5 recent studies exploring AI in healthcare

Medical researchers are becoming increasingly interested in artificial intelligence's potential to improve healthcare by reducing workflow inefficiencies, predicting health outcomes and speeding up diagnoses.

Below are five key AI studies that have been published recently:

  1. Machine learning of patient characteristics to predict admission outcomes in the undiagnosed diseases network: Researchers developed a machine learning algorithm to determine whether to accept patients to the Undiagnosed Diseases Network for extensive genome-scale evaluation. They found that the admission process could be accelerated by up to 68 percent using the algorithm, which would allow for more applications processed in a given time frame.

  2. Use of machine learning models to predict death after acute myocardial infarction: The research team developed machine learning methods to improve the prediction of in-hospital death after hospitalization for acute myocardial infarction. They found their models were not associated with significantly better prediction of risk of death after acute myocardial infarction, but they could improve the resolution of risk, which can better clarify individuals' risk for adverse outcomes.

  3. A survey of clinicians on the use of artificial intelligence in ophthalmology, dermatology, radiology and radiation oncology: A survey about personal use of clinical AI tools was conducted among fellows and trainees of three specialty colleges in Australia and New Zealand. The majority of respondents, who specialized in either ophthalmology, dermatology or radiology, believed the integration of such tools would improve their field.

  4. Semi-automated tracking of pain in critical care patients using artificial intelligence: a retrospective observational study: The research team conducted pain assessments on critically ill patients using an AI-powered tool they had developed, which used three machine learning methods: random forest, support vector machine and logistic regression. Random forest demonstrated the highest efficacy.

  5. Morphological and molecular breast cancer profiling through explainable machine learning: Researchers presented a machine-learning approach to integrated profiling of morphological, molecular and clinical features from breast cancer histology. Their approach's olecular predictions reached balanced accuracies up to 78 percent overall and 95 percent for certain subgroups of patients. 

More articles on artificial intelligence: 
AMA's 7 tips for responsible AI use in healthcare
5 data issues limiting AI's potential in healthcare
AI vs. machine learning vs. algorithms: Providence exec explains the differences, their healthcare applications

© Copyright ASC COMMUNICATIONS 2021. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.