Semiparametric Mixed Models for Nested Repeated Measures Applied to Ambulatory Blood Pressure Monitoring Data

Area of Science:

• Biostatistics
• Medical Device Analysis
• Clinical Research Statistics

Background:

• Semiparametric mixed models are valuable for analyzing complex longitudinal data in medical device studies.
• Nested repeated measures (NRM) arise from monitoring individuals over time, particularly before and after interventions.
• Accurate statistical modeling is crucial for interpreting outcomes in clinical trials.

Purpose of the Study:

• To apply semiparametric regression with penalized splines to model NRM data.
• To estimate mean profiles and account for complex covariance structures in longitudinal measurements.
• To evaluate the impact of surgical intervention for obstructive sleep apnea on 24-hour ambulatory blood pressure.

Main Methods:

• Utilized semiparametric mixed models incorporating penalized splines for mean profile estimation.
• Developed and applied covariance models specifically designed to handle nested repeated measures (NRM).
• Analyzed prospective data from a study on 24-hour ambulatory blood pressure following surgical intervention.

Main Results:

• The proposed covariance models effectively accounted for NRM in the blood pressure data.
• Penalized splines provided flexible estimation of the mean blood pressure profiles over time.
• Significant changes in ambulatory blood pressure were observed post-surgical intervention for sleep apnea.

Conclusions:

• Semiparametric mixed models offer a robust framework for analyzing NRM in medical device studies.
• The methodology successfully demonstrated the impact of surgical intervention on blood pressure regulation.
• This approach enhances the statistical rigor for evaluating interventions in longitudinal clinical studies.