DRRisk: A Web-based tool to Assess the Risk of Diabetic Retinopathy through Machine Learning on Electronic Health Records.

Objective: We developed a web-based tool for diabetic retinopathy (DR) risk assessment called DRRisk (https://drandml.cdrewu.edu/) using machine learning on electronic health record (EHR) data, with a goal of preventing vision loss in persons with diabetes, especially in underserved settings. Methods: DRRisk uses Python's Flask framework. Its user-interface is implemented using HTML, CSS and Javascript. Clinical experts were consulted on the tool's design. Results: DRRisk assesses current DR risk for people with diabetes, categorizing their risk level as low, moderate, or high, depending on the percentage of DR risk assigned by the underlying machine learning model. Discussion: A goal of our tool is to help providers prioritize patients at high risk for DR in order to prevent blindness. Conclusion: Our tool uses DR risk factors from EHR data to calculate a diabetic person's current DR risk. It may be useful for identifying unscreened diabetic patients who have undiagnosed DR.

Detecting diabetic retinopathy through machine learning on electronic health record data from an urban, safety net healthcare system.

OBJECTIVE: Clinical guidelines recommend annual eye examinations to detect diabetic retinopathy (DR) in patients with diabetes. However, timely DR detection remains a problem in medically underserved and under-resourced settings in the United States. Machine learning that identifies patients with latent/undiagnosed DR could help to address this problem. MATERIALS AND METHODS: Using electronic health record data from 40 631 unique diabetic patients seen at Los Angeles County Department of Health Services healthcare facilities between January 1, 2015 and December 31, 2017, we compared ten machine learning environments, including five classifier models, for assessing the presence or absence of DR. We also used data from a distinct set of 9300 diabetic patients seen between January 1, 2018 and December 31, 2018 as an external validation set. RESULTS: Following feature subset selection, the classifier with the best AUC on the external validation set was a deep neural network using majority class undersampling, with an AUC of 0.8, the sensitivity of 72.17%, and specificity of 74.2%. DISCUSSION: A deep neural network produced the best AUCs and sensitivity results on the test set and external validation set. Models are intended to be used to screen guideline noncompliant diabetic patients in an urban safety-net setting. CONCLUSION: Machine learning on diabetic patients' routinely collected clinical data could help clinicians in safety-net settings to identify and target unscreened diabetic patients who potentially have undiagnosed DR.

Predictive Models for Diabetic Retinopathy from Non-Image Teleretinal Screening Data.

Introduction: Timely diabetic retinopathy detection remains a problem in medically underserved settings in the US; diabetic patients in these locales have limited access to eye specialists. Teleretinal screening programs have been introduced to address this problem. Methods: Using data on ethnicity, gender, age, hemoglobin A1C, insulin dependence, time since last eye examination, subjective diabetes control, and years with diabetes from 27,116 diabetic patients participating in a Los Angeles County teleretinal screening program, we compared different machine learning methods for predicting retinopathy. The dataset exhibited a class imbalance. Results: Six classifiers learned on the data were predictive of retinopathy. The best model had an AUC of 0.754, sensitivity of 58% and specificity of 80%. Discussion: Successfully detecting retinopathy from diabetic patients' routinely collected clinical data could help clinicians in medically underserved areas identify unscreened diabetic patients who are at risk of developing retinopathy. This work is a step towards that goal.

2018 Salary Survey of AMIA Members: Factors Associated with Higher Salaries.

Greater transparency in salaries overall and in factors associated with differing salaries can help students and professionals plan their careers, discover biases and obstacles, and help advance professional disciplines broadly. In March 2018, we conducted the first salary survey of American Medical Informatics Association members. Our goal was to summarize salary information and provide a nuanced view pertaining to the diverse biomedical informatics community. To identify factors associated with higher salaries, we reviewed average salaries for different groups (physician status, academic status, and different leadership positions) by gender. We also fitted multiple linear regression models for all participants (N = 201) and for gender, physician- and academic-status subgroup. The mean (standard deviation) salary was $181,774 ($99,566). Men earned more than women on average, and especially among professionals from academic settings. More years working in informatics and full-time employment were two factors that were consistently associated with higher salary.

Understanding the Knowledge Gap Experienced by U.S. Safety Net Patients in Teleretinal Screening.

Safety-net patients' socioeconomic barriers interact with limited digital and health literacies to produce a "knowledge gap" that impacts the delivery of healthcare via telehealth technologies. Six focus groups (2 African- American and 4 Latino) were conducted with patients who received teleretinal screening in a U.S. urban safety-net setting. Focus groups were analyzed using a modified grounded theory methodology. Findings indicate that patients' knowledge gap is primarily produced at three points during the delivery of care: (1) exacerbation of patients' pre-existing personal barriers in the clinical setting; (2) encounters with technology during screening; and (3) lack of follow up after the visit. This knowledge gap produces confusion, potentially limiting patients' perceptions of care and their ability to manage their own care. It may be ameliorated through delivery of patient education focused on both disease pathology and specific role of telehealth technologies in disease management.

Identifying appropriate reference data models for comparative effectiveness research (CER) studies based on data from clinical information systems.

