Maintenance Notice

Due to necessary scheduled maintenance, the JMIR Publications website will be unavailable from Monday, March 11, 2019 at 4:00 PM to 4:30 PM EST. We apologize in advance for any inconvenience this may cause you.

Who will be affected?

Advertisement

Journal Description

JMIR Medical Informatics (JMI, ISSN 2291-9694; Impact Factor: 3.188) (Editor-in-chief: Christian Lovis MD MPH FACMI) is a Pubmed/SCIE-indexed, top-rated, tier A journal which focuses on clinical informatics, big data in health and health care, decision support for health professionals, electronic health records, ehealth infrastructures and implementation. It has a focus on applied, translational research, with a broad readership including clinicians, CIOs, engineers, industry and health informatics professionals.

Published by JMIR Publications, publisher of the Journal of Medical Internet Research (JMIR), JMIR Med Inform has a slightly different scope (putting more emphasis on applications for clinicians and health professionals rather than consumers/citizens), publishes even faster, and also allows papers which are more technical or more formative than what would be published in JMIR.

JMIR Med Inform is indexed in PubMed Central/PubMed and has also been accepted for SCIE. JMIR Med Inform received an inaugural Journal Impact Factor for 2018 (released June 2019) of 3.188.

JMIR Medical Informatics adheres to the same quality standards as JMIR and all articles published here are also cross-listed in the Table of Contents of JMIR (http://www.jmir.org/issue/current).

 

Recent Articles:

  • Impact on Readmission Reduction Among Heart Failure Patients Using Digital Health Monitoring: Feasibility and Adoptability Study

    Abstract:

    Background: Heart failure (HF) is a condition that affects approximately 6.2 million people in the United States and has a 5-year mortality rate of approximately 42%. With the prevalence expected to exceed 8 million cases by 2030, projections estimate that total annual HF costs will increase to nearly US $70 billion. Recently, the advent of remote monitoring technology has significantly broadened the scope of the physician’s reach in chronic disease management. Objective: The goal of our program, named the Heart Health Program, was to examine the feasibility of using digital health monitoring in real-world home settings, ascertain patient adoption, and evaluate impact on 30-day readmission rate. Methods: A digital medicine software platform developed at Mount Sinai Health System, called RxUniverse, was used to prescribe a digital care pathway including the HealthPROMISE digital therapeutic and iHealth mobile apps to patients’ personal smartphones. Vital sign data, including blood pressure (BP) and weight, were collected through an ambulatory remote monitoring system that comprised a mobile app and complementary consumer-grade Bluetooth-connected smart devices (BP cuff and digital scale) that send data to the provider care teams. Care teams were alerted via a Web-based dashboard of abnormal patient BP and weight change readings, and further action was taken at the clinicians’ discretion. We used statistical analyses to determine risk factors associated with 30-day all-cause readmission. Results: Overall, the Heart Health Program included 58 patients admitted to the Mount Sinai Hospital for HF. The 30-day hospital readmission rate was 10% (6/58), compared with the national readmission rates of approximately 25% and the Mount Sinai Hospital’s average of approximately 23%. Single marital status (P=.06) and history of percutaneous coronary intervention (P=.08) were associated with readmission. Readmitted patients were also less likely to have been previously prescribed angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers (P=.02). Notably, readmitted patients utilized the BP and weight monitors less than nonreadmitted patients, and patients aged younger than 70 years used the monitors more frequently on average than those aged over 70 years, though these trends did not reach statistical significance. The percentage of the 58 patients using the monitors at least once dropped from 83% (42/58) in the first week after discharge to 46% (23/58) in the fourth week. Conclusions: Given the increasing burden of HF, there is a need for an effective and sustainable remote monitoring system for HF patients following hospital discharge. We identified clinical and social factors as well as remote monitoring usage trends that identify targetable patient populations that could benefit most from integration of daily remote monitoring. In addition, we demonstrated that interventions driven by real-time vital sign data may greatly aid in reducing hospital readmissions and costs while improving patient outcomes.

  • Source: istockphoto; Copyright: Deagreez; URL: https://www.istockphoto.com/photo/top-view-cropped-photo-of-doctor-in-formal-wear-writing-down-a-prescription-for-a-gm926638678-254249544; License: Licensed by the authors.

