This article is more than 5 years old. Information may no longer be current.
Stroke, bleeding risk estimations by Canadian PCPs lacking in number, accuracy
Click Here to Manage Email Alerts
In a large population of patients with atrial fibrillation in Canada, primary care physicians failed to assign estimated risks for stroke or bleeding in a significant percentage, according to recent findings.
“There is underestimation of stroke and bleeding risk and overestimation of stroke and bleeding risk in a significant proportion of patients,” the researchers wrote. “An opportunity exists to enhance knowledge translation to PCPs.”
In a chart audit of data from the FREEDOM AF registry — a national knowledge translation program aimed at improving the treatment of AF in Canada — Paul Angaran, MD, of the University of Toronto, and colleagues analyzed data on 4,670 patients aged at least 18 years without significant valvular heart disease from 474 primary care practices.
The investigators conducted descriptive analysis of demographic variables, AF-related variables and risk score calculations. For the evaluation of PCP and calculated risk agreement, CHADS2 and HAS-BLED scores (0-1, low risk; 2, intermediate risk; ≥ 3, high risk) were classified. The researchers used the weighted kappa statistic to evaluate agreement between PCP-estimated risks for stroke and bleeding and the calculated scores.
PCP-estimated stroke risk was low in 864 (18.5%) patients, intermediate in 1,606 (34.4%) and high in 1,523 (32.6%) patients. In 15% of patients, a stroke risk estimate was not provided; 9.3% were designated as “do not know” and 5.2% had “missing” information.
The breakdown of CHADS2 score in the study population was: 0 (6.3%); 1 (23%); 2 (32.4%); 3 (19.1%); 4 (12.6%); 5 (4.7%); and 6 (1.1%). The median CHADS2 score was 2. The investigators found only 50% of cases were evaluated using a formal stroke risk assessment tool. Of the 864 patients classified as low risk by PCP estimate, 561 (65%) were calculated as low risk based on CHADS2 score, indicating that stroke risk was underestimated in 35% of patients.
In terms of bleeding risk, PCPs estimated low risk in 1,952 (41.8%) patients, intermediate risk in 1,133 (24.3%) patients and high risk in 437 (9.4%) patients. In 24.6%, the PCPs did not provide a bleeding risk estimate; 15.3% were classified as “do not know” and 9.3% had “missing” information.
The distribution of the calculated HAS-BLED scores was: 0-1 (17.2%); 2 (37.2%); and at least 3 (45.6%). Overall, the median HAS-BLED score was 2. A formal bleeding risk evaluation tool was used by PCPs in only 26% of patients. Of the 1,942 patients classified as low bleeding risk by their PCPs, 428 (21.9%) were deemed at low risk by HAS-BLED score; thus, PCPs underestimated the risk of bleeding in 78.1% of this patient group.
Warfarin was used for antithrombotic treatment in 90% of patients, and the distribution of time in the therapeutic range among the population was: 24% with TTR < 50; 9% with TTR between 50% and 60%; 11% with therapeutic range between 60% and 70%; and 56% with therapeutic range of at least 70%.
In a related editorial, Laurent Macle, MD, and Jason G. Andrade, MD, both of the Universite de Montreal, wrote that these findings suggest a disparity between empirical risk prediction tools and physician assessments.
“The study suggests that the decision to initiate [oral anticoagulants] is complex and considers many factors beyond simple risk prediction tools,” Macle and Andrade wrote. “In other words, the assessment of risk involves an educated, informed, and balanced assessment of individualized risk with a weighting of consequences. Given this complexity, the need exists for future knowledge translation activities with respect to the management of AF and stroke prevention, as well as follow-up studies to ensure that these knowledge translation activities are effecting appropriate changes in practice.” - by Jennifer Byrne
Disclosure: Angaran, Macle and Andrade report no relevant financial disclosures.
Click Here to Manage Email Alerts