7 × 10-3). The model applied to laboratory-confirmed positive and negative pertussis episodes had a specificity of 0.846. Predictive probability for pertussis increased with presence of whooping cough, whoop, and post-tussive vomiting in clinician notes, but decreased with gastrointestinal bleeding, sepsis, pulmonary symptoms, and fever. In conclusion, machine learning can help identify pertussis episodes among those diagnosed with ARD.

Verification and Validation of Computational Models Used in Biopharmaceutical Manufacturing: Potential Application of the ASME Verification and Validation 40 Standard and FDA Proposed AI/ML Model Life Cycle Management Framework.

A wide variety of computational models covering statistical, mechanistic, and machine learning (locked and adaptive) methods are explored for use in biopharmaceutical manufacturing. Limited discussion exists on how to establish the credibility of a computational model for application in biopharmaceutical manufacturing. In this work, we tried to use the American Society of Mechanical Engineers (ASME) Verification and Validation 40 (V&V 40) standard and FDA proposed AI/ML model life cycle management framework for Software as a Medical Device (SaMD) in biopharmaceutical manufacturing use cases, by applying to a set of curated hypothetical examples. We discussed the need for standardized frameworks to facilitate consistent decision making to enable efficient adoption of computational models in biopharmaceutical manufacturing and alignment of existing good practices with existing frameworks. In the study of our examples, we anticipate existing frameworks like V&V 40 can be adopted.

Predicting potential adverse events using safety data from marketed drugs.

BACKGROUND: While clinical trials are considered the gold standard for detecting adverse events, often these trials are not sufficiently powered to detect difficult to observe adverse events. We developed a preliminary approach to predict 135 adverse events using post-market safety data from marketed drugs. Adverse event information available from FDA product labels and scientific literature for drugs that have the same activity at one or more of the same targets, structural and target similarities, and the duration of post market experience were used as features for a classifier algorithm. The proposed method was studied using 54 drugs and a probabilistic approach of performance evaluation using bootstrapping with 10,000 iterations. RESULTS: Out of 135 adverse events, 53 had high probability of having high positive predictive value. Cross validation showed that 32% of the model-predicted safety label changes occurred within four to nine years of approval (median: six years). CONCLUSIONS: This approach predicts 53 serious adverse events with high positive predictive values where well-characterized target-event relationships exist. Adverse events with well-defined target-event associations were better predicted compared to adverse events that may be idiosyncratic or related to secondary target effects that were poorly captured. Further enhancement of this model with additional features, such as target prediction and drug binding data, may increase accuracy.

A framework to characterize the performance of early warning index alarm systems for patient monitoring.

In [Scully, C.G., and Daluwatte, C., Evaluating performance of early warning indices to predict physiological instabilities. J Biomed Inform. 75 (2017) 14-21], a framework was presented to characterize the performance of warning indices to provide information on the 1) probability a critical health event will occur when a warning is given (analogous to positive predictive value) and 2) proportion of warned events to all events (analogous to sensitivity). This framework also provides information about the timeliness of the warnings with respect to event occurrence and the warning burden of the system. •In the current work, we provide information on how this framework can be used when cases without events are present in a dataset to examine the proportion of warned non-events to all non-events (analogous to false positive rate).•Information on steps to apply the method, software, data and results for the case study are also provided to enable implementation of the framework.•Application and extension of the framework is demonstrated and discussed by adding non-event records to our previous case study comparing two warning strategies to predict physiologic instabilities.

Clinical Trial in a Dish: Personalized Stem Cell-Derived Cardiomyocyte Assay Compared With Clinical Trial Results for Two QT-Prolonging Drugs.

