Identifying risk of opioid use disorder for patients taking opioid medications with deep learning

ObjectiveThe United States is experiencing an opioid epidemic. In recent years, there were more than 10 million opioid misusers aged 12 years or older annually. Identifying patients at high risk of opioid use disorder (OUD) can help to make early clinical interventions to reduce the risk of OUD. Our goal is to develop and evaluate models to predict OUD for patients on opioid medications using electronic health records and deep learning methods. The resulting models help us to better understand OUD, providing new insights on the opioid epidemic. Further, these models provide a foundation for clinical tools to predict OUD before it occurs, permitting early interventions.MethodsElectronic health records of patients who have been prescribed with medications containing active opioid ingredients were extracted from Cerner’s Health Facts database for encounters between January 1, 2008, and December 31, 2017. Long short-term memory models were applied to predict OUD risk based on five recent prior encounters before the target encounter and compared with logistic regression, random forest, decision tree, and dense neural network. Prediction performance was assessed using F1 score, precision, recall, and area under the receiver-operating characteristic curve.ResultsThe long short-term memory (LSTM) model provided promising prediction results which outperformed other methods, with an F1 score of 0.8023 (about 0.016 higher than dense neural network (DNN)) and an area under the receiver-operating characteristic curve (AUROC) of 0.9369 (about 0.145 higher than DNN).ConclusionsLSTM–based sequential deep learning models can accurately predict OUD using a patient’s history of electronic health records, with minimal prior domain knowledge. This tool has the potential to improve clinical decision support for early intervention and prevention to combat the opioid epidemic.     

Bayesian logistic regression for online recalibration and revision of risk prediction models with performance guarantees

ObjectiveAfter deploying a clinical prediction model, subsequently collected data can be used to fine-tune its predictions and adapt to temporal shifts. Because model updating carries risks of over-updating/fitting, we study online methods with performance guarantees. Materials and MethodsWe introduce 2 procedures for continual recalibration or revision of an underlying prediction model: Bayesian logistic regression (BLR) and a Markov variant that explicitly models distribution shifts (MarBLR). We perform empirical evaluation via simulations and a real-world study predicting Chronic Obstructive Pulmonary Disease (COPD) risk. We derive “Type I and II” regret bounds, which guarantee the procedures are noninferior to a static model and competitive with an oracle logistic reviser in terms of the average loss.ResultsBoth procedures consistently outperformed the static model and other online logistic revision methods. In simulations, the average estimated calibration index (aECI) of the original model was 0.828 (95%CI, 0.818–0.938). Online recalibration using BLR and MarBLR improved the aECI towards the ideal value of zero, attaining 0.265 (95%CI, 0.230–0.300) and 0.241 (95%CI, 0.216–0.266), respectively. When performing more extensive logistic model revisions, BLR and MarBLR increased the average area under the receiver-operating characteristic curve (aAUC) from 0.767 (95%CI, 0.765–0.769) to 0.800 (95%CI, 0.798–0.802) and 0.799 (95%CI, 0.797–0.801), respectively, in stationary settings and protected against substantial model decay. In the COPD study, BLR and MarBLR dynamically combined the original model with a continually refitted gradient boosted tree to achieve aAUCs of 0.924 (95%CI, 0.913–0.935) and 0.925 (95%CI, 0.914–0.935), compared to the static model’s aAUC of 0.904 (95%CI, 0.892–0.916).DiscussionDespite its simplicity, BLR is highly competitive with MarBLR. MarBLR outperforms BLR when its prior better reflects the data.ConclusionsBLR and MarBLR can improve the transportability of clinical prediction models and maintain their performance over time.     

Importance-aware personalized learning for early risk prediction using static and dynamic health data

