Key factors help AI predict whether patients will follow up for incidental lung nodules
A few key factors can help artificial intelligence predict whether patients will return for additional services after imaging unearths incidental findings.
That’s according to new research from the University of Pennsylvania, published recently in the Journal of the American College of Radiology [1]. Incidental pulmonary nodules are a common finding on chest CT—occurring in upward of 51% cases. However, as few as 40% of patients address this concern, with poor adherence leading to delayed diagnoses and worse outcomes.
To improve upon these numbers, Penn researchers used an AI algorithm on a dataset of over 1,600 patients, hoping to forecast patients’ chances of addressing their incidental findings.
“To our knowledge, our study is the first in the field to construct prediction models that integrate patient demographic and socioeconomic in addition to clinical information and clinical setting to predict adherence to follow-up visits for IPNs,” Eduardo J. Mortani Barbosa Jr., MD, MBA, with Penn’s Department of Radiology, and co-authors wrote March 8. “[Machine learning] models like the ones developed and discussed in this study can become a key component to improve early diagnosis of malignancies with potential to reduce morbidity and mortality, while saving downstream and future healthcare costs.”
The study assessed 13 potential predictors and their influence on patients’ actions. Those included sex, race, insurance, employment, marriage, education and income. All except for comorbidity proved to have significant association with follow-up, Mortani et al. reported. Inpatient or emergency clinical context, along with high nodule risk, were the most meaningful predictors of patient follow-up. Sex, race and marital status, meanwhile, became more impactful if clinical context was removed from the AI model. Clinical context was associated with sex, race, insurance, employment, marriage, income, nodule risk and smoking status, “suggesting its role in mediating socioeconomic inequities.”
Mortani and co-investigators also tested three additional machine learning models, and all four exhibited solid, comparable predictive performance.
“In summary, our study demonstrates that clinical context and socioeconomic factors can predict a patient’s IPN follow-up adherence across four different machine learning models, demonstrating the utility of artificial intelligence tools in improving patient management and outcomes, through predictive analytics and targeted health systemwide interventions,” the authors concluded.
Read more, including potential study limitations, at the link below.
Marty Stempniak has covered healthcare since 2012, with his byline appearing in the American Hospital Association's member magazine, Modern Healthcare and McKnight's. Prior to that, he wrote about village government and local business for his hometown newspaper in Oak Park, Illinois. He won a Peter Lisagor and Gold EXCEL awards in 2017 for his coverage of the opioid epidemic.