Induced pluripotent stem cells (iPSCs) have shown promise in investigating donor-specific phenotypes and pathologies. The iPSC-derived cardiomyocytes (iPSC-CMs) could potentially be utilized in personalized cardiotoxicity studies, assessing individual proarrhythmic risk. However, it is unclear how closely iPSC-CMs derived from healthy subjects can recapitulate a range of responses to drugs. It is well known that QT-prolonging drugs induce subject-specific clinical response and that all healthy subjects do not necessarily develop arrhythmias or exhibit similar amounts of QT prolongation. We previously reported this variability in a study of four human ether-a-go-go-related gene (hERG) potassium channel-blocking drugs in which each subject underwent intensive pharmacokinetic and pharmacodynamic sampling such that subjects had 15 time-matched plasma drug concentration and electrocardiogram measurements throughout 24 hours after dosing in a phase I clinical research unit. In this study, iPSC-CMs were generated from those subjects. Their drug-concentration-dependent QT prolongation response from the clinic was compared with in vitro drug-concentration-dependent action potential duration (APD) prolongation response to the same two hERG-blocking drugs, dofetilide and moxifloxacin. Comparative results showed no significant correlation between the subject-specific APD response slopes and clinical QT response slopes to either moxifloxacin (P = 0.75) or dofetilide (P = 0.69). Similarly, no significant correlation was found between baseline QT and baseline APD measurements (P = 0.93). This result advances our current understanding of subject-specific iPSC-CMs and facilitates discussion into factors obscuring correlation and considerations for future studies of subject-specific phenotypes in iPSC-CMs.

Credibility Evidence for Computational Patient Models Used in the Development of Physiological Closed-Loop Controlled Devices for Critical Care Medicine.

Physiological closed-loop controlled medical devices automatically adjust therapy delivered to a patient to adjust a measured physiological variable. In critical care scenarios, these types of devices could automate, for example, fluid resuscitation, drug delivery, mechanical ventilation, and/or anesthesia and sedation. Evidence from simulations using computational models of physiological systems can play a crucial role in the development of physiological closed-loop controlled devices; but the utility of this evidence will depend on the credibility of the computational model used. Computational models of physiological systems can be complex with numerous non-linearities, time-varying properties, and unknown parameters, which leads to challenges in model assessment. Given the wide range of potential uses of computational patient models in the design and evaluation of physiological closed-loop controlled systems, and the varying risks associated with the diverse uses, the specific model as well as the necessary evidence to make a model credible for a use case may vary. In this review, we examine the various uses of computational patient models in the design and evaluation of critical care physiological closed-loop controlled systems (e.g., hemodynamic stability, mechanical ventilation, anesthetic delivery) as well as the types of evidence (e.g., verification, validation, and uncertainty quantification activities) presented to support the model for that use. We then examine and discuss how a credibility assessment framework (American Society of Mechanical Engineers Verification and Validation Subcommittee, V&V 40 Verification and Validation in Computational Modeling of Medical Devices) for medical devices can be applied to computational patient models used to test physiological closed-loop controlled systems.

Sex Differences of Patients With Systemic Hypertension (From the Analysis of the Systolic Blood Pressure Intervention Trial [SPRINT]).

There are differences in the incidence, pathophysiology, and long-term effects of hypertension between women and men. We assessed sex-specific benefit-risk tradeoffs of different blood pressure (BP) goals in patients enrolled in the Systolic Blood Pressure Intervention Trial (SPRINT) after propensity score matching those with standard therapy (systolic BP <140 mm Hg) to those with intensive therapy (systolic BP <120 mm Hg; n = 9,106). Cox regression was conducted to compare standard versus intensive therapy in women and men with the composite outcome of myocardial infarction, other acute coronary syndromes, stroke, heart failure, or death from cardiovascular causes. Women were generally healthier at baseline and had a lower cardiovascular risk. Men on intensive therapy had a lower risk of the composite outcome compared to those on standard therapy (hazard ratio [HR] 0.70, 95% confidence interval [CI] 0.57 to 0.86, p = 0.001) while in women no differences between therapy groups were observed (HR 0.82 [0.60 to 1.12], p = 0.206). For safety outcomes, women and men had increased risk of related serious adverse events with intensive treatment (HR 1.52 [1.06 to 2.18], p = 0.023 and HR 2.07 [1.55 to2.77], p < 0.001, respectively). In conclusion, our study demonstrated that women did not benefit from intensive compared to standard BP control. A potential explanation for this may be the lower baseline cardiovascular risk in women.

Multivariate physiological recordings in an experimental hemorrhage model.

