Mitigating Racial And Ethnic Bias And Advancing Health Equity In Clinical Algorithms: A Scoping Review.

In August 2022 the Department of Health and Human Services (HHS) issued a notice of proposed rulemaking prohibiting covered entities, which include health care providers and health plans, from discriminating against individuals when using clinical algorithms in decision making. However, HHS did not provide specific guidelines on how covered entities should prevent discrimination. We conducted a scoping review of literature published during the period 2011-22 to identify health care applications, frameworks, reviews and perspectives, and assessment tools that identify and mitigate bias in clinical algorithms, with a specific focus on racial and ethnic bias. Our scoping review encompassed 109 articles comprising 45 empirical health care applications that included tools tested in health care settings, 16 frameworks, and 48 reviews and perspectives. We identified a wide range of technical, operational, and systemwide bias mitigation strategies for clinical algorithms, but there was no consensus in the literature on a single best practice that covered entities could employ to meet the HHS requirements. Future research should identify optimal bias mitigation methods for various scenarios, depending on factors such as patient population, clinical setting, algorithm design, and types of bias to be addressed.

Movement Patterns of Transient and Prolonged Positioning Events in Nursing Home Residents: Results from the TEAM-UP Trial.

OBJECTIVE: To characterize transient and prolonged body position patterns in a large sample of nursing home (NH) residents and describe the variability in movement patterns based on time of occurrence. METHODS: This study is a descriptive, exploratory analysis of up to 28 days of longitudinal accelerometer data for 1,100 NH residents from the TEAM-UP (Turn Everyone and Move for Ulcer Prevention) clinical trial. Investigators analyzed rates of transient events (TEs; less than 60 seconds) and prolonged events (PEs; 60 seconds or longer) and their interrelationships by nursing shift. RESULTS: Residents' positions changed for at least 1 minute (PEs) nearly three times per hour. Shorter-duration movements (TEs) occurred almost eight times per hour. Residents' PE rates were highest in shift 2 (3 pm to 11 pm ), when the median duration and maximum lengths of PEs were lowest; the least active time of day was shift 3 (11 pm to 7 am ). Three-quarters of all PEs lasted less than 15 minutes. The rate of TEs within PEs decreased significantly as the duration of PEs increased. CONCLUSIONS: The NH residents demonstrate complex patterns of movements of both short and prolonged duration while lying and sitting. Findings represent how NH residents naturally move in real-world conditions and provide a new set of metrics to study tissue offloading and its role in pressure injury prevention.

Technical considerations for evaluating clinical prediction indices: a case study for predicting code blue events with MEWS.

Objective.There have been many efforts to develop tools predictive of health deterioration in hospitalized patients, but comprehensive evaluation of their predictive ability is often lacking to guide implementation in clinical practice. In this work, we propose new techniques and metrics for evaluating the performance of predictive alert algorithms and illustrate the advantage of capturing the timeliness and the clinical burden of alerts through the example of the modified early warning score (MEWS) applied to the prediction of in-hospital code blue events.Approach. Different implementations of MEWS were calculated from available physiological parameter measurements collected from the electronic health records of ICU adult patients. The performance of MEWS was evaluated using conventional and a set of non-conventional metrics and approaches that take into account the timeliness and practicality of alarms as well as the false alarm burden.Main results. MEWS calculated using the worst-case measurement (i.e. values scoring 3 points in the MEWS definition) over 2 h intervals significantly reduced the false alarm rate by over 50% (from 0.19/h to 0.08/h) while maintaining similar sensitivity levels as MEWS calculated from raw measurements (∼80%). By considering a prediction horizon of 12 h preceding a code blue event, a significant improvement in the specificity (∼60%), the precision (∼155%), and the work-up to detection ratio (∼50%) could be achieved, at the cost of a relatively marginal decrease in sensitivity (∼10%).Significance. Performance aspects pertaining to the timeliness and burden of alarms can aid in understanding the potential utility of a predictive alarm algorithm in clinical settings.