ObjectiveAccurate risk prediction is important for evaluating early medical treatment effects and improving health care quality. Existing methods are usually designed for dynamic medical data, which require long-term observations. Meanwhile, important personalized static information is ignored due to the underlying uncertainty and unquantifiable ambiguity. It is urgent to develop an early risk prediction method that can adaptively integrate both static and dynamic health data.Materials and MethodsData were from 6367 patients with Peptic Ulcer Bleeding between 2007 and 2016. This article develops a novel End-to-end Importance-Aware Personalized Deep Learning Approach (eiPDLA) to achieve accurate early clinical risk prediction. Specifically, eiPDLA introduces a long short-term memory with temporal attention to learn sequential dependencies from time-stamped records and simultaneously incorporating a residual network with correlation attention to capture their influencing relationship with static medical data. Furthermore, a new multi-residual multi-scale network with the importance-aware mechanism is designed to adaptively fuse the learned multisource features, automatically assigning larger weights to important features while weakening the influence of less important features.ResultsExtensive experimental results on a real-world dataset illustrate that our method significantly outperforms the state-of-the-arts for early risk prediction under various settings (eg, achieving an AUC score of 0.944 at 1 year ahead of risk prediction). Case studies indicate that the achieved prediction results are highly interpretable.ConclusionThese results reflect the importance of combining static and dynamic health data, mining their influencing relationship, and incorporating the importance-aware mechanism to automatically identify important features. The achieved accurate early risk prediction results save precious time for doctors to timely design effective treatments and improve clinical outcomes.     

可解释机器学习方法在疾病预测中的应用:脓毒血症患者死亡风险研究

目的 探索可解释机器学习方法在疾病预测中的应用。方法 本研究以脓毒血症死亡风险预测为例,从重症监护医学数据库(Medical Information Mart for Intensive Care, MIMIC)-Ⅳ中采集符合纳排标准的19 903例脓毒血症(sepsis-3)患者的临床数据,利用决策树、逻辑回归、随机森林、XGBoost、轻量梯度提升机(light gradient boosting machine,LightGBM)模型分别构建脓毒血症死亡预测模型。在此基础上,利用全局可解释方法(特征重要性、部分依赖图、个体条件期望、全局代理模型)和局部可解释方法(局部代理模型和Shapely值)对复杂机器学习模型进行解释,探索影响脓毒血症患者预后的危险因素。结果 解释性差的机器学习模型的预测性能[模型LightGBM、随机森林、XGBoost的曲线下面积(area under curve,AUC)值分别为0.913、0.892、0.872]高于具有内在解释性的模型(逻辑回归模型AUC=0.779,决策树模型AUC=0.791),并利用全局解释性方法、局部可解释性方法两种类型的解释方法对机器学习模型决策过程进行解释。结论 利用全局解释性方法可以解释在整个特征空间内机器学习模型的响应趋势,利用局部可解释性方法可以解释机器学习模型对特定病例的决策过程。     

Application of Bayesian networks to generate synthetic health data

ObjectiveThis study seeks to develop a fully automated method of generating synthetic data from a real dataset that could be employed by medical organizations to distribute health data to researchers, reducing the need for access to real data. We hypothesize the application of Bayesian networks will improve upon the predominant existing method, medBGAN, in handling the complexity and dimensionality of healthcare data.Materials and MethodsWe employed Bayesian networks to learn probabilistic graphical structures and simulated synthetic patient records from the learned structure. We used the University of California Irvine (UCI) heart disease and diabetes datasets as well as the MIMIC-III diagnoses database. We evaluated our method through statistical tests, machine learning tasks, preservation of rare events, disclosure risk, and the ability of a machine learning classifier to discriminate between the real and synthetic data.ResultsOur Bayesian network model outperformed or equaled medBGAN in all key metrics. Notable improvement was achieved in capturing rare variables and preserving association rules.DiscussionBayesian networks generated data sufficiently similar to the original data with minimal risk of disclosure, while offering additional transparency, computational efficiency, and capacity to handle more data types in comparison to existing methods. We hope this method will allow healthcare organizations to efficiently disseminate synthetic health data to researchers, enabling them to generate hypotheses and develop analytical tools. ConclusionWe conclude the application of Bayesian networks is a promising option for generating realistic synthetic health data that preserves the features of the original data without compromising data privacy.     

