Skip to main content

UCSF Study Combines AI, Chest Radiographs to Map Out Health Concerns, Expenses

Analysis  |  By Eric Wicklund  
   May 19, 2022

Researchers using deep learning tools to analyze chest radiographs found that they could predict patient costs at one, three and five years and identify high-risk patients, enabling health systems and payers to target care management and preventive health and develop cost and reimbursement plans.

Researchers at the University of California at San Francisco (UCSF) have combined AI tools with chest radiographs to not only help identify patients with potentially serious health issues, but accurately map out their healthcare costs for as much as five years.

The study, published this week in Nature, aims to help healthcare organizations identify patients who will need expensive treatment, allowing them to map out care management plans as well as health and wellness plans. It could also help health systems and payers develop accurate budgeting models for reimbursement.

It also points to the power of machine learning and AI technology in analyzing massive amounts of data to improve not only clinical outcomes but business models.

"This study confirms that radiological imaging indeed contains rich information that may not be routinely extracted by human radiologists but can be analyzed by the power of big data and deep learning," the researchers concluded. "Successfully predicting healthcare expenditure can potentially be an important first step towards improving health policy and medical interventions to address patient care and societal costs."

The study, conducted by a team led by Jae Ho Son and Yixin Chen of UCSF's Center for Intelligent Imaging, used AI technology on 21,872 frontal chest radiographs (CXR) collected from 19,524 patients with at least one year of spending data between 2012 and 2016. The patients were non-obstetric adults who visited the emergency department and received a chest radiograph at the ED or an outpatient facility on that same day.

"The models were developed to identify patients who are likely to incur high healthcare expenditure and predict their subsequent amount of healthcare spending within 1, 3, and 5 years," the study noted. "Unlike physicians who are trained to identify only a handful of imaging biomarkers known to medical literature, our deep learning algorithm is able to take into account thousands of imaging features of weak to moderate correlations with healthcare spending as presented in the training set."

"When a CXR is evaluated by the deep learning algorithm, its pixels are aggregated, transformed, and passed through many layers of filters with each layer extracting different lines, angles, patterns, and associations," the research team said. "As those extracted features are then passed upstream to higher-level filters, they are compared to the thousands of CXR that the algorithm was trained on. All these numbers finally converge to the estimated cost. Considering that CXR tends to be standardized, deep learning algorithms are trained to be extremely sensitive to details that clinical radiologists may not typically recognize."

The researchers noted that the AI platform combines demographic factors, baseline health factors and clinical diseases to map out a patient's current and future cost predictions. This, in turn, can be used to identify high-risk patients who account for the health system's biggest medical expenditures and potentially change that pattern.

"Such predictions can provide an important starting point in identifying high risk patients to achieve reduction in their healthcare spending and encouraging lifestyle modifications and more intensive medical management to achieve better medical and financial outcomes," they noted.

"We believe the use case of the model can go beyond simple actuarial calculation purposes," they wrote. "Though such a model would not be able to provide the precise diagnosis, it can sound an alarm to the patient and primary care doctor that the patient will likely have high healthcare spending in the future. Furthermore, our algorithm could be used in outpatient settings to estimate approximate future healthcare costs such that patients, doctors, and insurance companies would have a reliable indicator to consider when making patient treatment and financial decisions. The identified high-risk patients could be subject to more intensive preventive medical interventions and close follow-up visits to modify patient outcomes."

"The algorithm could also be used to identify patients with CXR that appear normal according to current clinical radiological standards but are still at risk for high medical costs," they added. "Similar to most deep learning algorithms, the application of ours can potentially be automatic, fast, scalable, and relatively low cost when compared to other services in the healthcare system."

Eric Wicklund is the associate content manager and senior editor for Innovation at HealthLeaders.

Get the latest on healthcare leadership in your inbox.