Explainability Metrics of Deep Convolutional Networks for Photoplethysmography Quality Assessment.

Photoplethysmography (PPG) is a noninvasive way to monitor various aspects of the circulatory system, and is becoming more and more widespread in biomedical processing. Recently, deep learning methods for analyzing PPG have also become prevalent, achieving state of the art results on heart rate estimation, atrial fibrillation detection, and motion artifact identification. Consequently, a need for interpretable deep learning has arisen within the field of biomedical signal processing. In this paper, we pioneer novel explanatory metrics which leverage domain-expert knowledge to validate a deep learning model. We visualize model attention over a whole testset using saliency methods and compare it to human expert annotations. Congruence, our first metric, measures the proportion of model attention within expert-annotated regions. Our second metric, Annotation Classification, measures how much of the expert annotations our deep learning model pays attention to. Finally, we apply our metrics to compare between a signal based model and an image based model for PPG signal quality classification. Both models are deep convolutional networks based on the ResNet architectures. We show that our signal-based one dimensional model acts in a more explainable manner than our image based model; on average 50.78% of the one dimensional model's attention are within expert annotations, whereas 36.03% of the two dimensional model's attention are within expert annotations. Similarly, when thresholding the one dimensional model attention, one can more accurately predict if each pixel of the PPG is annotated as artifactual by an expert. Through this testcase, we demonstrate how our metrics can provide a quantitative and dataset-wide analysis of how explainable the model is.

Causal relationship between neuronal activity and cerebral hemodynamics in patients with ischemic stroke.

OBJECTIVE: Neurovascular coupling enables rapid adaptation of cerebral blood flow (CBF) to support neuronal activity. Modern techniques enable the simultaneous recording of neuronal activities and hemodynamic parameters. However, the causal relationship between electrical brain activity and CBF is still unclarified. In this study, we investigated the causal relationship between surface electroencephalogram (EEG) and cerebral blood flow velocity (FV) from transcranial Doppler using Granger causality (GC) analysis. APPROACH: Twenty simultaneous recordings of EEG and FV from 17 acute ischemic stroke patients were studied. Each patient had simultaneous, continuous monitoring of EEG and bilateral FVs in either the middle cerebral arteries or posterior cerebral arteries. The causal interactions between FV (0.006-0.4 Hz) and EEG (delta, theta, alpha, beta and gamma bands) were investigated through GC index (GCI). In order to make the GCIs comparable, the proportion of GCI (PGCI) values where G-causality is statistically significant were calculated. Scores on the NIH Stroke Scale (NIHSS) and the modified Rankin Scale (mRS) for neurologic disability were recorded respectively at discharge. Patients were divided into a deceased (mRS = 6) and a survival group (mRS = 1 to 5), and a favorable (mRS: 1 to 2) and unfavorable outcome group (mRS: 3 ~ 6). MAIN RESULTS: This study identified a causal relationship from EEG→FV, indicating EEG contained information that can be used for FV prediction. PGCI was negatively related with mRS (p < 0.05), indicating that stronger causalities between EEG and FV exist in patients with better outcome. The NIHSS was negatively related with the asymmetry of the two-side PGCI, calculated as the difference between the lesional side and non-lesional side PGCI. SIGNIFICANCE: A causal relationship from EEG→FV may exist in patients with ischemic stroke. The strength of G-causality may be related to stroke severity at discharge.

Photoplethysmography based atrial fibrillation detection: a review.

Atrial fibrillation (AF) is a cardiac rhythm disorder associated with increased morbidity and mortality. It is the leading risk factor for cardioembolic stroke and its early detection is crucial in both primary and secondary stroke prevention. Continuous monitoring of cardiac rhythm is today possible thanks to consumer-grade wearable devices, enabling transformative diagnostic and patient management tools. Such monitoring is possible using low-cost easy-to-implement optical sensors that today equip the majority of wearables. These sensors record blood volume variations-a technology known as photoplethysmography (PPG)-from which the heart rate and other physiological parameters can be extracted to inform about user activity, fitness, sleep, and health. Recently, new wearable devices were introduced as being capable of AF detection, evidenced by large prospective trials in some cases. Such devices would allow for early screening of AF and initiation of therapy to prevent stroke. This review is a summary of a body of work on AF detection using PPG. A thorough account of the signal processing, machine learning, and deep learning approaches used in these studies is presented, followed by a discussion of their limitations and challenges towards clinical applications.

