80 family physicians to participate in study of DI ordering patterns

TORONTO – St. Michael’s Hospital plans to install a clinical decision support system in the hospital’s family practice clinics to study how to improve the appropriateness of imaging tests ordered by physicians.

Healthcare providers across the country suspect a large number of MRIs and CTs ordered and performed in hospitals are unnecessary.

As imaging technologies have advanced, more and more detail has become visible on the scans they produce. This sometimes leads to additional scans for things that aren’t clinically relevant, driving up imaging costs on a per-patient basis.

“Unnecessary tests will also contribute to longer wait times and potentially decrease access for patients who need the tests urgently,” said Kate MacGregor, the Quality Improvement and Radiation Protection Manager in the Department of Medical Imaging.

“Unnecessary tests can also potentially put patients’ health at risk,” she said. “Some tests use ionizing radiation, so from that standpoint it’s a public health problem.”

The clinical decision support system will service 80 family practice physicians and 60,000 patients. The system will use OrderWise software, developed by the Toronto-based company MedCurrent, to gather additional information from family physicians to determine if the test they want to order is appropriate.

“In imaging, sometimes the dilemma is when you think someone has a problem, what’s the best test to do?” said Dr. Bruce Gray, a radiologist at St. Michael’s and one of the project leads. “When it’s decided that an imaging test will be useful for the management of the patient, it raises the question: Is that really true? What is the evidence?”

The concept of the clinical decision support system is relatively simple. When ordering a test, clinicians will enter patient information into a computer, such as symptoms and relevant medical history. The software will then display whether the test should be ordered, or whether a different test would be more appropriate. The clinician will have the option to override the recommendation.

“We’ll then be able to identify how often each clinician is overriding the system, and what types of cases they are overriding most often,” said Dr. Gray. “When we see they’ve been overriding, we can ask, why is that? Maybe they have a legitimate reason for overriding it. We hope to make this a positive and engaging experience for the family practice physicians.”

The tool is designed to be combined with the patient’s electronic health records, said Dr. Gray, with only a few clicks needed to determine whether a test is appropriate.

A number of hospitals throughout North America use OrderWise or similar software to determine whether an imaging test is appropriate. What’s different about the St. Michael’s project, according to Dr. Gray, is that it will be the first site in Canada to use a clinical decision support tool embedded with appropriateness rules tailored for the Canadian context.

The system will also improve the ordering process for imaging tests for both patients and clinicians. Currently, the ordering of imaging tests is almost entirely paper-based and goes through a number of people and departments before an appointment is scheduled.

The clinical decision support system will be added to the ordering process electronically in order to give the patient an appointment at the same time the test is ordered.

Dr. Gray said this increase in efficiency will be a main selling point for family physicians to use the system.

But before the group leading the project can assess the effectiveness of the clinical decision support tool, they first have to determine what the current ordering patterns are.

“One of the biggest problems in radiology is a lack of standardization with regard to how doctors order tests, typically they give very vague reasons on the request form, and then those request forms are scanned in, they’re not put into a database of any sort,” said MacGregor.

“Then when the radiologist dictates the report, after reviewing the imaging tests, there’s no structure for that either, everything is free text. So there is no way to fully understand what the current ordering patterns are, and whether or not they’re appropriate.”

The team is using a natural language processing tool called Montage to understand these baseline ordering patterns and better evaluate the effectiveness of the clinical decision support system when it’s installed in the spring.

Using Montage, the team is finding patterns in the tests ordered by the family health team and individual physicians. They hope to use this information to see if patterns change after the clinical decision support system is installed in the spring.