    Combining Contextualized Embeddings and Prior Knowledge for Clinical Named Entity Recognition: Evaluation Study

    Abstract:

    Background: Named entity recognition (NER) is a key step in clinical natural language processing (NLP). Traditionally, rule-based systems leverage prior knowledge to define rules to identify named entities. Recently, deep learning–based NER systems have become more and more popular. Contextualized word embedding, as a new type of representation of the word, has been proposed to dynamically capture word sense using context information and has proven successful in many deep learning–based systems in either general domain or medical domain. However, there are very few studies that investigate the effects of combining multiple contextualized embeddings and prior knowledge on the clinical NER task. Objective: This study aims to improve the performance of NER in clinical text by combining multiple contextual embeddings and prior knowledge. Methods: In this study, we investigate the effects of combining multiple contextualized word embeddings with classic word embedding in deep neural networks to predict named entities in clinical text. We also investigate whether using a semantic lexicon could further improve the performance of the clinical NER system. Results: By combining contextualized embeddings such as ELMo and Flair, our system achieves the F-1 score of 87.30% when only training based on a portion of the 2010 Informatics for Integrating Biology and the Bedside NER task dataset. After incorporating the medical lexicon into the word embedding, the F-1 score was further increased to 87.44%. Another finding was that our system still could achieve an F-1 score of 85.36% when the size of the training data was reduced to 40%. Conclusions: Combined contextualized embedding could be beneficial for the clinical NER task. Moreover, the semantic lexicon could be used to further improve the performance of the clinical NER system.

  • Student using an EHR system. Source: Image created by the Authors; Copyright: The Authors; URL: https://medinform.jmir.org/4/12648; License: Creative Commons Attribution (CC-BY).

    Navigating Through Electronic Health Records: Survey Study on Medical Students’ Perspectives in General and With Regard to a Specific Training

    Abstract:

    Background: An electronic health record (EHR) is the state-of-the-art method for ensuring all data concerning a given patient are up to date for use by multidisciplinary hospital teams. Therefore, medical students need to be trained to use health information technologies within this environment from the early stages of their education. Objective: As little is known about the effects of specific training within the medical curriculum, this study aimed to develop a course module and evaluate it to offer best practice teaching for today’s students. Moreover, we looked at the acceptance of new technologies such as EHRs. Methods: Fifth-year medical students (N=104) at the University of Tübingen took part in a standardized two-day training procedure about the advantages and risks of EHR use. After the training, students performed their own EHR entries on hypothetical patient cases in a safe practice environment. In addition, questionnaires—standardized and with open-ended questions—were administered to assess students’ experiences with a new teaching module, a newly developed EHR simulator, the acceptance of the health technology, and their attitudes toward it before and after training. Results: After the teaching, students rated the benefit of EHR training for medical knowledge significantly higher than before the session (mean 3.74, SD 1.05). However, they also had doubts about the long-term benefit of EHRs for multidisciplinary coworking after training (mean 1.96, SD 0.65). The special training with simulation software was rated as helpful for preparing students (88/102, 86.2%), but they still did not feel safe in all aspects of EHR. Conclusions: A specific simulated training on using EHRs helped students improve their knowledge and become more aware of the risks and challenges of such a system. Overall, students welcomed the new training module and supported the integration of EHR teaching into the medical curriculum. Further studies are needed to optimize training modules and make use of long-term feedback opportunities a simulated system offers.

  • Source: Foter; Copyright: Foter; URL: https://foter.com/photo3/notebook-person-man-laptops/; License: Public Domain (CC0).

    Developing a Reproducible Microbiome Data Analysis Pipeline Using the Amazon Web Services Cloud for a Cancer Research Group: Proof-of-Concept Study

    Abstract:

    Background: Cloud computing for microbiome data sets can significantly increase working efficiencies and expedite the translation of research findings into clinical practice. The Amazon Web Services (AWS) cloud provides an invaluable option for microbiome data storage, computation, and analysis. Objective: The goals of this study were to develop a microbiome data analysis pipeline by using AWS cloud and to conduct a proof-of-concept test for microbiome data storage, processing, and analysis. Methods: A multidisciplinary team was formed to develop and test a reproducible microbiome data analysis pipeline with multiple AWS cloud services that could be used for storage, computation, and data analysis. The microbiome data analysis pipeline developed in AWS was tested by using two data sets: 19 vaginal microbiome samples and 50 gut microbiome samples. Results: Using AWS features, we developed a microbiome data analysis pipeline that included Amazon Simple Storage Service for microbiome sequence storage, Linux Elastic Compute Cloud (EC2) instances (ie, servers) for data computation and analysis, and security keys to create and manage the use of encryption for the pipeline. Bioinformatics and statistical tools (ie, Quantitative Insights Into Microbial Ecology 2 and RStudio) were installed within the Linux EC2 instances to run microbiome statistical analysis. The microbiome data analysis pipeline was performed through command-line interfaces within the Linux operating system or in the Mac operating system. Using this new pipeline, we were able to successfully process and analyze 50 gut microbiome samples within 4 hours at a very low cost (a c4.4xlarge EC2 instance costs $0.80 per hour). Gut microbiome findings regarding diversity, taxonomy, and abundance analyses were easily shared within our research team. Conclusions: Building a microbiome data analysis pipeline with AWS cloud is feasible. This pipeline is highly reliable, computationally powerful, and cost effective. Our AWS-based microbiome analysis pipeline provides an efficient tool to conduct microbiome data analysis.

  • Detecting hypoglycemic events from EHR notes. Source: Freepik; Copyright: pressfoto; URL: https://www.freepik.com/free-photo/interns-consultng-with-doctor_5634018.htm#query=doctor%20tablet&position=40; License: Licensed by JMIR.

    Automatic Detection of Hypoglycemic Events From the Electronic Health Record Notes of Diabetes Patients: Empirical Study

    Abstract:

    Background: Hypoglycemic events are common and potentially dangerous conditions among patients being treated for diabetes. Automatic detection of such events could improve patient care and is valuable in population studies. Electronic health records (EHRs) are valuable resources for the detection of such events. Objective: In this study, we aim to develop a deep-learning–based natural language processing (NLP) system to automatically detect hypoglycemic events from EHR notes. Our model is called the High-Performing System for Automatically Detecting Hypoglycemic Events (HYPE). Methods: Domain experts reviewed 500 EHR notes of diabetes patients to determine whether each sentence contained a hypoglycemic event or not. We used this annotated corpus to train and evaluate HYPE, the high-performance NLP system for hypoglycemia detection. We built and evaluated both a classical machine learning model (ie, support vector machines [SVMs]) and state-of-the-art neural network models. Results: We found that neural network models outperformed the SVM model. The convolutional neural network (CNN) model yielded the highest performance in a 10-fold cross-validation setting: mean precision=0.96 (SD 0.03), mean recall=0.86 (SD 0.03), and mean F1=0.91 (SD 0.03). Conclusions: Despite the challenges posed by small and highly imbalanced data, our CNN-based HYPE system still achieved a high performance for hypoglycemia detection. HYPE can be used for EHR-based hypoglycemia surveillance and population studies in diabetes patients.

  • Electronic health information exchange by ambulatory providers. Source: Freepik; Copyright: pressfoto; URL: https://www.freepik.com/free-photo/doctor-filling-application_5766725.htm; License: Licensed by JMIR.

    Key Factors Affecting Ambulatory Care Providers’ Electronic Exchange of Health Information With Affiliated and Unaffiliated Partners: Web-Based Survey Study

    Abstract:

    Background: Despite the potential benefits of electronic health information exchange (HIE) to improve the quality and efficiency of care, HIE use by ambulatory providers remains low. Ambulatory providers can greatly improve the quality of care by electronically exchanging health information with affiliated providers within their health care network as well as with unaffiliated, external providers. Objective: This study aimed to examine the extent of electronic HIE use by ambulatory clinics with affiliated providers within their health system and with external providers, as well as the key technological, organizational, and environmental factors affecting the extent of HIE use within and outside the health system. Methods: A Web-based survey of 320 ambulatory care providers was conducted in the state of Illinois. The study examined the extent of HIE usage by ambulatory providers with hospitals, clinics, and other facilities within and outside their health care system–encompassing seven kinds of health care data. Ten factors pertaining to technology (IT [information technology] Compatibility, External IT Support, Security & Privacy Safeguards), organization (Workflow Adaptability, Senior Leadership Support, Clinicians Health-IT Knowledge, Staff Health-IT Knowledge), and environment (Government Efforts & Incentives, Partner Readiness, Competitors and Peers) were assessed. A series of multivariate regressions were used to examine predictor effects. Results: The 6 regressions produced adjusted R-squared values ranging from 0.44 to 0.63. We found that ambulatory clinics exchanged more health information electronically with affiliated entities within their health system as compared with those outside their health system. Partner readiness emerged as the most significant predictor of HIE usage with all entities. Governmental initiatives for HIE, clinicians’ prior familiarity and knowledge of health IT systems, implementation of appropriate security, and privacy safeguards were also significant predictors. External information technology support and workflow adaptability emerged as key predictors for HIE use outside a clinic’s health system. Differences based on clinic size, ownership, and specialty were also observed. Conclusions: This study provides exploratory insights into HIE use by ambulatory providers within and outside their health care system and differential predictors that impact HIE use. HIE use can be further improved by encouraging large-scale interoperability efforts, improving external IT support, and redesigning adaptable workflows.

  • Source: Shutterstock; Copyright: Gorodenkoff; URL: https://www.shutterstock.com/image-photo/surgeon-female-doctor-walk-through-hospital-1189798327?src=1xDp0icwbYi7lOPwWQGK4w-1-10; License: Licensed by the authors.

    Usability Factors Associated With Physicians’ Distress and Information System–Related Stress: Cross-Sectional Survey

    Abstract:

    Background: Constantly changing and difficult-to-use information systems have arisen as a significant source of stress in physicians’ work. Physicians have reported several usability problems, system failures, and a lack of integration between the systems and have experienced that systems poorly support the documentation and retrieval of patient data. This stress has kept rising in the 21st century, and it seems that it may also affect physicians’ well-being. Objective: This study aimed to examine the associations of (1) usability variables (perceived benefits, technical problems, support for feedback, and user-friendliness), (2) the number of systems in daily use, (3) experience of using information systems, and (4) participation in information systems development work with physicians’ distress and levels of stress related to information systems (SRIS) levels. Methods: A cross-sectional survey was conducted among 4018 Finnish physicians (64.82%, 2572 out of 3968 women) aged between 24 and 64 years (mean 46.8 years) in 2017. The analyses of covariance were used to examine the association of independent variables with SRIS and distress (using the General Health Questionnaire) adjusted for age, gender, employment sector, specialization status, and the electronic health record system in use. Results: High levels of technical problems and a high number of systems in daily use were associated with high levels of SRIS, whereas high levels of user-friendliness, perceived benefits, and support for feedback were associated with low levels of SRIS. Moreover, high levels of technical problems were associated with high levels of psychological distress, whereas high levels of user-friendliness were associated with low distress levels. Those who considered themselves experienced users of information systems had low levels of both SRIS and distress. Conclusions: It seems that by investing in user-friendly systems with better technical quality and good support for feedback that professionals perceive as being beneficial would improve the work-related well-being and overall well-being of physicians. Moreover, improving physicians’ skills related to information systems by giving them training could help to lessen the stress that results from poorly functioning information systems and improve physicians’ well-being.

  • An automated medical history–taking device (DIAANA) helping physicians in their differential diagnose by providing an exhaustive, high-quality, standardized anamnesis and differential diagnosis. Source: Pxhere; Copyright: rawpixels.com; URL: https://pxhere.com/en/photo/1430845; License: Public Domain (CC0).

    Differential Diagnosis Assessment in Ambulatory Care With an Automated Medical History–Taking Device: Pilot Randomized Controlled Trial

    Abstract:

    Background: Automated medical history–taking devices (AMHTDs) are emerging tools with the potential to increase the quality of medical consultations by providing physicians with an exhaustive, high-quality, standardized anamnesis and differential diagnosis. Objective: This study aimed to assess the effectiveness of an AMHTD to obtain an accurate differential diagnosis in an outpatient service. Methods: We conducted a pilot randomized controlled trial involving 59 patients presenting to an emergency outpatient unit and suffering from various conditions affecting the limbs, the back, and the chest wall. Resident physicians were randomized into 2 groups, one assisted by the AMHTD and one without access to the device. For each patient, physicians were asked to establish an exhaustive differential diagnosis based on the anamnesis and clinical examination. In the intervention group, residents read the AMHTD report before performing the anamnesis. In both the groups, a senior physician had to establish a differential diagnosis, considered as the gold standard, independent of the resident’s opinion and AMHTD report. Results: A total of 29 patients were included in the intervention group and 30 in the control group. Differential diagnosis accuracy was higher in the intervention group (mean 75%, SD 26%) than in the control group (mean 59%, SD 31%; P=.01). Subgroup analysis showed a between-group difference of 3% (83% [17/21]-80% [14/17]) for low complexity cases (1-2 differential diagnoses possible) in favor of the AMHTD (P=.76), 31% (87% [13/15]-56% [18/33]) for intermediate complexity (3 differential diagnoses; P=.02), and 24% (63% [34/54]-39% [14/35]) for high complexity (4-5 differential diagnoses; P=.08). Physicians in the intervention group (mean 4.3, SD 2) had more years of clinical practice compared with the control group (mean 5.5, SD 2; P=.03). Differential diagnosis accuracy was negatively correlated to case complexity (r=0.41; P=.001) and the residents’ years of practice (r=0.04; P=.72). The AMHTD was able to determine 73% (SD 30%) of correct differential diagnoses. Patient satisfaction was good (4.3/5), and 26 of 29 patients (90%) considered that they were able to accurately describe their symptomatology. In 8 of 29 cases (28%), residents considered that the AMHTD helped to establish the differential diagnosis. Conclusions: The AMHTD allowed physicians to make more accurate differential diagnoses, particularly in complex cases. This could be explained not only by the ability of the AMHTD to make the right diagnoses, but also by the exhaustive anamnesis provided.

  • Source: Unsplash; Copyright: Marcelo Leal; URL: https://unsplash.com/photos/6pcGTJDuf6M; License: Licensed by JMIR.

    Digital Health and the State of Interoperable Electronic Health Records

    Abstract:

    Digital health systems and innovative care delivery within these systems have great potential to improve national health care and positively impact the health outcomes of patients. However, currently, very few countries have systems that can implement digital interventions at scale. This is partly because of the lack of interoperable electronic health records (EHRs). It is difficult to make decisions for an individual or population when the data on that person or population are dispersed over multiple incompatible systems. This viewpoint paper has highlighted some key obstacles of current EHRs and some promising successes, with the goal of promoting EHR evolution and advocating for frameworks that develop digital health systems that serve populations—a critical goal as we move further into this data-rich century with an ever-increasing number of patients who live longer and depend on health care services where resources may already be strained. This paper aimed to analyze the evolution, obstacles, and current landscape of EHRs and identify fundamental areas of hindrance for interoperability. It also aimed to highlight countries where advances have been made and extract best practices from these examples. The obstacles to EHR interoperability are not easily solved, but improving the current situation in countries where a national policy is not in place will require a focused inquiry into solutions from various sources in the public and private sector. Effort must be made on a national scale to seek solutions for optimally interoperable EHRs beyond status quo solutions. A list of considerations for best practices is suggested.

  • Source: iStock by Getty Images; Copyright: mediaphotos; URL: https://www.istockphoto.com/ca/photo/doctor-using-tablet-gm468864398-61240656; License: Licensed by the authors.

    Extracting Clinical Features From Dictated Ambulatory Consult Notes Using a Commercially Available Natural Language Processing Tool: Pilot, Retrospective,...

    Abstract:

    Background: The increasing adoption of electronic health records (EHRs) in clinical practice holds the promise of improving care and advancing research by serving as a rich source of data, but most EHRs allow clinicians to enter data in a text format without much structure. Natural language processing (NLP) may reduce reliance on manual abstraction of these text data by extracting clinical features directly from unstructured clinical digital text data and converting them into structured data. Objective: This study aimed to assess the performance of a commercially available NLP tool for extracting clinical features from free-text consult notes. Methods: We conducted a pilot retrospective cross-sectional study of the accuracy of NLP from dictated consult notes from our tuberculosis clinic with manual chart abstraction as the reference standard. Consult notes for 130 patients were extracted and processed using NLP. We extracted 15 clinical features from these consult notes and grouped them a priori into categories of simple, moderate, and complex for analysis. Results: For the primary outcome of overall accuracy, NLP performed best for features classified as simple, achieving an overall accuracy of 96% (95% CI 94.3-97.6). Performance was slightly lower for features of moderate clinical and linguistic complexity at 93% (95% CI 91.1-94.4), and lowest for complex features at 91% (95% CI 87.3-93.1). Conclusions: The findings of this study support the use of NLP for extracting clinical features from dictated consult notes in the setting of a tuberculosis clinic. Further research is needed to fully establish the validity of NLP for this and other purposes.

