Optimizing warfarin dosing for patients with atrial fibrillation using machine learning.

While novel oral anticoagulants are increasingly used to reduce risk of stroke in patients with atrial fibrillation, vitamin K antagonists such as warfarin continue to be used extensively for stroke prevention across the world. While effective in reducing the risk of strokes, the complex pharmacodynamics of warfarin make it difficult to use clinically, with many patients experiencing under- and/or over- anticoagulation. In this study we employed a novel implementation of deep reinforcement learning to provide clinical decision support to optimize time in therapeutic International Normalized Ratio (INR) range. We used a novel semi-Markov decision process formulation of the Batch-Constrained deep Q-learning algorithm to develop a reinforcement learning model to dynamically recommend optimal warfarin dosing to achieve INR of 2.0-3.0 for patients with atrial fibrillation. The model was developed using data from 22,502 patients in the warfarin treated groups of the pivotal randomized clinical trials of edoxaban (ENGAGE AF-TIMI 48), apixaban (ARISTOTLE) and rivaroxaban (ROCKET AF). The model was externally validated on data from 5730 warfarin-treated patients in a fourth trial of dabigatran (RE-LY) using multilevel regression models to estimate the relationship between center-level algorithm consistent dosing, time in therapeutic INR range (TTR), and a composite clinical outcome of stroke, systemic embolism or major hemorrhage. External validation showed a positive association between center-level algorithm-consistent dosing and TTR (R2 = 0.56). Each 10% increase in algorithm-consistent dosing at the center level independently predicted a 6.78% improvement in TTR (95% CI 6.29, 7.28; p < 0.001) and a 11% decrease in the composite clinical outcome (HR 0.89; 95% CI 0.81, 1.00; p = 0.015). These results were comparable to those of a rules-based clinical algorithm used for benchmarking, for which each 10% increase in algorithm-consistent dosing independently predicted a 6.10% increase in TTR (95% CI 5.67, 6.54, p < 0.001) and a 10% decrease in the composite outcome (HR 0.90; 95% CI 0.83, 0.98, p = 0.018). Our findings suggest that a deep reinforcement learning algorithm can optimize time in therapeutic range for patients taking warfarin. A digital clinical decision support system to promote algorithm-consistent warfarin dosing could optimize time in therapeutic range and improve clinical outcomes in atrial fibrillation globally.

Machine Learning Approaches for the Image-Based Identification of Surgical Wound Infections: Scoping Review.

BACKGROUND: Surgical site infections (SSIs) occur frequently and impact patients and health care systems. Remote surveillance of surgical wounds is currently limited by the need for manual assessment by clinicians. Machine learning (ML)-based methods have recently been used to address various aspects of the postoperative wound healing process and may be used to improve the scalability and cost-effectiveness of remote surgical wound assessment. OBJECTIVE: The objective of this review was to provide an overview of the ML methods that have been used to identify surgical wound infections from images. METHODS: We conducted a scoping review of ML approaches for visual detection of SSIs following the JBI (Joanna Briggs Institute) methodology. Reports of participants in any postoperative context focusing on identification of surgical wound infections were included. Studies that did not address SSI identification, surgical wounds, or did not use image or video data were excluded. We searched MEDLINE, Embase, CINAHL, CENTRAL, Web of Science Core Collection, IEEE Xplore, Compendex, and arXiv for relevant studies in November 2022. The records retrieved were double screened for eligibility. A data extraction tool was used to chart the relevant data, which was described narratively and presented using tables. Employment of TRIPOD (Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis) guidelines was evaluated and PROBAST (Prediction Model Risk of Bias Assessment Tool) was used to assess risk of bias (RoB). RESULTS: In total, 10 of the 715 unique records screened met the eligibility criteria. In these studies, the clinical contexts and surgical procedures were diverse. All papers developed diagnostic models, though none performed external validation. Both traditional ML and deep learning methods were used to identify SSIs from mostly color images, and the volume of images used ranged from under 50 to thousands. Further, 10 TRIPOD items were reported in at least 4 studies, though 15 items were reported in fewer than 4 studies. PROBAST assessment led to 9 studies being identified as having an overall high RoB, with 1 study having overall unclear RoB. CONCLUSIONS: Research on the image-based identification of surgical wound infections using ML remains novel, and there is a need for standardized reporting. Limitations related to variability in image capture, model building, and data sources should be addressed in the future.

