Stroke patients at a US hospital underwent more than twice as many magnetic resonance imaging (MRI) studies as similar patients treated in a Canadian hospital. At the same time, Canadian stroke patients underwent significantly more CT scans than U.S. patients, a new study shows.
The researchers led by Dr. Max P. Rosen of the Department of Radiology, Beth Israel Deaconess Medical Center, Harvard Medical School, concluded that the availability of imaging modality and patient age were the significant factors in imaging usage for diagnosing stroke patients.
"To better understand some of the determinants of imaging ordering behavior, we analyzed the effect of differential capacity on the imaging workup of patients with acute non-hemorrhagic stroke," Rosen said in a prepared statement. "Our study demonstrates that for patients presenting with symptoms of acute stroke, differences in scanning capacity (CT and MRI) may shape aspects of clinical management," said Rosen.
The study involved a "natural experiment" between the United States and Canada, by comparing the imaging orders of doctors at two large metropolitan hospitals – Beth Israel Deaconess Medical Center, Boston, and Ottawa Hospital, Ottawa, which has two separate inpatient campuses. They retrospectively analyzed the data from the inpatient admissions via the emergency department between Oct. 1, 2002, and Dec. 31, 2006.
To reduce the number of confounding factors they compared only those patients treated at the two hospitals for nonhemorrhagic stroke because this population is relatively homogenous and is easily identified retrospectively through hospital administrative records. After excluding patients who had not been admitted for the first time for stroke, there were 918 patients in the U.S. group and 1, 759 patients in the Canadian group analyzed for this study.
The US hospital had six MRI scanners and seven CT scanners, both were available 24 hours per day, seven days per week. The Canadian hospital had two MRI scanners and five CT scanners. MRI was available 24 hours per day Monday through Friday and eight hours per day on Saturday and Sunday. CT was available 16 hours per day Monday through Friday and eight hours per day on Saturday and Sunday.
Compared with the Canadian hospital, the US hospital study population had a greater proportion of patients with atrial fibrillation, coronary artery disease, diabetes, and hypertension. However, the mean Charlson indices were identical for the two populations, indicating that the burden of disease in the two study populations was similar.
The rate of MRI scans at the US hospital was more than twice that at either of the Canadian hospitals with 95.75 scans per 100 patients vs 41.39 scans per 100 patients. Patients at the Canadian hospital underwent a significantly greater number of head and neck CT scans with 162.59 studies per 100 patients compared to patients at the US hospital with109.48 studies per 100 patients.
Comparing overall imaging, the US hospital utilized MRI significantly more with MRI composing 20.98 percent of all imaging compared to 9.11 percent of all imaging at the Canadian hospital. The Canadian hospital conducted the greatest proportion of CT studies with 36 percent of all imaging at the Canadian hospitals compared to 26 percent at the US hospital. X-rays composed 49 percent of the imaging done at the US hospital compared to 39 percent of the Canadian hospital.
Of the patients admitted to the US hospital 89.4 percent underwent CT or MRI studies of the head and neck compared to 78.5 percent patients admitted to the Canadian hospital. Patients at the US hospital were also significantly more likely to undergo both CT and MRI (44.9 percent vs 8.0 percent).
Imaging intensity, meaning the number of imaging studies a patient underwent while in the hospital, was substantially higher at the US hospital. Although neurologic imaging per patient was roughly equivalent across all sites, differences in average length of stay suggest that a patient at the US hospital underwent on average 2 neurologic imaging examinations every 5 days, whereas patients at the Canadian hospital underwent on average 1 imaging study every 6 days. The average length of stay in the US hospital was 4.69 days compared to 13.3 days in the Canadian hospital.
While the study was not designed to compare outcomes, the researchers noted that the inpatient mortality rate at the US hospital was 7.1 percent, significantly lower than the 11.3 percent inpatient mortality rate at the Canadian hospital. Significantly more patients at the US hospital were discharged home with support services, rehabilitation or acute care, and long-term care. Whether this outcome was linked to imaging difference, the researchers could not say.
“Our study design allowed us to capture outcome data occurring while patients were in the hospital,” they wrote. “Because patients at the US site were discharged earlier, it is possible that we missed more outcome events in this patient population. Alternatively, it is possible that the overall quality of care had an impact on outcomes. For example, a stroke protocol was used at the US site, but not at the Canadian site, and the US site demonstrated a greater use of anticoagulants, two factors shown to be associated with improved clinical outcomes.”
The researchers noted that despite the differences in modality, the average number of neurologic imaging studies per patient (all modalities) was roughly equal at both sites. There were 205.23 studies per 100 patients at the US hospital vs 203.98 studies per 100 patients at the Canadian hospital. They concluded that access to imaging services may affect utilization.
According to background information in the study, in 2005, Canada had 11.3 CT scanners per million people, compared to 32.2 CT scanners per million people in the United States. The US-Canadian disparity in MRI scanners is much larger: 5.5 scanners per million people in Canada in 2005 compared with 26.6 scanners per million people in the United States in 2004. Compared with a typical US physician, a typical Canadian physician must deal with relatively limited total scanning capacity, in particular limited MRI capacity.
By Michael O’Leary, contributing writer, Health Imaging Hub