INTRODUCTION: The need for a common format for electronic exchange of clinical data prompted federal endorsement of applicable standards. However, despite obvious similarities, a consensus standard has not yet been selected in the comparative effectiveness research (CER) community. METHODS: Using qualitative metrics for data retrieval and information loss across a variety of CER topic areas, we compare several existing models from a representative sample of organizations associated with clinical research: the Observational Medical Outcomes Partnership (OMOP), Biomedical Research Integrated Domain Group, the Clinical Data Interchange Standards Consortium, and the US Food and Drug Administration. RESULTS: While the models examined captured a majority of the data elements that are useful for CER studies, data elements related to insurance benefit design and plans were most detailed in OMOP's CDM version 4.0. Standardized vocabularies that facilitate semantic interoperability were included in the OMOP and US Food and Drug Administration Mini-Sentinel data models, but are left to the discretion of the end-user in Biomedical Research Integrated Domain Group and Analysis Data Model, limiting reuse opportunities. Among the challenges we encountered was the need to model data specific to a local setting. This was handled by extending the standard data models. DISCUSSION: We found that the Common Data Model from the OMOP met the broadest complement of CER objectives. Minimal information loss occurred in mapping data from institution-specific data warehouses onto the data models from the standards we assessed. However, to support certain scenarios, we found a need to enhance existing data dictionaries with local, institution-specific information.

A comparison of methods for assessing penetrating trauma on retrospective multi-center data.

OBJECTIVE: TraumaSCAN-Web (TSW) is a computerized decision support system for assessing chest and abdominal penetrating trauma which utilizes 3D geometric reasoning and a Bayesian network with subjective probabilities obtained from an expert. The goal of the present study is to determine whether a trauma risk prediction approach using a Bayesian network with a predefined structure and probabilities learned from penetrating trauma data is comparable in diagnostic accuracy to TSW. METHODS: Parameters for two Bayesian networks with expert-defined structures were learned from 637 gunshot and stab wound cases from three hospitals, and diagnostic accuracy was assessed using 10-fold cross-validation. The first network included information on external wound locations, while the second network did not. Diagnostic accuracy of learned networks was compared to that of TSW on 194 previously evaluated cases. RESULTS: For 23 of the 24 conditions modeled by TraumaSCAN-Web, 16 conditions had Areas Under the ROC Curve (AUCs) greater than 0.90 while 21 conditions had AUCs greater than 0.75 for the first network. For the second network, 16 and 20 conditions had AUCs greater than 0.90 and 0.75, respectively. AUC results for learned networks on 194 previously evaluated cases were better than or equal to AUC results for TSW for all diagnoses evaluated except diaphragm and heart injuries. CONCLUSIONS: For 23 of the 24 penetrating trauma conditions studied, a trauma diagnosis approach using Bayesian networks with predefined structure and probabilities learned from penetrating trauma data was better than or equal in diagnostic accuracy to TSW. In many cases, information on wound location in the first network did not significantly add to predictive accuracy. The study suggests that a decision support approach that uses parameter-learned Bayesian networks may be sufficient for assessing some penetrating trauma conditions.

Evaluating the discriminatory power of a computer-based system for assessing penetrating trauma on retrospective multi-center data.

OBJECTIVE: To evaluate the discriminatory power of TraumaSCAN-Web, a system for assessing penetrating trauma, using retrospective multi-center case data for gunshot and stab wounds to the thorax and abdomen. METHODS: 80 gunshot and 114 stab cases were evaluated using TraumaSCAN-Web. Areas under the Receiver Operator Characteristic Curves (AUC) were calculated for each condition modeled in TraumaSCAN-Web. RESULTS: Of the 23 conditions modeled by TraumaSCAN-Web, 19 were present in either the gunshot or stab case data. The gunshot AUCs ranged from 0.519 (pericardial tamponade) to 0.975 (right renal injury). The stab AUCs ranged from 0.701 (intestinal injury) to 1.000 (tracheal injury).

Combining geometric and probabilistic reasoning for computer-based penetrating-trauma assessment.

OBJECTIVE: To ascertain whether three-dimensional geometric and probabilistic reasoning methods can be successfully combined for computer-based assessment of conditions arising from ballistic penetrating trauma to the chest and abdomen. DESIGN: The authors created a computer system (TraumaSCAN) that integrates three-dimensional geometric reasoning about anatomic likelihood of injury with probabilistic reasoning about injury consequences using Bayesian networks. Preliminary evaluation of TraumaSCAN was performed via a retrospective study testing performance of the system on data from 26 cases of actual gunshot wounds. MEASUREMENTS: Areas under the receiver operating characteristics (ROC) curve were calculated for each condition modeled in TraumaSCAN that was present in the 26 cases. The comprehensiveness and relevance of the TraumaSCAN diagnosis for the 26 cases were used to assess the overall performance of the system. To test the ability of TraumaSCAN to handle limited findings, these measurements were calculated both with and without input of observed findings into the Bayesian network. RESULTS: For the 11 conditions assessed, the worst area under the ROC curve with no observed findings input into the Bayesian network was 0.542 (95% CI, 0.146-0.937), the median was 0.883 (95% CI, 0.713-1.000), and the best was 1.00 (95% CI, 1.000-1.000). The worst area under the ROC curve with all observed findings input into the Bayesian network was 0.835 (95% CI, 0.602-1.000), the median was 0.941 (95% CI, 0.827-1.000), and the best was 0.992 (95% CI, 0.965-1.000). A comparison of the areas under the curve obtained with and without input of observed findings into the Bayesian network showed that there were significant differences for 2 of the 11 conditions assessed. CONCLUSION: A computer-based method that combines geometric and probabilistic reasoning shows promise as a tool for assessing ballistic penetrating trauma to the chest and abdomen.