This article is more than 5 years old. Information may no longer be current.
Allocation bias may mask true benefit of primary PCI in registries
Click Here to Manage Email Alerts
Study findings have shown that allocation bias of high-risk STEMI patients may explain why PCI in registry-based comparative effectiveness studies may appear to be less effective than in randomized controlled trials.
Researchers analyzed 55,022 patients from 11 registries to determine whether higher-risk patients (Killip class ≥1) were being preferentially allocated to one therapy. Eight of the 11 registries had documented sufficient data to be included in the analysis.
The difference in mortality between primary PCI and fibrinolysis depended strongly on how the high-risk patients were allocated, with higher-risk patients being preferentially allocated to PCI. Unequal distribution of high-risk patients accounted for most of the variation between registries (adjusted R2 meta=83.1%). After adjusting for unequal allocation of high-risk patients, primary angioplasty had a 22% lower mortality rate (OR=0.78; 95% CI, 0.64-0.97).
The researchers also derived a formula, termed the number needed to abolish, in which they detect when a disease is vulnerable to this form of bias in evaluation of its treatments.
“The conflict between [randomized controlled trial] and registry meta-analyses of PCI vs. fibrinolysis can be explained by allocation bias of just a few higher-risk patients to the primary PCI arm in registries,” the researchers concluded. “We present a quantitative formula highlighting two factors that make registry studies comparing interventions vulnerable to bias: the ability of clinicians on the spot to identify a high-risk subset and a tendency to preferentially allocate them to one therapy.”
Disclosure: The researchers report no relevant financial disclosures.
Click Here to Manage Email Alerts