Machine Learning: The Next Paradigm Shift in Medical Education.

Machine learning (ML) algorithms are powerful prediction tools with immense potential in the clinical setting. There are a number of existing clinical tools that use ML, and many more are in development. Physicians are important stakeholders in the health care system, but most are not equipped to make informed decisions regarding deployment and application of ML technologies in patient care. It is of paramount importance that ML concepts are integrated into medical curricula to position physicians to become informed consumers of the emerging tools employing ML. This paradigm shift is similar to the evidence-based medicine (EBM) movement of the 1990s. At that time, EBM was a novel concept; now, EBM is considered an essential component of medical curricula and critical to the provision of high-quality patient care. ML has the potential to have a similar, if not greater, impact on the practice of medicine. As this technology continues its inexorable march forward, educators must continue to evaluate medical curricula to ensure that physicians are trained to be informed stakeholders in the health care of tomorrow.

Drug repurposing: Iron in the fire for older drugs.

Repositioning or "repurposing" of existing therapies for indications of alternative disease is an attractive approach that can generate lower costs and require a shorter approval time than developing a de novo drug. The development of experimental drugs is time-consuming, expensive, and limited to a fairly small number of targets. The incorporation of separate and complementary data should be used, as each type of data set exposes a specific feature of organism knowledge Drug repurposing opportunities are often focused on sporadic findings or on time-consuming pre-clinical drug tests which are often not guided by hypothesis. In comparison, repurposing in-silico drugs is a new, hypothesis-driven method that takes advantage of big-data use. Nonetheless, the widespread use of omics technology, enhanced data storage, data sense, machine learning algorithms, and computational modeling all give unparalleled knowledge of the methods of action of biological processes and drugs, providing wide availability, for both disease-related data and drug-related data. This review has taken an in-depth look at the current state, possibilities, and limitations of further progress in the field of drug repositioning.

A Racially Unbiased, Machine Learning Approach to Prediction of Mortality: Algorithm Development Study.

BACKGROUND: Racial disparities in health care are well documented in the United States. As machine learning methods become more common in health care settings, it is important to ensure that these methods do not contribute to racial disparities through biased predictions or differential accuracy across racial groups. OBJECTIVE: The goal of the research was to assess a machine learning algorithm intentionally developed to minimize bias in in-hospital mortality predictions between white and nonwhite patient groups. METHODS: Bias was minimized through preprocessing of algorithm training data. We performed a retrospective analysis of electronic health record data from patients admitted to the intensive care unit (ICU) at a large academic health center between 2001 and 2012, drawing data from the Medical Information Mart for Intensive Care-III database. Patients were included if they had at least 10 hours of available measurements after ICU admission, had at least one of every measurement used for model prediction, and had recorded race/ethnicity data. Bias was assessed through the equal opportunity difference. Model performance in terms of bias and accuracy was compared with the Modified Early Warning Score (MEWS), the Simplified Acute Physiology Score II (SAPS II), and the Acute Physiologic Assessment and Chronic Health Evaluation (APACHE). RESULTS: The machine learning algorithm was found to be more accurate than all comparators, with a higher sensitivity, specificity, and area under the receiver operating characteristic. The machine learning algorithm was found to be unbiased (equal opportunity difference 0.016, P=.20). APACHE was also found to be unbiased (equal opportunity difference 0.019, P=.11), while SAPS II and MEWS were found to have significant bias (equal opportunity difference 0.038, P=.006 and equal opportunity difference 0.074, P<.001, respectively). CONCLUSIONS: This study indicates there may be significant racial bias in commonly used severity scoring systems and that machine learning algorithms may reduce bias while improving on the accuracy of these methods.

Retrospective analysis of the accuracy of predicting the alert level of COVID-19 in 202 countries using Google Trends and machine learning.

BACKGROUND: Internet search engine data, such as Google Trends, was shown to be correlated with the incidence of COVID-19, but only in several countries. We aim to develop a model from a small number of countries to predict the epidemic alert level in all the countries worldwide. METHODS: The "interest over time" and "interest by region" Google Trends data of Coronavirus, pneumonia, and six COVID symptom-related terms were searched. The daily incidence of COVID-19 from 10 January to 23 April 2020 of 202 countries was retrieved from the World Health Organization. Three alert levels were defined. Ten weeks' data from 20 countries were used for training with machine learning algorithms. The features were selected according to the correlation and importance. The model was then tested on 2830 samples of 202 countries. RESULTS: Our model performed well in 154 (76.2%) countries, of which each had no more than four misclassified samples. In these 154 countries, the accuracy was 0.8133, and the kappa coefficient was 0.6828. While in all 202 countries, the accuracy was 0.7527, and the kappa coefficient was 0.5841. The proposed algorithm based on Random Forest Classification and nine features performed better compared to other machine learning methods and the models with different numbers of features. CONCLUSIONS: Our result suggested that the model developed from 20 countries with Google Trends data and Random Forest Classification can be applied to predict the epidemic alert levels of most countries worldwide.

Using a simple open-source automated machine learning algorithm to forecast COVID-19 spread: A modelling study.

INTRODUCTION: Machine learning algorithms have been used to develop prediction models in various infectious and non-infectious settings including interpretation of images in predicting the outcome of diseases. We demonstrate the application of one such simple automated machine learning algorithm to a dataset obtained about COVID-19 spread in South Korea to better understand the disease dynamics. MATERIAL AND METHODS: Data from 20th January 2020 (when the first case of COVID-19 was detected in South Korea) to 4th March 2020 was accessed from Korea's centre for disease control (KCDC). A future time-series of specified length (taken as 7 days in our study) starting from 5th March 2020 to 11th March 2020 was generated and fed to the model to generate predictions with upper and lower trend bounds of 95% confidence intervals. The model was assessed for its ability to reliably forecast using mean absolute percentage error (MAPE) as the metric. RESULTS: As on 4th March 2020, 145,541 patients were tested for COVID-19 (in 45 days) in South Korea of which 5166 patients tested positive. The predicted values approximated well with the actual numbers. The difference between predicted and observed values ranged from 4.08% to 12.77% . On average, our predictions differed from actual values by 7.42% (MAPE) over the same period. CONCLUSION: Open source and automated machine learning tools like Prophet can be applied and are effective in the context of COVID-19 for forecasting spread in naïve communities. It may help countries to efficiently allocate healthcare resources to contain this pandemic.