A framework for employing longitudinally collected multicenter electronic health records to stratify heterogeneous patient populations on disease history

ObjectiveTo facilitate patient disease subset and risk factor identification by constructing a pipeline which is generalizable, provides easily interpretable results, and allows replication by overcoming electronic health records (EHRs) batch effects.Material and MethodsWe used 1872 billing codes in EHRs of 102 880 patients from 12 healthcare systems. Using tools borrowed from single-cell omics, we mitigated center-specific batch effects and performed clustering to identify patients with highly similar medical history patterns across the various centers. Our visualization method (PheSpec) depicts the phenotypic profile of clusters, applies a novel filtering of noninformative codes (Ranked Scope Pervasion), and indicates the most distinguishing features.ResultsWe observed 114 clinically meaningful profiles, for example, linking prostate hyperplasia with cancer and diabetes with cardiovascular problems and grouping pediatric developmental disorders. Our framework identified disease subsets, exemplified by 6 “other headache” clusters, where phenotypic profiles suggested different underlying mechanisms: migraine, convulsion, injury, eye problems, joint pain, and pituitary gland disorders. Phenotypic patterns replicated well, with high correlations of ≥0.75 to an average of 6 (2–8) of the 12 different cohorts, demonstrating the consistency with which our method discovers disease history profiles.DiscussionCostly clinical research ventures should be based on solid hypotheses. We repurpose methods from single-cell omics to build these hypotheses from observational EHR data, distilling useful information from complex data.ConclusionWe establish a generalizable pipeline for the identification and replication of clinically meaningful (sub)phenotypes from widely available high-dimensional billing codes. This approach overcomes datatype problems and produces comprehensive visualizations of validation-ready phenotypes.     

Development and validation of prediction models for mechanical ventilation,renal replacement therapy,and readmission in COVID-19 patients

ObjectiveCoronavirus disease 2019 (COVID-19) patients are at risk for resource-intensive outcomes including mechanical ventilation (MV), renal replacement therapy (RRT), and readmission. Accurate outcome prognostication could facilitate hospital resource allocation. We develop and validate predictive models for each outcome using retrospective electronic health record data for COVID-19 patients treated between March 2 and May 6, 2020.Materials and MethodsFor each outcome, we trained 3 classes of prediction models using clinical data for a cohort of SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2)–positive patients (n = 2256). Cross-validation was used to select the best-performing models per the areas under the receiver-operating characteristic and precision-recall curves. Models were validated using a held-out cohort (n = 855). We measured each model’s calibration and evaluated feature importances to interpret model output.ResultsThe predictive performance for our selected models on the held-out cohort was as follows: area under the receiver-operating characteristic curve—MV 0.743 (95% CI, 0.682-0.812), RRT 0.847 (95% CI, 0.772-0.936), readmission 0.871 (95% CI, 0.830-0.917); area under the precision-recall curve—MV 0.137 (95% CI, 0.047-0.175), RRT 0.325 (95% CI, 0.117-0.497), readmission 0.504 (95% CI, 0.388-0.604). Predictions were well calibrated, and the most important features within each model were consistent with clinical intuition.DiscussionOur models produce performant, well-calibrated, and interpretable predictions for COVID-19 patients at risk for the target outcomes. They demonstrate the potential to accurately estimate outcome prognosis in resource-constrained care sites managing COVID-19 patients.ConclusionsWe develop and validate prognostic models targeting MV, RRT, and readmission for hospitalized COVID-19 patients which produce accurate, interpretable predictions. Additional external validation studies are needed to further verify the generalizability of our results.     

Using nursing notes to improve clinical outcome prediction in intensive care patients: A retrospective cohort study

ObjectiveElectronic health record documentation by intensive care unit (ICU) clinicians may predict patient outcomes. However, it is unclear whether physician and nursing notes differ in their ability to predict short-term ICU prognosis. We aimed to investigate and compare the ability of physician and nursing notes, written in the first 48 hours of admission, to predict ICU length of stay and mortality using 3 analytical methods.Materials and MethodsThis was a retrospective cohort study with split sampling for model training and testing. We included patients ≥18 years of age admitted to the ICU at Beth Israel Deaconess Medical Center in Boston, Massachusetts, from 2008 to 2012. Physician or nursing notes generated within the first 48 hours of admission were used with standard machine learning methods to predict outcomes.ResultsFor the primary outcome of composite score of ICU length of stay ≥7 days or in-hospital mortality, the gradient boosting model had better performance than the logistic regression and random forest models. Nursing and physician notes achieved area under the curves (AUCs) of 0.826 and 0.796, respectively, with even better predictive power when combined (AUC, 0.839).DiscussionModels using only nursing notes more accurately predicted short-term prognosis than did models using only physician notes, but in combination, the models achieved the greatest accuracy in prediction. ConclusionsOur findings demonstrate that statistical models derived from text analysis in the first 48 hours of ICU admission can predict patient outcomes. Physicians’ and nurses’ notes are both uniquely important in mortality prediction and combining these notes can produce a better predictive model.     