  • Source: Bigstock; Copyright: Proxima Studio; URL: https://www.bigstockphoto.com/image-315228529/stock-photo-doctor-and-senior-woman-patient-medical-consultation%2C-telehealth%2C-telemedicine%2C-remote-health-care-c; License: Licensed by the authors.

    Opportunities and Challenges of Telehealth in Remote Communities: Case Study of the Yukon Telehealth System

    Abstract:

    Background: Telehealth has been shown to improve access to health care and to reduce costs to the patient and health care system, especially for patients living in rural settings. However, unique challenges arise when implementing telehealth in remote communities. Objective: The study aimed to explore the current use, challenges, and opportunities of the Yukon Telehealth System. The lessons learned from this study were used to determine important factors to consider when attempting to advance and expand telehealth programs in remote communities. Methods: A mixed methods approach was used to evaluate the Yukon Telehealth System and to determine possible future advances. Quantitative data were obtained through usage logs. Web-based questionnaires were administered to nurses in each of the 14 Yukon community health centers outside of Whitehorse and patients who had used telehealth. Qualitative data included focus groups and semistructured interviews with 36 telehealth stakeholders. Results: Since 2008, there has been a consistent number of telehealth sessions of about 1000 per year, with clinical care as the main use (69.06% [759/1099] of all sessions in 2015). From the questionnaire (11 community nurses and 10 patients) and the interview data, there was a consensus among the clinicians and patients that the system provided timely access and cost savings from reduced travel. However, they believed that it was underutilized, and the equipment was outdated. The following 4 factors were identified, which should be considered when trying to advance and expand a telehealth program: (1) patient and clinician buy-in: past telehealth experiences (eg, negative clinician experiences with outdated technology) should be considered when advancing the system. Expansion of services in orthopedics, dermatology, and psychiatry were found to be particularly feasible and beneficial in Yukon; (2) workflow: the use and scheduling of telehealth should be streamlined and automated as much as possible to reduce dependencies on the single Yukon telehealth coordinator; (3) access to telehealth technology: clinicians and patients should have easy access to up-to-date telehealth technology. The use of consumer products, such as mobile technology, should be leveraged as appropriate; and (4) infrastructure: the required human resources and technology need to be established when expanding and advancing telehealth. Conclusions: While clinicians and patients had generally positive perceptions of the Yukon Telehealth System, there was consensus that it was underutilized. Many opportunities exist to expand the types of telehealth services and the number of telehealth sessions, including the expansion of services in several new specialty areas, updating telehealth equipment to streamline workflows and increase convenience and uptake, and integrating novel technologies. The identified barriers and recommendations from this evaluation can be applied to the development and expansion of telehealth in other remote communities to realize telehealth’s potential for providing efficient, safe, convenient, and cost-effective care delivery.

  • Source: Pixabay; Copyright: rawpixel; URL: https://pixabay.com/photos/computer-business-office-technology-3343887/; License: Public Domain (CC0).

    Cohort Selection for Clinical Trials From Longitudinal Patient Records: Text Mining Approach

    Abstract:

    Background: Clinical trials are an important step in introducing new interventions into clinical practice by generating data on their safety and efficacy. Clinical trials need to ensure that participants are similar so that the findings can be attributed to the interventions studied and not to some other factors. Therefore, each clinical trial defines eligibility criteria, which describe characteristics that must be shared by the participants. Unfortunately, the complexities of eligibility criteria may not allow them to be translated directly into readily executable database queries. Instead, they may require careful analysis of the narrative sections of medical records. Manual screening of medical records is time consuming, thus negatively affecting the timeliness of the recruitment process. Objective: Track 1 of the 2018 National Natural Language Processing Clinical Challenge focused on the task of cohort selection for clinical trials, aiming to answer the following question: Can natural language processing be applied to narrative medical records to identify patients who meet eligibility criteria for clinical trials? The task required the participating systems to analyze longitudinal patient records to determine if the corresponding patients met the given eligibility criteria. We aimed to describe a system developed to address this task. Methods: Our system consisted of 13 classifiers, 1 for each eligibility criterion. All classifiers used a bag-of-words (BoW) document representation model. To prevent the loss of relevant contextual information associated with such representation, a pattern-matching approach was used to extract context-sensitive features. They were embedded back into the text as lexically distinguishable tokens, which were consequently featured in the BoW representation. Supervised machine learning was chosen wherever a sufficient number of both positive and negative instances was available to learn from. A rule-based approach focusing on a small set of relevant features was chosen for the remaining criteria. Results: The system was evaluated using microaveraged F measure. Overall, 4 machine algorithms, including support vector machine, logistic regression, naïve Bayesian classifier, and gradient tree boosting (GTB), were evaluated on the training data using 10–fold cross-validation. Overall, GTB demonstrated the most consistent performance. Its performance peaked when oversampling was used to balance the training data. The final evaluation was performed on previously unseen test data. On average, the F measure of 89.04% was comparable to 3 of the top ranked performances in the shared task (91.11%, 90.28%, and 90.21%). With an F measure of 88.14%, we significantly outperformed these systems (81.03%, 78.50%, and 70.81%) in identifying patients with advanced coronary artery disease. Conclusions: The holdout evaluation provides evidence that our system was able to identify eligible patients for the given clinical trial with high accuracy. Our approach demonstrates how rule-based knowledge infusion can improve the performance of machine learning algorithms even when trained on a relatively small dataset.

Citing this Article

Right click to copy or hit: ctrl+c (cmd+c on mac)

Latest Submissions Open for Peer-Review:

View All Open Peer Review Articles
  • Intelligent alarms using principal component analysis on hemodynamic monitoring

    Date Submitted: Nov 12, 2019

    Open Peer Review Period: Nov 10, 2019 - Jan 5, 2020

    Background: Patient monitors in modern hospitals give heartbeat waveform data that is reduced to aggregated variables and simple thresholds for alarms. Often, the monitors give a steady stream of non...

    Background: Patient monitors in modern hospitals give heartbeat waveform data that is reduced to aggregated variables and simple thresholds for alarms. Often, the monitors give a steady stream of non-specific alarms, leading to alarm fatigue in clinicians. An alarm can be seen as a classification problem, and by applying Principal Component Analysis (PCA) to the heart rate waveform of readily available monitoring devices, the accuracy of the classification of abnormality could be highly increased. Objective: To investigate whether physiological changes could be detected by looking at the heart rate waveform. Methods: A dataset of a healthy volunteer monitored with electrocardiography (ECG) and invasive blood pressure (BP) experiencing several tilts on a tilting table was investigated. A novel way of splitting continuous data based on the heartbeat was introduced. PCA was applied to classify the heartbeats. Results: A classification using only the aggregated variables heart rate (HR) and BP was able to correctly identify 20.7% of the heartbeats in the vertical tilt as abnormal. A classification using the full waveforms and combining the ECG and BP signals was able to correctly identify 83.5% of the heartbeats in the vertical tilt as abnormal. A humanistic machine learning (ML) method is then proposed based on the PCA classification. Conclusions: A ML method for classification of physiological variability was described. The main novelty lies in the splitting of an ECG and BP signal by the heart rate and performing a PCA on the data-table.

  • A review and comparison of third-party software applications for electronic health records

    Date Submitted: Nov 8, 2019

    Open Peer Review Period: Nov 8, 2019 - Jan 3, 2020

    Background: Third-party electronic health record (EHR) apps are available to allow healthcare organizations to extend the capabilities and features of their EHR. Given the widespread utilization of E...

