ACR Appropriateness Criteria get new relational database model

February 10, 2003

The American College of Radiology Appropriateness Criteria (ACRAC) are intended to guide radiologists and referring physicians in making initial decisions about diagnostic imaging and therapeutic techniques.For the ACRAC to be relevant to current

The American College of Radiology Appropriateness Criteria (ACRAC) are intended to guide radiologists and referring physicians in making initial decisions about diagnostic imaging and therapeutic techniques.

For the ACRAC to be relevant to current practice, however, it must be in a format that allows it to be used with current electronic systems for patient record keeping, billing/reimbursement, and research databases, said Dr. Chris Sistrom, an assistant professor of radiology at the University of Florida.

Most such computerized systems use a relational database as the main repository for data and much of the business logic. Sistrom proposes a data model that completely encodes all attributes and domains of the published guidelines and is suitable for translation into any industry standard relational database system (Digit Imaging 2002 Nov 6; [epub ahead of print]).

"In the U.S., ICD9 and CPT codes are the standard way to represent symptoms/diseases and medical interventions, respectively," Sistrom said. "Therefore, our data model incorporates these coding schemes and provides for mappings between numeric codes and descriptive information contained in the current ACRAC."

Sistrom has invited the ACR to help fund an effort to encode the entire ACRAC into this database format. One important use would be to maintain the "master" copy of the ACRAC at the ACR. This copy would be maintained online and updated as new sections are added and prior ones modified.

From such a database, it could be easy to produce a "report" essentially matching the current printed document. The database could also serve as the back end for a Web-based decision support tool, he said.

"Newly developing computerized physician order entry systems for radiology will need 'rules' tables, and the ACRAC in database form with ICD9 and CPT code mappings would be ideal for this purpose," Sistrom said.

Sistrom is starting research to translate the neuro section of the ACRAC into his database format. With this in hand, physicians will be able to query the Medicare reimbursement tables to see if the Medicare allowances and exclusions for imaging test/indication pairs have any relationship with the ACRAC recommendation.

"The answer to the question 'if payers reimburse for what' is considered to be standard/best imaging practice by leading experts in radiology," he said.