Estimating the association between blood pressure variability and cardiovascular disease: An application using the...

Area of Science:

• Cardiovascular Medicine
• Biostatistics
• Epidemiology

Background:

• Visit-to-visit systolic blood pressure variability is linked to cardiovascular events.
• Current estimation methods are prone to significant measurement error.
• This error can bias risk estimates due to regression dilution.

Purpose of the Study:

• To demonstrate bias in current blood pressure variability estimation methods.
• To propose and evaluate alternative statistical methods for reducing bias.
• To improve the accuracy of predicting cardiovascular events using blood pressure variability.

Main Methods:

• A two-stage method using mixed-effects models and time-to-event analysis.
• A joint model simultaneously fitting mixed-effects and time-to-event submodels.
• Utilizing data from the Atherosclerosis Risk in Communities (ARIC) study.

Main Results:

• Standard methods for estimating blood pressure variability are subject to regression dilution bias.
• The proposed two-stage and joint models effectively reduce this bias.
• Accurate estimation of blood pressure variability improves cardiovascular risk prediction.

Conclusions:

• Addressing measurement error in blood pressure variability is crucial.
• The two-stage and joint models offer improved statistical approaches.
• These methods enhance the reliability of associations between blood pressure variability and cardiovascular outcomes.