In this paper we describe a data set of multivariate physiological measurements recorded from conscious sheep (N = 8; 37.4 ± 1.1 kg) during hemorrhage. Hemorrhage was experimentally induced in each animal by withdrawing blood from a femoral artery at two different rates (fast: 1.25 mL/kg/min; and slow: 0.25 mL/kg/min). Data, including physiological waveforms and continuous/intermittent measurements, were transformed to digital file formats (European Data Format [EDF] for waveforms and Comma-Separated Values [CSV] for continuous and intermittent measurements) as a comprehensive data set and stored and publicly shared here (Appendix A). The data set comprises experimental information (e.g., hemorrhage rate, animal weight, event times), physiological waveforms (arterial and central venous blood pressure, electrocardiogram), time-series records of non-invasive physiological measurements (SpO2, tissue oximetry), intermittent arterial and venous blood gas analyses (e.g., hemoglobin, lactate, SaO2, SvO2) and intermittent thermodilution cardiac output measurements. A detailed explanation of the hemodynamic and pulmonary changes during hemorrhage is available in a previous publication (Scully et al., 2016) [1].

Detecting atrial fibrillation from short single lead ECGs using statistical and morphological features.

OBJECTIVE: Point of care ECG devices can improve the early detection of atrial fibrillation (AF). The efficiency of such devices depends on the capability of automatic AF detection against normal sinus rhythm and other arrhythmias from a short single lead ECG record in the presence of noise and artifacts. The objective of this study was to develop an algorithm that classifies a short single lead ECG record into 'Normal', 'AF', 'Other' and 'Noisy' classes, and identify the challenges in developing such algorithms and potential mitigation steps. APPROACH: Rule-based identification was used to detect lead inversion and records too noisy to be of immediate use. A set of statistical and morphological features describing the rhythm was then extracted, and support vector machine classifiers were used to classify records into three classes: 'Normal', 'AF' or 'Other'. The algorithm was trained and tested using 12 186 short single lead ECGs recorded on a point of care device made available via the Computing in Cardiology Challenge 2017. MAIN RESULTS: The algorithm achieved a sensitivity of 77.5%, a specificity of 97.9% and an accuracy of 96.1% in the detection of AF from a non-AF rhythm in a five-fold cross validation. It achieved F1 measures of 89%, 78% and 67% for 'Normal', 'AF' and 'Other' classes, respectively, when evaluated with a hidden test set. The overall challenge score was 78%. SIGNIFICANCE: Most existing algorithms can distinguish the AF rhythm from the normal sinus rhythm when ECG recordings are clean and are obtained with multi-lead systems, while their ability to discriminate against other arrhythmias and noise remains largely unknown. This study proposes an algorithm that classifies a short single lead ECG record from point of care devices into 'Normal', 'AF', 'Other' and 'Noisy' classes and discusses computational approaches to mitigate any unique challenges such as lead inversion, low amplitude signals, noise and artifacts.

A novel ECG detector performance metric and its relationship with missing and false heart rate limit alarms.

PURPOSE: Performance of ECG beat detectors is traditionally assessed on long intervals (e.g.: 30min), but only incorrect detections within a short interval (e.g.: 10s) may cause incorrect (i.e., missed+false) heart rate limit alarms (tachycardia and bradycardia). We propose a novel performance metric based on distribution of incorrect beat detection over a short interval and assess its relationship with incorrect heart rate limit alarm rates. BASIC PROCEDURES: Six ECG beat detectors were assessed using performance metrics over long interval (sensitivity and positive predictive value over 30min) and short interval (Area Under empirical cumulative distribution function (AUecdf) for short interval (i.e., 10s) sensitivity and positive predictive value) on two ECG databases. False heart rate limit and asystole alarm rates calculated using a third ECG database were then correlated (Spearman's rank correlation) with each calculated performance metric. MAIN FINDINGS: False alarm rates correlated with sensitivity calculated on long interval (i.e., 30min) (ρ=-0.8 and p<0.05) and AUecdf for sensitivity (ρ=0.9 and p<0.05) in all assessed ECG databases. Sensitivity over 30min grouped the two detectors with lowest false alarm rates while AUecdf for sensitivity provided further information to identify the two beat detectors with highest false alarm rates as well, which was inseparable with sensitivity over 30min. PRINCIPAL CONCLUSIONS: Short interval performance metrics can provide insights on the potential of a beat detector to generate incorrect heart rate limit alarms.

