Collaboration creates guidelines to avoid unnecessary testing on prostate cancer patients
Bone and CT scans can tell us if a patient’s prostate cancer has spread beyond the prostate and into surrounding bone or tissue. However, the scans themselves are expensive, and expose patients to harmful radiation. The scans often result in false positives, which lead to additional expensive and sometimes painful procedures.
But most importantly, the scans are not always necessary. Yet, there hasn’t been a consistent set of guidelines to help physicians determine which patients need the scans.
So participants in the Michigan Urological Surgery Improvement Collaborative (MUSIC) CQI set out to develop a tool that would help participants determine which patients need the additional scans, and which do not.
Working with the University of Michigan College of Engineering, a team of urologists, engineers and researchers used data analytics and predictive modeling to determine the maximum thresholds for a patient’s PSA level and cancer stage. Patients with values below those thresholds do not need the scans. Those with one or two values above the threshold should be tested.
A pilot study to test the guidelines resulted in a 40 percent reduction of bone and CT scans, a prevention of $275,000 in testing costs, and an increase in metastatic cancers detected.
The data for the study came from the MUSIC initiative. With close to 90 percent of urologists in Michigan participating in the initiative, years of extensive data has been collected on patients throughout the state. The group looked at medical records from patients who had undergone a bone or CT scan, and considered many different factors including age, levels of PSA, risk factors and types of prostate cancer.
Then, they looked at which patients tested positive. Next, they used predictive modeling, using past results to predict which future patients are likely to be at high risk. Finally, they created a simple set of guidelines to determine which patients should be screened.
“I think it shows the power of collaboration between medicine and engineering, and I’m looking forward to seeing how data and predictive analytics can help improve outcomes and efficiency in other areas as well,” says David Miller, MD, a project director with MUSIC and urologist at Michigan Medicine.
The group hopes to adapt the guidelines to other regions of the country. They also are looking at how to use predictive modeling to help determine when patients should get a prostate biopsy.
Michigan Engineering produced a video about the project. See it here.