A versatile, fast and unbiased method for estimation of gene-by-environment interaction effects on biobank-scale datasets.

Identification of gene-by-environment interactions (GxE) is crucial to understand the interplay of environmental effects on complex traits. However, current methods evaluating GxE on biobank-scale datasets have limitations. We introduce MonsterLM, a multiple linear regression method that does not rely on model specification and provides unbiased estimates of variance explained by GxE. We demonstrate robustness of MonsterLM through comprehensive genome-wide simulations using real genetic data from 325,989 individuals. We estimate GxE using waist-to-hip-ratio, smoking, and exercise as the environmental variables on 13 outcomes (N = 297,529-325,989) in the UK Biobank. GxE variance is significant for 8 environment-outcome pairs, ranging from 0.009 - 0.071. The majority of GxE variance involves SNPs without strong marginal or interaction associations. We observe modest improvements in polygenic score prediction when incorporating GxE. Our results imply a significant contribution of GxE to complex trait variance and we show MonsterLM to be well-purposed to handle this with biobank-scale data.

Machine learning for detecting centre-level irregularities in randomized controlled trials: A pilot study.

Centralized statistical monitoring is sometimes employed as an alternative to onsite monitoring for randomized control trials. Current central monitoring methods have limitations, in that they are relatively resource intensive and do not necessarily generalize to studies where an irregularity pattern has not been observed before. Machine learning has been effective in detecting irregularities in industries such as finance and manufacturing, but to date none have been applied to clinical trials. We conducted a pilot study for the use of machine learning to identify center-level irregularities in data from multicenter clinical trials. We employed unsupervised machine learning methods, which do not rely on labelled data, and therefore allow for the automated discovery of previously unseen irregularity patterns while maintaining flexibility when applied to new data with different structures. This pilot study employs unsupervised machine learning to compute distance matrices between centres, which we used to produce centre-level continuous features. We then used a one-class support vector machine to learn the underlying distribution of each data set to identify data that was substantially different from these distributions. We evaluated our approach against current automatable centralized monitoring methods on two trials with known irregularities. While current approaches performed well on one trial (AUROC 0.752 for monitoring vs. 0.584 for machine learning), our techniques performed substantially better on the other (AUROC 0.140 for monitoring vs 0.728 for machine learning). The results of this pilot study suggest both the feasibility and the potential value of a machine learning-based approach to irregularity detection in RCTs.

Optimizing Health Coaching for Patients With Type 2 Diabetes Using Machine Learning: Model Development and Validation Study.

BACKGROUND: Health coaching is an emerging intervention that has been shown to improve clinical and patient-relevant outcomes for type 2 diabetes. Advances in artificial intelligence may provide an avenue for developing a more personalized, adaptive, and cost-effective approach to diabetes health coaching. OBJECTIVE: We aim to apply Q-learning, a widely used reinforcement learning algorithm, to a diabetes health-coaching data set to develop a model for recommending an optimal coaching intervention at each decision point that is tailored to a patient's accumulated history. METHODS: In this pilot study, we fit a two-stage reinforcement learning model on 177 patients from the intervention arm of a community-based randomized controlled trial conducted in Canada. The policy produced by the reinforcement learning model can recommend a coaching intervention at each decision point that is tailored to a patient's accumulated history and is expected to maximize the composite clinical outcome of hemoglobin A1c reduction and quality of life improvement (normalized to [ ​0, 1 ​], with a higher score being better). Our data, models, and source code are publicly available. RESULTS: Among the 177 patients, the coaching intervention recommended by our policy mirrored the observed diabetes health coach's interventions in 17.5% (n=31) of the patients in stage 1 and 14.1% (n=25) of the patients in stage 2. Where there was agreement in both stages, the average cumulative composite outcome (0.839, 95% CI 0.460-1.220) was better than those for whom the optimal policy agreed with the diabetes health coach in only one stage (0.791, 95% CI 0.747-0.836) or differed in both stages (0.755, 95% CI 0.728-0.781). Additionally, the average cumulative composite outcome predicted for the policy's recommendations was significantly better than that of the observed diabetes health coach's recommendations (tn-1=10.040; P<.001). CONCLUSIONS: Applying reinforcement learning to diabetes health coaching could allow for both the automation of health coaching and an improvement in health outcomes produced by this type of intervention.