A Supervised Approach to Robust Photoplethysmography Quality Assessment.

Early detection of Atrial Fibrillation (AFib) is crucial to prevent stroke recurrence. New tools for monitoring cardiac rhythm are important for risk stratification and stroke prevention. As many of new approaches to long-term AFib detection are now based on photoplethysmogram (PPG) recordings from wearable devices, ensuring high PPG signal-to-noise ratios is a fundamental requirement for a robust detection of AFib episodes. Traditionally, signal quality assessment is often based on the evaluation of similarity between pulses to derive signal quality indices. There are limitations to using this approach for accurate assessment of PPG quality in the presence of arrhythmia, as in the case of AFib, mainly due to substantial changes in pulse morphology. In this paper, we first tested the performance of algorithms selected from a body of studies on PPG quality assessment using a dataset of PPG recordings from patients with AFib. We then propose machine learning approaches for PPG quality assessment in 30-s segments of PPG recording from 13 stroke patients admitted to the University of California San Francisco (UCSF) neuro intensive care unit and another dataset of 3764 patients from one of the five UCSF general intensive care units. We used data acquired from two systems, fingertip PPG (fPPG) from a bedside monitor system, and radial PPG (rPPG) measured using a wearable commercial wristband. We compared various supervised machine learning techniques including k-nearest neighbors, decisions trees, and a two-class support vector machine (SVM). SVM provided the best performance. fPPG signals were used to build the model and achieved 0.9477 accuracy when tested on the data from the fPPG exclusive to the test set, and 0.9589 accuracy when tested on the rPPG data.

Deep learning approaches for plethysmography signal quality assessment in the presence of atrial fibrillation.

OBJECTIVE: Photoplethysmography (PPG) monitoring has been implemented in many portable and wearable devices we use daily for health and fitness tracking. Its simplicity and cost-effectiveness has enabled a variety of biomedical applications, such as continuous long-term monitoring of heart arrhythmias, fitness, and sleep tracking, and hydration monitoring. One major issue that can hinder PPG-based applications is movement artifacts, which can lead to false interpretations. In many implementations, noisy PPG signals are discarded. Misinterpreted or discarded PPG signals pose a problem in applications where the goal is to increase the yield of detecting physiological events, such as in the case of paroxysmal atrial fibrillation (AF)-a common episodic heart arrhythmia and a leading risk factor for stroke. In this work, we compared a traditional machine learning and deep learning approaches for PPG quality assessment in the presence of AF, in order to find the most robust method for PPG quality assessment. APPROACH: The training data set was composed of 78 278 30 s long PPG recordings from 3764 patients using bedside patient monitors. Two different representations of PPG signals were employed-a time-series based (1D) one and an image-based (2D) one. Trained models were tested on an independent set of 2683 30 s PPG signals from 13 stroke patients. MAIN RESULTS: ResNet18 showed a higher performance (0.985 accuracy, 0.979 specificity, and 0.988 sensitivity) than SVM and other deep learning approaches. 2D-based models were generally more accurate than 1D-based models. SIGNIFICANCE: 2D representation of PPG signal enhances the accuracy of PPG signal quality assessment.

Continuous monitoring of cerebrovascular reactivity through pulse transit time and intracranial pressure.