Applying a Computational Model Credibility Framework to Physiological Closed-loop Controlled Medical Device Testing.

Physiological closed-loop controlled medical devices are safety-critical systems that combine patient monitors with therapy delivery devices to automatically titrate therapy to meet a patient's current need. Computational models of physiological systems can be used to test these devices and generate pre-clinical evidence of safety and performance before using the devices on patients. The credibility, utility, and acceptability of such model-based test results will depend on, among other factors, the computational model used. We examine how a recently developed risk-informed framework for establishing the credibility of computational models in medical device applications can be applied in the evaluation of physiological closed-loop controlled devices.

Evaluating performance of early warning indices to predict physiological instabilities.

Patient monitoring algorithms that analyze multiple features from physiological signals can produce an index that serves as a predictive or prognostic measure for a specific critical health event or physiological instability. Classical detection metrics such as sensitivity and positive predictive value are often used to evaluate new patient monitoring indices for such purposes, but since these metrics do not take into account the continuous nature of monitoring, the assessment of a warning system to notify a user of a critical health event remains incomplete. In this article, we present challenges of assessing the performance of new warning indices and propose a framework that provides a more complete characterization of warning index performance predicting a critical event that includes the timeliness of the warning. The framework considers 1) an assessment of the sensitivity to provide a notification within a meaningful time window, 2) the cumulative sensitivity leading up to an event, 3) characteristics on if the warning stays on until the event occurs once a warning has been activated, and 4) the distribution of warning times and the burden of additional warnings (e.g., false-alarm rate) throughout monitoring that may or may not be associated with the event of interest. Using an example from an experimental study of hemorrhage, we examine how this characterization can differentiate two warning systems in terms of timeliness of warnings and warning burden.

Progression and variability of physiologic deterioration in an ovine model of lung infection sepsis.

In this study, a lung infection model of pneumonia in sheep (n = 12) that included smoke inhalation injury followed by methicillin-resistant Staphylococcus aureus placement into the lungs was used to investigate hemodynamic and pulmonary dysfunctions during the course of sepsis progression. To assess the variability in disease progression, animals were retrospectively divided into survivor (n = 6) and nonsurvivor (n = 6) groups, and a range of physiological indexes reflecting hemodynamic and pulmonary function were estimated and compared to evaluate variability in dynamics underlying sepsis development. Blood pressure and heart rate variability analyses were performed to assess whether they discriminated between the survivor and nonsurvivor groups early on and after intervention. Results showed hemodynamic deterioration in both survivor and nonsurvivor animals during sepsis along with a severe oxygenation disruption (decreased peripheral oxygen saturation) in nonsurvivors separating them from survivor animals of this model. Variability analysis of beat-to-beat heart rate and blood pressure reflected physiologic deterioration during infection for all animals, but these analyses did not discriminate the nonsurvivor animals from survivor animals.NEW & NOTEWORTHY Variable pulmonary response to injury results in varying outcomes in a previously reported animal model of lung injury and methicillin-resistant Staphylococcus aureus-induced sepsis. Heart rate and blood pressure variability analyses were investigated to track the varying levels of physiologic deterioration but did not discriminate early nonsurvivors from survivors.

Effect of hemorrhage rate on early hemodynamic responses in conscious sheep.