    Background: Third-party electronic health record (EHR) apps are available to allow healthcare organizations to extend the capabilities and features of their EHR. Given the widespread utilization of EHRs, and the emergence of third-party apps in EHR marketplaces, it has become necessary to conduct a systematic review and analysis of apps in EHR app marketplaces. Objective: The goal of this review is to organize, categorize, and characterize availability of third-party apps in EHR marketplaces. Methods: Two informaticists (JR & BW) used grounded theory principles to review and categorize EHR apps listed in top EHR vendors’ public-facing marketplaces. Results: We categorized a total of 471 EHR apps into a taxonomy consisting of 3 primary categories, 15 secondary categories, and 55 tertiary categories. The three primary categories were administrative (203 apps; 43.1%), provider support (159 apps; 33.8%), and patient care (109 apps; 23.1%). Within administrative apps, we split the apps into four secondary categories: front office (77 apps), financial (53 apps), office administration (49 apps), and office device integration (17 apps). Within the provider support primary classification, we split the apps into eight secondary categories: documentation (34 apps), records management (27 apps), care coordination (23 apps), population health (18 apps), EHR efficiency (16 apps), ordering & prescribing (15 apps), medical device integration (13 apps) and specialty EHR (12 apps). Within the patient care primary classification, we split the apps into three secondary categories: patient engagement (50 apps), clinical decision support (40 apps), and remote care (18 apps). Total app counts varied substantially across EHR vendors. Overall distribution of apps across primary categories were relatively similar with a few exceptions. Conclusions: We characterized and organized a diverse and rich set of third-party EHR apps. This work provides an important reference for developers, researchers, and EHR customers to more easily search, review, and compare apps in EHR app stores.

  • Which characteristics of primary care doctors determine their use of teleconsultations in the catalan public Health system? A retrospective descriptive cross-sectional study

    Date Submitted: Oct 3, 2019

    Open Peer Review Period: Oct 3, 2019 - Nov 28, 2019

    Background: eConsulta is a teleconsultation service involving doctors and patients which is part of Catalonia's public health system's IT system. It has been in operation since the end of 2015 as an a...

    Background: eConsulta is a teleconsultation service involving doctors and patients which is part of Catalonia's public health system's IT system. It has been in operation since the end of 2015 as an adjunct to face-to-face consultations. A key factor in understanding the barriers and facilitators to the acceptance of the tool is understanding the socio-demographic characteristics of General Practitioners who determine its use. Objective: to analyse the socio-demographic factors which affect the likelihood of doctors using eConsulta. Methods: using a retrospective cross-sectional analysis of administrative data, a multivariate logistic regression analysis of the use of eConsulta in relation to a set of socio-demographic variables is performed. Results: the model shows that the characteristics of the doctors who use eConsulta are that they are between 45 and 54 years of age, score higher than the 80th percentile on the quality of care index, have a high degree of accessibility, are involved in teaching and work on a health team in an urban setting which also has a high socio-economic level. Conclusions: the results suggest that certain socio-demographic characteristics associated with General Practitioners determine whether they use eConsulta. These must be taken into account if its deployment is to be encouraged in the context of a public health system.

  • Is YouTube Portuguese videos useful as a source of information on diabetes foot care?

    Date Submitted: Sep 30, 2019

    Open Peer Review Period: Oct 2, 2019 - Dec 2, 2019

    Background: Studies have been assessed the importance of YouTube as a source of information for some health conditions, diseases or procedures. However, this platform provides an ever-growing, unregul...

    Background: Studies have been assessed the importance of YouTube as a source of information for some health conditions, diseases or procedures. However, this platform provides an ever-growing, unregulated source, and some of their information may cause health risks to patients. Objective: The aim of this study was to evaluate the use of Brazilian YouTube videos as a source of useful in-formation about diabetes foot care. Methods: The website www.youtube.com was searched on November 30, 2016, for the term “diabetes foot care” in Brazilian Portuguese to assess their usefulness as an information source. The videos were categorized as very useful, moderately useful, somewhat useful, and not useful. The video sources were categorized into 4 groups: organizational, professional, personal and advertisement. Ethics ap-proval was not required and descriptive statistics were calculated for all variables. Results: Our search resulted in 8.080 videos, of which 200 were reviewed, and 159 videos uploaded on YouTube between 2008 and 2016 were analysed. Videos were categorized as very useful (6.29%), moderately useful (16.35%), somewhat useful (24.35%), and not useful (52.83%). The video source revealed the following classification: organizational, n = 76; professional, n = 11; personal, n= 46; and advertisement, n = 26. Conclusions: YouTube's Brazilian videos on diabetic foot care are popular, with varied sources and content. However, most of their content is not useful. Therefore, YouTube videos in Portuguese cannot be considered a good source of information about diabetes foot care.

Advertisement