OBJECTIVE: Cerebrovascular reactivity (CR) is a mechanism that maintains stable blood flow supply to the brain. Pressure reactivity index (PRx), the correlation coefficient between slow waves of invasive arterial blood pressure (ABP) and intracranial pressure (ICP) has been validated for CR assessment. However, in clinical ward, not every subarachnoid hemorrhage (SAH) patient has invasive ABP monitoring. Pulse transit time (PTT), the propagation time of a pulse wave travelling from the heart to peripheral arteries, has been suggested as a surrogate measure of ABP. In this study, we proposed to use PTT instead of invasive ABP to monitor CR. APPROACH: Forty-five SAH patients with simultaneous recordings of invasive ABP, ICP, oxygen saturation level (SpO2) and electrocardiograph (ECG) were included. PTT was calculated as the time from the ECG R-wave peak to the onset of SpO2. PTT based pressure reactivity index (tPRx) was calculated as the correlation coefficient between slow waves of PTT and ICP. Wavelet tPRx (wtRx) was calculated as the cosine of wavelet phase shift between PTT and ICP. Meanwhile, PRx and wPRx were also calculated using invasive ABP and ICP as input. MAIN RESULTS: The result showed a negative relationship between PTT and ABP (r = -0.58, p  < 0.001). tPRx negatively correlated with PRx (r = -0.51, p  = 0.003). Wavelet method correlated well with correlation method demonstrated through positive relationship between wPRx and PRx (r = 0.82, p  < 0.001) as well as wtPRx and tPRx (r = 0.84, p  < 0.001). SIGNIFICANCE: PTT demonstrates great potential as a useful tool for CR assessment when invasive ABP is unavailable. Key points • Pulse transit time (PTT), defined as the propagation time of a pulse wave travelling from the heart to the peripheral arteries, has been proposed as a surrogate measure of ABP. The relationship between PTT and ABP in SAH patients remains unknown. • Cerebrovascular reactivity (CR) assessment through PTT has advantages over invasive ABP, as it avoids bleeding and infection risk, and can be used outside of the ICU. • We introduced a new method to assess CR using PTT and ICP through correlation based method and wavelet based method. • We found that beat-to-beat PTT was negatively related with invasive ABP in SAH patients. A significant linear relationship exists between PTT-based CR parameter and a well validated method, PRx. PTT demonstrates great potential as a useful tool for CR assessment when invasive ABP is unavailable in SAH patients.

Wavelet leader multifractal analysis of heart rate variability in atrial fibrillation.

BACKGROUND: Accurate and timely detection of atrial fibrillation (AF) episodes is important in primarily and secondary prevention of ischemic stroke and heart-related problems. In this work, heart rate regularity of ECG inter-beat intervals was investigated in episodes of AF and other rhythms using a wavelet leader based multifractal analysis. Our aim was to improve the detectability of AF episodes. METHODS: Inter-beat intervals from 25 ECG recordings available in the MIT-BIH atrial fibrillation database were analysed. Four types of annotated rhythms (atrial fibrillation, atrial flutter, AV junctional rhythm, and other rhythms) were available. A wavelet leader based multifractal analysis was applied to 5 min non-overlapping windows of each recording to estimate the multifractal spectrum in each window. The width of the multifractal spectrum was analysed for its discrimination power between rhythm episodes. RESULTS: In 10 of 25 recordings, the width of multifractal spectrum was significantly lower in episodes of AF than in other rhythms indicating increased regularity during AF. High classification accuracy (95%) of AF episodes was achieved using a combination of features derived from the multifractal analysis and statistical central moment features. CONCLUSIONS: An increase in the regularity of inter-beat intervals was observed during AF episodes by means of multifractal analysis. Multifractal features may be used to improve AF detection accuracy.

Seizure prediction for therapeutic devices: A review.

Research in seizure prediction has come a long way since its debut almost 4 decades ago. Early studies suffered methodological caveats leading to overoptimistic results and lack of statistical significance. The publication of guidelines addressing mainly the question of performance evaluation and statistical validation in seizure prediction helped revising the status of the field. While many studies failed to prove that above chance prediction is possible by applying these guidelines, other studies were successful. Methods based on EEG analysis using linear and nonlinear measures were reportedly successful in detecting preictal changes and using them to predict seizures above chance. In this review, we present a selection of studies in seizure prediction published in the last decade. The studies were selected based on the validity of the methods and the statistical significance of performance results. These results varied between studies and many showed acceptable levels of sensitivity and specificity that could be appealing for therapeutic devices. The relatively large prediction horizon and early preictal changes reported in most studies suggest that seizure prediction may work better in closed loop seizure control devices rather than as seizure advisory devices. The emergence of a large database of annotated long-term EEG recordings should help prospective assessment of prediction methods. Some questions remain to be addressed before large clinical trials involving seizure prediction can be carried out.