Physiological compensatory mechanisms can mask the extent of hemorrhage in conscious mammals, which can be further complicated by individual tolerance and variations in hemorrhage onset and duration. We assessed the effect of hemorrhage rate on tolerance and early physiologic responses to hemorrhage in conscious sheep. Eight Merino ewes (37.4 ± 1.1 kg) were subjected to fast (1.25 mL/kg/min) and slow (0.25 mL/kg/min) hemorrhages separated by at least 3 days. Blood was withdrawn until a drop in mean arterial pressure (MAP) of >30 mmHg and returned at the end of the experiment. Continuous monitoring includedMAP, central venous pressure, pulmonary artery pressure, pulse oximetry, and tissue oximetry. Cardiac output by thermodilution and arterial blood samples were also measured. The effects of fast versus slow hemorrhage rates were compared for total volume of blood removed and stoppage time (whenMAP < 30 mmHg of baseline) and physiological responses during and after the hemorrhage. Estimated blood volume removed whenMAPdropped 30 mmHg was 27.0 ± 4.2% (mean ± standard error) in the slow and 27.3 ± 3.2% in the fast hemorrhage (P = 0.47, pairedttest between rates). Pressure and tissue oximetry responses were similar between hemorrhage rates. Heart rate increased at earlier levels of blood loss during the fast hemorrhage, but hemorrhage rate was not a significant factor for individual hemorrhage tolerance or hemodynamic responses. In 5/16 hemorrhages MAP stopping criteria was reached with <25% of blood volume removed. This study presents the physiological responses leading up to a significant drop in blood pressure in a large conscious animal model and how they are altered by the rate of hemorrhage.

Atypical pupillary light reflex and heart rate variability in children with autism spectrum disorder.

We investigated pupillary light reflex (PLR) in 152 children with ASD, 116 typically developing (TD) children, and 36 children with non-ASD neurodevelopmental disorders (NDDs). Heart rate variability (HRV) was measured simultaneously to study potential impairments in the autonomic nervous system (ANS) associated with ASD. The results showed that the ASD group had significantly longer PLR latency, reduced relative constriction amplitude, and shorter constriction/redilation time than those of the TD group. Similar atypical PLR parameters were observed in the NDD group. A significant age effect on PLR latency was observed in children younger than 9 years in the TD group, but not in the ASD and NDD groups. Atypical HRV parameters were observed in the ASD and NDD groups. A significant negative correlation existed between the PLR constriction amplitude and average heart rate in children with an ASD, but not in children with typical development.

Age-dependent pupillary light reflex parameters in children.

Pupillary light reflex (PLR) refers to the phenomenon where pupil size changes in response to stimulation with a flash of light. It is a simple functional test that can reveal dysfunctions associated with the PLR pathway. Although abnormal PLR responses have been reported in many neurological disorders, few studies investigated neurodevelopmental effects on PLR parameters. We studied the effect of age on PLR in a group of 6 to 17 year old children with typical development. A significant and consistent age effect was found on PLR latency in children younger than 10 years old. Age effects were also observed in resting pupil diameter and constriction amplitude. However such age related trends were not observed in children with neurodevelopment disorders. These results suggest that PLR has the potential to be used as a simple noninvasive tool for monitoring neurodevelopment in children.

Photon diffusion in a homogeneous medium bounded externally or internally by an infinitely long circular cylindrical applicator. II. Quantitative examinations of the steady-state theory.

This is Part II of the work that examines photon diffusion in a homogenous medium enclosed by a concave circular cylindrical applicator or enclosing a convex circular cylindrical applicator. Part I of this work [J. Opt. Soc. Am. A 27, 648 (2010)] analytically examined the steady-state photon diffusion between a source and a detector for two specific cases: (1) the detector is placed only azimuthally with respect to the source, and (2) the detector is placed only longitudinally with respect to the source, in the infinitely long concave and convex applicator geometries. For the first case, it was predicted that the decay rate of photon fluence would become smaller in the concave geometry and greater in the convex geometry than that in the semi-infinite geometry for the same source-detector distance. For the second case, it was projected that the decay rate of photon fluence would be greater in the concave geometry and smaller in the convex geometry than that in the semi-infinite geometry for the same source-detector distance. This Part II of the work quantitatively examines these predictions from Part I through several approaches, including (a) the finite-element method, (b) the Monte Carlo simulation, and (c) experimental measurement. Despite that the quantitative examinations have to be conducted for finite cylinder applicators with large length-to-radius ratio to approximate the infinite-length condition modeled in Part I, the results obtained by these quantitative methods for two concave and three convex applicator dimensions validated the qualitative trend predicted by Part I and verified the quantitative accuracy of the analytic treatment of Part I in the diffusion regime of the measurement, at a given set of absorption and reduced scattering coefficients of the medium.