STAN: spatio-temporal attention network for pandemic prediction using real-world evidence

ObjectiveWe aim to develop a hybrid model for earlier and more accurate predictions for the number of infected cases in pandemics by (1) using patients’ claims data from different counties and states that capture local disease status and medical resource utilization; (2) utilizing demographic similarity and geographical proximity between locations; and (3) integrating pandemic transmission dynamics into a deep learning model.Materials and MethodsWe proposed a spatio-temporal attention network (STAN) for pandemic prediction. It uses a graph attention network to capture spatio-temporal trends of disease dynamics and to predict the number of cases for a fixed number of days into the future. We also designed a dynamics-based loss term for enhancing long-term predictions. STAN was tested using both real-world patient claims data and COVID-19 statistics over time across US counties.ResultsSTAN outperforms traditional epidemiological models such as susceptible-infectious-recovered (SIR), susceptible-exposed-infectious-recovered (SEIR), and deep learning models on both long-term and short-term predictions, achieving up to 87% reduction in mean squared error compared to the best baseline prediction model.ConclusionsBy combining information from real-world claims data and disease case counts data, STAN can better predict disease status and medical resource utilization.     

Healthcare Process Modeling to Phenotype Clinician Behaviors for Exploiting the Signal Gain of Clinical Expertise (HPM-ExpertSignals): Development and evaluation of a conceptual framework

ObjectiveThere are signals of clinicians’ expert and knowledge-driven behaviors within clinical information systems (CIS) that can be exploited to support clinical prediction. Describe development of the Healthcare Process Modeling Framework to Phenotype Clinician Behaviors for Exploiting the Signal Gain of Clinical Expertise (HPM-ExpertSignals).Materials and MethodsWe employed an iterative framework development approach that combined data-driven modeling and simulation testing to define and refine a process for phenotyping clinician behaviors. Our framework was developed and evaluated based on the Communicating Narrative Concerns Entered by Registered Nurses (CONCERN) predictive model to detect and leverage signals of clinician expertise for prediction of patient trajectories.ResultsSeven themes—identified during development and simulation testing of the CONCERN model—informed framework development. The HPM-ExpertSignals conceptual framework includes a 3-step modeling technique: (1) identify patterns of clinical behaviors from user interaction with CIS; (2) interpret patterns as proxies of an individual’s decisions, knowledge, and expertise; and (3) use patterns in predictive models for associations with outcomes. The CONCERN model differentiated at risk patients earlier than other early warning scores, lending confidence to the HPM-ExpertSignals framework.DiscussionThe HPM-ExpertSignals framework moves beyond transactional data analytics to model clinical knowledge, decision making, and CIS interactions, which can support predictive modeling with a focus on the rapid and frequent patient surveillance cycle.ConclusionsWe propose this framework as an approach to embed clinicians’ knowledge-driven behaviors in predictions and inferences to facilitate capture of healthcare processes that are activated independently, and sometimes well before, physiological changes are apparent.     

Development of a “meta-model” to address missing data,predict patient-specific cancer survival and provide a foundation for clinical decision support

ObjectiveLike most real-world data, electronic health record (EHR)–derived data from oncology patients typically exhibits wide interpatient variability in terms of available data elements. This interpatient variability leads to missing data and can present critical challenges in developing and implementing predictive models to underlie clinical decision support for patient-specific oncology care. Here, we sought to develop a novel ensemble approach to addressing missing data that we term the “meta-model” and apply the meta-model to patient-specific cancer prognosis.Materials and MethodsUsing real-world data, we developed a suite of individual random survival forest models to predict survival in patients with advanced lung cancer, colorectal cancer, and breast cancer. Individual models varied by the predictor data used. We combined models for each cancer type into a meta-model that predicted survival for each patient using a weighted mean of the individual models for which the patient had all requisite predictors.ResultsThe meta-model significantly outperformed many of the individual models and performed similarly to the best performing individual models. Comparisons of the meta-model to a more traditional imputation-based method of addressing missing data supported the meta-model’s utility.ConclusionsWe developed a novel machine learning–based strategy to underlie clinical decision support and predict survival in cancer patients, despite missing data. The meta-model may more generally provide a tool for addressing missing data across a variety of clinical prediction problems. Moreover, the meta-model may address other challenges in clinical predictive modeling including model extensibility and integration of predictive algorithms trained across different institutions and datasets.     

