Although machine learning is increasingly being applied to support clinical decision making, there is a significant gap in understanding what it is and how nurses should adopt it in practice. The purpose of this case study is to show how one application of machine learning may support nursing work and to discuss how nurses can contribute to improving its relevance and performance. Using data from 130 specialized hospitals with 101 766 patients with diabetes, we applied various advanced statistical methods (known as machine learning algorithms) to predict early readmission. The best-performing machine learning algorithm showed modest predictive ability with opportunities for improvement. Nurses can contribute to machine learning algorithms by (1) filling data gaps with nursing-relevant data that provide personalized context about the patient, (2) improving data preprocessing techniques, and (3) evaluating potential value in practice. These findings suggest that nurses need to further process the information provided by machine learning and apply “Wisdom-in-Action” to make appropriate clinical decisions. Nurses play a pivotal role in ensuring that machine learning algorithms are shaped by their unique knowledge of each patient's personalized context. By combining machine learning with unique nursing knowledge, nurses can provide more visibility to nursing work, advance nursing science, and better individualize patient care. Therefore, to successfully integrate and maximize the benefits of machine learning, nurses must fully participate in its development, implementation, and evaluation.
Author Affiliations: Schools of Nursing (Mr Kwon and Dr Currie) and Population and Public Health (Dr Karim), University of British Columbia; and Centre for Health Evaluation and Outcome Sciences (CHÉOS), Providence Health Care Research Institute (Dr Karim), Vancouver, British Columbia, Canada; School of Nursing and Data Science Institute, Columbia University (Dr Topaz); and Brigham and Women's Health Hospital, Boston, MA (Dr Topaz).
The authors have disclosed that they have no significant relationships with, or financial interest in, any commercial companies pertaining to this article.
Corresponding author: Jae Yung Kwon, MSN, RN, University of British Columbia, T201-2211 Wesbrook Mall, Vancouver, British Columbia, Canada V6T 2B5 (firstname.lastname@example.org).