A machine learning-based clinical decision support algorithm for reducing unnecessary coronary angiograms.

Background: Conventional clinical risk scores and diagnostic algorithms are proving to be suboptimal in the prediction of obstructive coronary artery disease, contributing to the low diagnostic yield of invasive angiography. Machine learning could help better predict which patients would benefit from invasive angiography vs other noninvasive diagnostic modalities. Objective: To reduce patient risk and cost to the healthcare system by improving the diagnostic yield of invasive coronary angiography through optimized outpatient selection. Methods: Retrospective analysis of 12 years of referral data from a provincial cardiac registry, including all patients referred for invasive angiography of more than 1.4 million individuals in Ontario, Canada. Stable outpatients undergoing coronary angiography during the study period were included in the analysis. The training set (80% random sample, n = 23,750) was used to develop 8 prediction models in Python using grid-search cross-validation. The test set (20% random sample, n = 5938), evaluated the discrimination performance of each model. Results: The machine-learning model achieved a substantially better performance (area under the receiver operating characteristics curve: 0.81) than existing models for predicting obstructive coronary artery disease in patients referred for invasive angiography. It significantly outperformed both the reference model and current clinical practice with a net reclassification index of 27.8% (95% confidence interval [CI]: [24.9%-30.8%], P value <.01) and 44.7% (95% CI: [42.4%-47.0%], P value <.01), respectively. Conclusion: This prediction model, when coupled with a point-of-care, online decision support tool to be used by referring physicians, could improve the diagnostic yield of invasive coronary angiography in stable, elective outpatients, thus improving patient safety and reducing healthcare costs.

Opening the Black Box: The Promise and Limitations of Explainable Machine Learning in Cardiology.

Many clinicians remain wary of machine learning because of longstanding concerns about "black box" models. "Black box" is shorthand for models that are sufficiently complex that they are not straightforwardly interpretable to humans. Lack of interpretability in predictive models can undermine trust in those models, especially in health care, in which so many decisions are- literally-life and death issues. There has been a recent explosion of research in the field of explainable machine learning aimed at addressing these concerns. The promise of explainable machine learning is considerable, but it is important for cardiologists who may encounter these techniques in clinical decision-support tools or novel research papers to have critical understanding of both their strengths and their limitations. This paper reviews key concepts and techniques in the field of explainable machine learning as they apply to cardiology. Key concepts reviewed include interpretability vs explainability and global vs local explanations. Techniques demonstrated include permutation importance, surrogate decision trees, local interpretable model-agnostic explanations, and partial dependence plots. We discuss several limitations with explainability techniques, focusing on the how the nature of explanations as approximations may omit important information about how black-box models work and why they make certain predictions. We conclude by proposing a rule of thumb about when it is appropriate to use black- box models with explanations rather than interpretable models.

Machine Learning-Based Early Warning Systems for Clinical Deterioration: Systematic Scoping Review.

BACKGROUND: Timely identification of patients at a high risk of clinical deterioration is key to prioritizing care, allocating resources effectively, and preventing adverse outcomes. Vital signs-based, aggregate-weighted early warning systems are commonly used to predict the risk of outcomes related to cardiorespiratory instability and sepsis, which are strong predictors of poor outcomes and mortality. Machine learning models, which can incorporate trends and capture relationships among parameters that aggregate-weighted models cannot, have recently been showing promising results. OBJECTIVE: This study aimed to identify, summarize, and evaluate the available research, current state of utility, and challenges with machine learning-based early warning systems using vital signs to predict the risk of physiological deterioration in acutely ill patients, across acute and ambulatory care settings. METHODS: PubMed, CINAHL, Cochrane Library, Web of Science, Embase, and Google Scholar were searched for peer-reviewed, original studies with keywords related to "vital signs," "clinical deterioration," and "machine learning." Included studies used patient vital signs along with demographics and described a machine learning model for predicting an outcome in acute and ambulatory care settings. Data were extracted following PRISMA, TRIPOD, and Cochrane Collaboration guidelines. RESULTS: We identified 24 peer-reviewed studies from 417 articles for inclusion; 23 studies were retrospective, while 1 was prospective in nature. Care settings included general wards, intensive care units, emergency departments, step-down units, medical assessment units, postanesthetic wards, and home care. Machine learning models including logistic regression, tree-based methods, kernel-based methods, and neural networks were most commonly used to predict the risk of deterioration. The area under the curve for models ranged from 0.57 to 0.97. CONCLUSIONS: In studies that compared performance, reported results suggest that machine learning-based early warning systems can achieve greater accuracy than aggregate-weighted early warning systems but several areas for further research were identified. While these models have the potential to provide clinical decision support, there is a need for standardized outcome measures to allow for rigorous evaluation of performance across models. Further research needs to address the interpretability of model outputs by clinicians, clinical efficacy of these systems through prospective study design, and their potential impact in different clinical settings.