Scale invariance properties of intracerebral EEG improve seizure prediction in mesial temporal lobe epilepsy.

Although treatment for epilepsy is available and effective for nearly 70 percent of patients, many remain in need of new therapeutic approaches. Predicting the impending seizures in these patients could significantly enhance their quality of life if the prediction performance is clinically practical. In this study, we investigate the improvement of the performance of a seizure prediction algorithm in 17 patients with mesial temporal lobe epilepsy by means of a novel measure. Scale-free dynamics of the intracerebral EEG are quantified through robust estimates of the scaling exponents--the first cumulants--derived from a wavelet leader and bootstrap based multifractal analysis. The cumulants are investigated for the discriminability between preictal and interictal epochs. The performance of our recently published patient-specific seizure prediction algorithm is then out-of-sample tested on long-lasting data using combinations of cumulants and state similarity measures previously introduced. By using the first cumulant in combination with state similarity measures, up to 13 of 17 patients had seizures predicted above chance with clinically practical levels of sensitivity (80.5%) and specificity (25.1% of total time under warning) for prediction horizons above 25 min. These results indicate that the scale-free dynamics of the preictal state are different from those of the interictal state. Quantifiers of these dynamics may carry a predictive power that can be used to improve seizure prediction performance.

Seizure prediction in patients with mesial temporal lobe epilepsy using EEG measures of state similarity.

OBJECTIVES: In patients with intractable epilepsy, predicting seizures above chance and with clinically acceptable performance has yet to be demonstrated. In this study, an intracranial EEG-based seizure prediction method using measures of similarity with a reference state is proposed. METHODS: 1565 h of continuous intracranial EEG data from 17 patients with mesial temporal lobe epilepsy were investigated. The recordings included 175 seizures. In each patient the data was split into a training set and a testing set. EEG segments were analyzed using continuous wavelet transform. During training, a reference state was defined in the immediate preictal data and used to derive three features quantifying the discrimination between preictal and interictal states. A classifier was then trained in the feature space. Its performance was assessed using testing set and compared with a random predictor for statistical validation. RESULTS: Better than random prediction performance was achieved in 7 patients. The sensitivity was higher than 85%, the warning rate was less than 0.35/h and the proportion of time under warning was less than 30%. CONCLUSION: Seizures are predicted above chance in 41% of patients using measures of state similarity. SIGNIFICANCE: Sensitivity and specificity levels are potentially interesting for closed-loop seizure control applications.

Discriminating preictal and interictal states in patients with temporal lobe epilepsy using wavelet analysis of intracerebral EEG.

OBJECTIVE: Identification of consistent distinguishing features between preictal and interictal periods in the EEG is an essential step towards performing seizure prediction. We propose a novel method to separate preictal and interictal states based on the analysis of the high frequency activity of intracerebral EEGs in patients with mesial temporal lobe epilepsy. METHODS: Wavelet energy and entropy were computed in sliding window fashion from preictal and interictal epochs. A comparison of their organization in a 2 dimensional space was carried out using three features quantifying the similarities between their underlying states and a reference state. A discriminant analysis was then used in the features space to classify epochs. Performance was assessed based on sensitivity and false positive rates and validation was performed using a bootstrapping approach. RESULTS: Preictal and interictal epochs were discriminable in most patients on a subset of channels that were found to be close or within the seizure onset zone. CONCLUSIONS: Preictal and interictal states were separable using measures of similarity with the reference state. Discriminability varies with frequency bands. SIGNIFICANCE: This method is useful to discriminate preictal from interictal states in intracerebral EEGs and could be useful for seizure prediction.