Dynamic modeling of hospitalized COVID-19 patients reveals disease state–dependent risk factors

ObjectiveThe study sought to investigate the disease state–dependent risk profiles of patient demographics and medical comorbidities associated with adverse outcomes of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections.Materials and MethodsA covariate-dependent, continuous-time hidden Markov model with 4 states (moderate, severe, discharged, and deceased) was used to model the dynamic progression of COVID-19 during the course of hospitalization. All model parameters were estimated using the electronic health records of 1362 patients from ProMedica Health System admitted between March 20, 2020 and December 29, 2020 with a positive nasopharyngeal PCR test for SARS-CoV-2. Demographic characteristics, comorbidities, vital signs, and laboratory test results were retrospectively evaluated to infer a patient’s clinical progression.ResultsThe association between patient-level covariates and risk of progression was found to be disease state dependent. Specifically, while being male, being Black or having a medical comorbidity were all associated with an increased risk of progressing from the moderate disease state to the severe disease state, these same factors were associated with a decreased risk of progressing from the severe disease state to the deceased state.DiscussionRecent studies have not included analyses of the temporal progression of COVID-19, making the current study a unique modeling-based approach to understand the dynamics of COVID-19 in hospitalized patients.ConclusionDynamic risk stratification models have the potential to improve clinical outcomes not only in COVID-19, but also in a myriad of other acute and chronic diseases that, to date, have largely been assessed only by static modeling techniques.     

Development of a novel machine learning model to predict presence of nonalcoholic steatohepatitis

ObjectiveTo develop a computer model to predict patients with nonalcoholic steatohepatitis (NASH) using machine learning (ML).Materials and MethodsThis retrospective study utilized two databases: a) the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) nonalcoholic fatty liver disease (NAFLD) adult database (2004-2009), and b) the Optum^® de-identified Electronic Health Record dataset (2007-2018), a real-world dataset representative of common electronic health records in the United States. We developed an ML model to predict NASH, using confirmed NASH and non-NASH based on liver histology results in the NIDDK dataset to train the model.ResultsModels were trained and tested on NIDDK NAFLD data (704 patients) and the best-performing models evaluated on Optum data (~3,000,000 patients). An eXtreme Gradient Boosting model (XGBoost) consisting of 14 features exhibited high performance as measured by area under the curve (0.82), sensitivity (81%), and precision (81%) in predicting NASH. Slightly reduced performance was observed with an abbreviated feature set of 5 variables (0.79, 80%, 80%, respectively). The full model demonstrated good performance (AUC 0.76) to predict NASH in Optum data.DiscussionThe proposed model, named NASHmap, is the first ML model developed with confirmed NASH and non-NASH cases as determined through liver biopsy and validated on a large, real-world patient dataset. Both the 14 and 5-feature versions exhibit high performance.ConclusionThe NASHmap model is a convenient and high performing tool that could be used to identify patients likely to have NASH in clinical settings, allowing better patient management and optimal allocation of clinical resources.     

Phosphate binders usage,patients knowledge,and adherence: A cross-sectional study in 4 centers at Qassim,Saudi Arabia

Objectives:To assess phosphate binders’ usage, knowledge regarding their utilization, and adherence among hemodialysis patients in Qassim, Saudi Arabia.Methods:A prospective cross-sectional study conducted at 4 dialysis centers in Qassim, Saudi Arabia with inclusion of 237 patients’ undergoing hemodialysis between November 2018 to January 2019. The study involved interviewing the patients, reviewing their medical records for biomarkers used to assess kidney function, and assessing the patients’ knowledge-based regarding dietary phosphate control, as well as adherence to phosphate binders’ usage.Results:Out of 237 included patients, male to female ratio was 54:46. The prevalence of prescribing non-calcium phosphate binders was 82.7% whereas prescribing calcium phosphate binders was 73.8%. A total of 63% of patients showed a medium level of adherence to phosphate binders. Although adherence level was not poor, therapeutic efficacy was affected by other factors such as administration time adherence positively correlated with the serum phosphate level (p=0.00).Conclusion:Phosphate binders usage is frequent among hemodialysis patients in Qassim centers. Circulating phosphate level was affected by the extent of patients’ knowledge of dietary control and adherence to the usage of phosphate binders. Thus, we recommend enhancing patient education in reference to high- and low- phosphate-rich diet to take wise dietary decisions, lower pill burden, and improve adherence toward the control of hyperphosphatemia     

