Generalized semiparametric varying-coefficient model for longitudinal data with applications to adaptive treatment...

Area of Science:

• Statistics
• Biostatistics
• Longitudinal Data Analysis

Background:

• Longitudinal data analysis requires flexible models to capture complex covariate effects.
• Existing models may not adequately address time-constant, time-varying, and covariate-varying effects simultaneously.
• Understanding treatment efficacy in HIV patients requires robust statistical methods.

Purpose of the Study:

• To develop and validate a generalized semiparametric varying-coefficient model for longitudinal data.
• To flexibly model time-constant, time-varying, and covariate-varying effects.
• To analyze the impact of antiretroviral treatment switching in HIV patients.

Main Methods:

• A generalized semiparametric varying-coefficient model was proposed.
• Local linear smoothing and profile weighted least squares estimation were employed.
• Hypothesis testing for parametric functions and bandwidth selection were developed.

Main Results:

• The proposed model demonstrated satisfactory finite sample performance in simulations.
• Analysis of the ACTG 244 trial revealed benefits of switching antiretroviral therapy.
• Treatment switching to combination therapies showed advantages over Zidovudine (ZDV) monotherapy.

Conclusions:

• The developed semiparametric varying-coefficient model offers a flexible approach for longitudinal data.
• Antiretroviral treatment switching is beneficial for HIV patients, particularly after resistance mutations emerge.
• The findings support strategic treatment modifications in managing HIV infection.