In a two-year study, researchers at Harvey L. Neiman Health Policy Institute were able to more easily and accurately determine the likelihood of a patient undergoing CT, MR, or PET scans using a new AI-powered comorbidity index that is the first specifically designed for advanced imaging, providing them with a better understanding of utilization rates among these modalities.

Dubbed the Neiman Imaging Comorbidity Index (NICI), the solution was trained and validated on claims data of 10.5 million individuals from Optum's deidentified Clinformatics Data Mart database to examine different variables that together can create an accurate picture of a patient's chances of undergoing advanced radiology scans.

Because it is trained on claims data, the NICI can statistically control the likelihood that a patient will undergo advanced imaging. Using data on acute and historical health conditions in these claims, the index calculates a score, identifying certain conditions as more indicative of the need for imaging than others. This allows it to statistically control and adjust for differences in a patient's chances of receiving advanced imaging, eliminating biases found with other comorbidity indexes, such as the Charlson comorbidity index (CCI).
Designed to predict mortality in cancer patients but not broad patient populations, CCI is often used in imaging research due to the absence of an image-specific comorbidity index. The researchers limited the number of comorbidities that the NICI evaluated to nine statistically correlated to advanced imaging, making it easier to use compared to the CCI.

"In comparison, the CCI has 17 comorbidities, yet the NICI is more predictive of advanced imaging use. This is not a criticism of the CCI, but rather, a benefit of the imaging-specific NICI, a simple and effective tool for researchers," said Casey Pelzl, lead author and senior economics and health services analyst at the Neiman Health Policy Institute.

Out of the 10.5 million individuals used as development and validation cohorts for the NICI, 2.1 million underwent advanced imaging. A numerical index was created in the development cohort, which consisted of 70% of the total data set, with different weights assigned to each comorbidity.

It was then pitted against the CCI to assess the other 30% of the database and proved better at distinguishing between those who received advanced imaging and those who did not. Among the factors it assessed were age and sex, allowing it to better discriminate between these two groups.

According to Pelzl, the index outperformed the CCI across all patient groups, regardless of age. “The NICI is most predictive of imaging use for older individuals who, on average, have more comorbidities recorded in the claims data that were used."

Future studies are needed to externally validate the NICI.

The findings were published in the Journal of the American College of Radiology.

CT Homepage