Abstract
Aim:
To asses the long-term effectiveness of asthma treatments by comparing the utility of repeated measure models (RMM) and multilevel models (MLM) in analyzing longitudinal data of pulmonary function measured by forced expiratory volume in one second (FEV1), over an extended periods.
Subject and Methods:
Seventy-two asthma patients were randomized into three groups: standard drug (a), test drug (c), and placebo (p), with 24 patients each. Forced expiratory volume (FEV1) was measured hourly for 8 hours post-treatment, plus a baseline measurement. Repeated measure models (RMM) and Multilevel models (MLM) were used to analyze forced expiratory volume (FEV1) changes over time.
Results:
The repeated measures model with an unstructured covariance matrix proved most effective, as indicated by Akaike Information Criterion (AIC) of 342.45, Bayesian Information Criterion (BIC) of 445, and corrected AIC (AICC) of 349.7. This model displayed a correlation decrease in forced expiratory volume (FEV1) from 0.7124 to 0.6429 over 8 hours, with a standard error of 0.1448.
Conclusion:
The study supports the use of repeated measures models with an unstructured covariance matrix for analyzing the efficacy of asthma treatments over time. This model effectively captured the dynamics of treatment effects on respiratory function, adhering to assumptions such as linearity, homoscedasticity, normality, and absence of significant outliers, thereby providing robust and reliable results.