Optimal thresholds for ultrasound attenuation parameter in the evaluation of hepatic steatosis severity: evidence from a cohort of patients with biopsy-proven fatty liver disease.

OBJECTIVE: FibroTouch is a newly developed device to assess ultrasound attenuation parameter (UAP) and liver stiffness measurement to quantify hepatic steatosis and fibrosis, respectively. However, there is currently a lack of defined thresholds of UAP to diagnose different stages of hepatic steatosis. We aimed to assess the optimal thresholds of UAP for hepatic steatosis in individuals with biopsy-proven fatty liver disease (FLD). METHODS: We enrolled 497 adults with FLD undergoing FibroTouch and liver biopsy. Area under the receiver operating characteristic curve (AUROC) was performed to calculate the performance of UAP in staging hepatic steatosis. Hepatic steatosis >33% was defined as significant steatosis. We determined the optimal cutoff values of UAP and the sensitivity or specificity higher than 90%. Sensitivity, specificity, positive predictive value and negative predictive value were subsequently calculated. RESULTS: The median UAP for the enrolled patients was 308 dB/m. Multivariable logistic regression analysis showed that UAP was associated with significant steatosis [adjusted-odds ratio 1.05, 95% confidence interval (CI), 1.02-1.09; P = 0.001]. The AUROCs for S ≥ 1, S ≥ 2 and S = 3 were 0.88 (95% CI, 0.84-0.91), 0.77 (95% CI, 0.73-0.81), and 0.70 (95% CI, 0.63-0.77), respectively. The optimal UAP cutoffs were 295 dB/m for S ≥ 1, 314 dB/m for S ≥ 2, and 324 dB/m for S = 3. Almost identical results were observed in the subgroup of patients with biopsy-confirmed nonalcoholic fatty liver disease (n = 435). CONCLUSION: We found that the AUROC values of UAP by FibroTouch were ranging from 0.70 to 0.88 for assessing hepatic steatosis severity. These UAP cutoffs could be applicable for clinical use.

Association between Hemoglobin Glycation Index and NAFLD in Chinese Nondiabetic Individuals.

Purpose: Limited studies have preliminarily identified a positive association between nonalcoholic fatty liver disease (NAFLD) and hemoglobin glycation index (HGI). However, this association has not been fully established. We aim to investigate the association between NAFLD and HGI in Chinese nondiabetic individuals and to construct a risk score based on HGI to predict a person's risk of NAFLD. Methods: After strict exclusion criteria, 5,903 individuals were included in this retrospective cross-sectional study. We randomly selected 1,967 subjects in the enrollment to obtain an equation of linear regression, which was used to calculate predicted HbA1c and drive HGI. The other subjects were classified into four categories according to HGI level (≤-0.22, -0.21∼0.02, 0.03∼0.28, and ≥0.29). All subjects retrospectively reviewed the baseline characteristics, laboratory examinations, and abdominal ultrasonography. Results: The prevalence of NAFLD in this population was 20.7%, which increases along with the growth of HGI levels (P < 0.001). Adjusted to multiple factors, this trend still remained significant (OR: 1.172 (95% CI, 1.074-1.279)). The combined NAFLD risk score based on HGI resulted in an area under the receiver operator characteristic curve (AUROC) of 0.85 provided sensitivity, specificity, positive predictive value, and a negative predictive value for NAFLD of 84.4%, 71.3%, 65.0%, and 88.0%, respectively. Conclusions: NAFLD is independently associated with HGI levels in Chinese nondiabetic individuals. And, NAFLD risk score may be used as one of the risk predictors of NAFLD in nondiabetic population.