CHICAGO – A clinical decision-support tool designed to identify hospitalized patients who need thromboembolism prophylaxis and embedded in a hospital’s electronic health record led to significantly more appropriate prophylaxis, compared with usual care, and significantly cut the 30-day rate of thromboembolism in a randomized, multicenter trial with more than 10,000 patients.
“This is the first time that a clinical decision support tool not only changed [thromboprophylaxis prescribing] behavior but also affected hard outcomes. That’s remarkable,” lead investigator Alex C. Spyropoulos, MD, said in an interview.
Even so, outside experts expressed concerns about certain results and the trial design.
Use of the decision-support risk calculator for thromboembolism in the IMPROVE-DD VTE trial significantly boosted use of appropriate inpatient thromboprophylaxis starting at hospital admission by a relative 52%, and significantly increased outpatient thromboprophylaxis prescribed at discharge by a relative 93% in the study’s two primary endpoints, Spyropoulos reported at the American Heart Association scientific sessions.
This intervention led to a significant 29% relative reduction in the incidence of total thromboembolic events, both venous and arterial, during hospitalization and through 30 days post discharge.
The absolute thromboembolic event rates were 2.9% among 5,249 patients treated at either of two U.S. hospitals that used the EHR-based risk calculator and 4.0% in 5,450 patients seen at either of two other U.S. hospitals that served as controls and where usual care method identified patients who needed thromboprophylaxis, said Spyropoulos, professor and director of the anticoagulation and clinical thrombosis services for Northwell Health in New York. This included a 2.7% rate of venous thromboembolism and a 0.25% rate of arterial thromboembolism in the intervention patients, and a 3.3% rate of venous events and a 0.7% rate of arterial events in the controls.
Patients treated at the hospitals that used the EHR-embedded risk calculator also has a numerically lower rate of major bleeding events during hospitalization and 30-day postdischarge follow-up, a 0.15% rate compared with a 0.22% rate in the control patients, a difference that was not significant.
A “Powerful Message”
“It’s a powerful message to see an absolute 1.1% difference in the rate of thromboembolism and a trend to fewer major bleeds. I think this will change practice,” Spyropoulos added in the interview. “The next step is dissemination.”
But thromboprophylaxis experts cautioned that, while the results looked promising, the findings need more analysis and review, and the intervention may need further testing before it’s ready for widespread use.
For example, one unexpected result was an unexpected 2.1 percentage point increase in all-cause mortality linked with use of the decision-support tool. Total deaths from admission to 30 days after discharge occurred in 9.1% of the patients treated at the two hospitals that used the risk calculator and 7.0% among the control patients, a difference that Dr, Spyropoulos said was likely the result of unbalanced outcomes from COVID-19 infections that had no relevance to the tested intervention. The trial ran during December 2020–January 2022.
But Wait — More Detail and Analysis Needed
“I’d like to see more analysis of the data from this trial,” and “there is the issue of increased mortality,” commented Gregory Piazza, MD, director of vascular medicine at Brigham and Women’s Hospital in Boston, and a specialist in thromboembolism prevention and management. He also highlighted the need for greater detail on the arterial thromboembolic events tallied during the study.
With more details and analysis of these findings “we’ll learn more about the true impact” of this intervention, Piazza said in an interview.
“The increased mortality in the intervention group may have been due to differential treatment and decision-making and confounding and warrants further investigation,” commented Elaine M. Hylek, MD, a professor at Boston University and designated discussant for the report. Selection bias may have contributed to this possible confounding, Hylek noted.
Other limitations of the study cited by Hylek included its reliance on individual clinician decision-making to actually prescribe thromboprophylaxis, a lack of information on patient adherence to their thromboprophylaxis prescription, and an overall low rate of appropriate thromboprophylaxis prescribed to patients at discharge. The rates were 7.5% among the controls and 13.6% among patients in the intervention arm. For prescription at the time of hospitalization, the rates were 72.5% among control patients and 80.1% for patients seen at the two hospitals that used the decision-support tool.
The IMPROVE-DD VTE Risk Assessment Tool
The clinical decision-support tool tested is called the IMPROVE-DD VTE risk assessment model, developed over several years by Spyropoulos and associates; they have also performed multiple validation studies. The model includes eight factors that score 1-3 points if positive that can add up to total scores of 0-14. A score of 0 or 1 is considered low risk, 2 or 3 intermediate risk, and 4 or more high risk. One of the scoring factors is the result of a D-dimer test, which explains the DD part of the name.
The eight factors and point assignments are prior venous thromboembolism: 3 points; known thrombophilia: 2 points; lower limb paralysis: 2 points; current cancer: 2 points; d-dimer level more than twofold the upper limit of normal: 2 points; immobilized for at least 7 days: 1 point; admitted to the ICU or coronary care unit: 1 point; and age greater than 60 years old: 1 point.
Development of the IMPROVE-DD VTE risk calculator received most of its funding from the U.S. Agency for Healthcare Research and Quality, and the risk tool will be available for hospitals and health systems to access at no charge through the agency’s website, Spyropoulos said. The researchers designed the calculator to operate in any EHR product.
IMPROVE-DD VTE “is a very valid, high-quality tool,” commented Piazza. “We’ve used some rather blunt tools in the past,” and especially praised inclusion of D-dimer results into the IMPROVE-DD VTE model.
“It’s nice to use a biomarker in addition to clinical factors,” he said. “A biomarker provides a more holistic picture; we can’t do genetic testing on every patient.”
Enrollment Focused on Higher-Risk Patients
The study ran at four academic, tertiary-care hospitals in the Northwell Health network in the New York region. It enrolled patients aged more than 60 years who were hospitalized for any of five diagnoses: heart failure; acute respiratory insufficiency, including chronic obstructive lung disease or asthma; acute infectious disease, including COVID-19; acute inflammatory disease, including rheumatic disease; or acute stroke. The study excluded patients with a history of atrial fibrillation, those who used an anticoagulant at home, or those who had received therapeutic anticoagulation within 24 hours of their hospital admission.
The anticoagulant prophylaxis that patients received depended on their calculated risk level – intermediate or high – and whether they were inpatients or being discharged. The anticoagulants that clinicians could prescribe included unfractionated heparin, enoxaparin, fondaparinux, rivaroxaban, and apixaban.
“We’ve been looking for a long time for a tool for medically ill patients that’s like the CHA2DS2-VASc score” for patients with atrial fibrillation. “These powerful data say we now have this, and the EHR provides a vehicle to easily implement it,” Spyropoulos said.
The IMPROVE-DD VTE study received partial funding from Janssen. Spyropoulos has been a consultant to Nayer, Boehringer Ingelheim, Bristol-Myers Squibb, Janssen, Pfizer, and Sanofi; adviser to the ATLAS Group; and has received research support from Janssen. Piazza has received research funding from Bayer, BIG/EKOS, BMS, Janssen, and Portola. Hylek had been a consultant to Bayer and Ionis, and has received honoraria from Boehringer Ingelheim and Pfizer.
This article originally appeared on MDedge.com, part of the Medscape Professional Network.
Source: Read Full Article