“…Therefore, the Generalized Estimating Equations (GEE) approach (Liang and Zeger, 1986) provides a convenient alternative to maximum likelihood estimation, especially for longitudinal categorical and ordinal data (Lumley, 1996;Parsons et al, 2006;Touloumis et al, 2013;Nooraee et al, 2014;Ditlhong et al, 2018;da Silva et al, 2019). Marginalized models, integrating the random effects in the likelihood functions to get the estimation for marginal regression parameters, have also been explored for longitudinal ordinal data (Lee and Daniels, 2008;Lee et al, 2016;Schildcrout et al, 2022).…”