Comparison of Mortality Predictive Models of Sepsis Patients Based on Machine Learning

ObjectiveTo compare the performance of five machine learning models and SAPS II score in predicting the 30-day mortality amongst patients with sepsis.MethodsThe sepsis patient-related data were extracted from the MIMIC-IV database. Clinical features were generated and selected by mutual information and grid search. Logistic regression, Random forest, LightGBM, XGBoost, and other machine learning models were constructed to predict the mortality probability. Five measurements including accuracy, precision, recall, F1 score, and area under curve (AUC) were acquired for model evaluation. An external validation was implemented to avoid conclusion bias.ResultsLightGBM outperformed other methods, achieving the highest AUC (0.900), accuracy (0.808), and precision (0.SS9). All machine learning models performed better than SAPS II score (AUC=0.748). LightGBM achieved 0.883 in AUC in the external data validation.ConclusionsThe machine learning models are more effective in predicting the 30-day mortality of patients with sepsis than the traditional SAPS II score.     

Machine-learning model to predict the cause of death using a stacking ensemble method for observational data

ObjectiveCause of death is used as an important outcome of clinical research; however, access to cause-of-death data is limited. This study aimed to develop and validate a machine-learning model that predicts the cause of death from the patient’s last medical checkup.Materials and MethodsTo classify the mortality status and each individual cause of death, we used a stacking ensemble method. The prediction outcomes were all-cause mortality, 8 leading causes of death in South Korea, and other causes. The clinical data of study populations were extracted from the national claims (n = 174 747) and electronic health records (n = 729 065) and were used for model development and external validation. Moreover, we imputed the cause of death from the data of 3 US claims databases (n = 994 518, 995 372, and 407 604, respectively). All databases were formatted to the Observational Medical Outcomes Partnership Common Data Model.ResultsThe generalized area under the receiver operating characteristic curve (AUROC) of the model predicting the cause of death within 60 days was 0.9511. Moreover, the AUROC of the external validation was 0.8887. Among the causes of death imputed in the Medicare Supplemental database, 11.32% of deaths were due to malignant neoplastic disease.DiscussionThis study showed the potential of machine-learning models as a new alternative to address the lack of access to cause-of-death data. All processes were disclosed to maintain transparency, and the model was easily applicable to other institutions.ConclusionA machine-learning model with competent performance was developed to predict cause of death.     

Bidirectional imputation of spatial GPS trajectories with missingness using sparse online Gaussian Process

ObjectiveWe propose a bidirectional GPS imputation method that can recover real-world mobility trajectories even when a substantial proportion of the data are missing. The time complexity of our online method is linear in the sample size, and it provides accurate estimates on daily or hourly summary statistics such as time spent at home and distance traveled.Materials and MethodsTo preserve a smartphone’s battery, GPS may be sampled only for a small portion of time, frequently <10%, which leads to a substantial missing data problem. We developed an algorithm that simulates an individual’s trajectory based on observed GPS location traces using sparse online Gaussian Process to addresses the high computational complexity of the existing method. The method also retains the spherical geometry of the problem, and imputes the missing trajectory in a bidirectional fashion with multiple condition checks to improve accuracy.ResultsWe demonstrated that (1) the imputed trajectories mimic the real-world trajectories, (2) the confidence intervals of summary statistics cover the ground truth in most cases, and (3) our algorithm is much faster than existing methods if we have more than 3 months of observations; (4) we also provide guidelines on optimal sampling strategies.ConclusionsOur approach outperformed existing methods and was significantly faster. It can be used in settings in which data need to be analyzed and acted on continuously, for example, to detect behavioral anomalies that might affect treatment adherence, or to learn about colocations of individuals during an epidemic.     

Incorporating real-world evidence into the development of patient blood glucose prediction algorithms for the ICU

ObjectiveGlycemic control is an important component of critical care. We present a data-driven method for predicting intensive care unit (ICU) patient response to glycemic control protocols while accounting for patient heterogeneity and variations in care.Materials and MethodsUsing electronic medical records (EMRs) of 18 961 ICU admissions from the MIMIC-III dataset, including 318 574 blood glucose measurements, we train and validate a gradient boosted tree machine learning (ML) algorithm to forecast patient blood glucose and a 95% prediction interval at 2-hour intervals. The model uses as inputs irregular multivariate time series data relating to recent in-patient medical history and glycemic control, including previous blood glucose, nutrition, and insulin dosing.ResultsOur forecasting model using routinely collected EMRs achieves performance comparable to previous models developed in planned research studies using continuous blood glucose monitoring. Model error, expressed as mean absolute percentage error is 16.5%–16.8%, with Clarke error grid analysis demonstrating that 97% of predictions would be clinically acceptable. The 95% prediction intervals achieve near intended coverage at 93%–94%.DiscussionML algorithms built on observational data sources, such as EMRs, present a promising approach for personalization and automation of glycemic control in critical care. Future research may benefit from applying a combination of methodologies and data sources to develop robust methodologies that account for the variations seen in ICU patients and difficultly in detecting the extremes of observed blood glucose values.ConclusionWe demonstrate that EMRs can be used to train ML algorithms that may be suitable for incorporation into ICU decision support systems.     

Artificial neural network for predicting pathological stage of clinically localized prostate cancer in a Taiwanese population

BackgroundWe developed an artificial neural network (ANN) model to predict prostate cancer pathological staging in patients prior to when they received radical prostatectomy as this is more effective than logistic regression (LR), or combined use of age, prostate-specific antigen (PSA), body mass index (BMI), digital rectal examination (DRE), trans-rectal ultrasound (TRUS), biopsy Gleason sum, and primary biopsy Gleason grade.MethodsOur study evaluated 299 patients undergoing retro-pubic radical prostatectomy or robotic-assisted laparoscopic radical prostatectomy surgical procedures with pelvic lymph node dissection. The results were intended to predict the pathological stage of prostate cancer (T2 or T3) after radical surgery. The predictive ability of ANN was compared with LR and validation of the 2007 Partin Tables was estimated by the areas under the receiving operating characteristic curve (AUCs).ResultsOf the 299 patients we evaluated, 109 (36.45%) displayed prostate cancer with extra-capsular extension (ECE), and 190 (63.55%) displayed organ-confined disease (OCD). LR analysis showed that only PSA and BMI were statistically significant predictors of prostate cancer with capsule invasion. Overall, ANN outperformed LR significantly (0.795 ± 0.023 versus 0.746 ± 0.025, p = 0.016). Validation using the current Partin Tables for the participants of our study was assessed, and the predictive capacity of AUC for OCD was 0.695.ConclusionANN was superior to LR at predicting OCD in prostate cancer. Compared with the validation of current Partin Tables for the Taiwanese population, the ANN model resulted in larger AUCs and more accurate prediction of the pathologic stage of prostate cancer.     

“A catalyst for action”: Factors for implementing clinical risk prediction models of infection in home care settings

ObjectiveThe study sought to outline how a clinical risk prediction model for identifying patients at risk of infection is perceived by home care nurses, and to inform how the output of the model could be integrated into a clinical workflow.Materials and MethodsThis was a qualitative study using semi-structured interviews with 50 home care nurses. Interviews explored nurses’ perceptions of clinical risk prediction models, their experiences using them in practice, and what elements are important for the implementation of a clinical risk prediction model focusing on infection. Interviews were audio-taped and transcribed, with data evaluated using thematic analysis.ResultsTwo themes were derived from the data: (1) informing nursing practice, which outlined how a clinical risk prediction model could inform nurse clinical judgment and be used to modify their care plan interventions, and (2) operationalizing the score, which summarized how the clinical risk prediction model could be incorporated in home care settings.DiscussionThe findings indicate that home care nurses would find a clinical risk prediction model for infection useful, as long as it provided both context around the reasons why a patient was deemed to be at high risk and provided some guidance for action.ConclusionsIt is important to evaluate the potential feasibility and acceptability of a clinical risk prediction model, to inform the intervention design and implementation strategy. The results of this study can provide guidance for the development of the clinical risk prediction tool as an intervention for integration in home care settings.