Análise das solicitações de acompanhamento funcional: um estudo na Universidade Federal de Minas Gerais

Abstract

This study examines the practice of Organizational Support (AF, from portuguese Acompanhamento Funcional) at Universidade Federal de Minas Gerais (UFMG), focusing on its relationship with sociodemographic and functional variables of staff members. AF is a multidisciplinary strategy primarily adopted by public institutions planned to contribute to and improve managerial, functional, and relational aspects in workplace environments. In addition to the absence of scientific literature on the subject, this study is justified by the growing need to use data for expanding and enhancing preventive actions. The study aimed to understand the factors influencing the request for AF at UFMG. Secondary data from Technical-Administrative in Education staff (TAE) and Teaching staff (MAG) were used, covering the years from 2017 to 2023. The analysis was conducted using Generalized Linear Mixed Models (GLMM), which are flexible tools for incorporating fixed and random effect variables in complex, repeated, clustered or longitudinal data, where the distribution is not necessarily Normal. Residual analysis was performed using the Probability Integral Transformation (PIT) method and quantile-quantile (Q-Q) plot enveloping, which indicated the model's ability to capture relationships between explanatory variables and the response variable, demonstrating strong robustness for discrimination. Results for the TAE group indicated that staff members aged 40 to 49, who are unmarried, with less than three years at the institution were associated with higher chances of request AF. Conversely, an age above 50 and higher educational levels, such as a university degree or postgraduate studies, were associated with lower chances. For the MAG group, unmarried staff members aged 40 to 49 were associated with higher chances of request AF, while male staff members were associated with lower chances. Estimates of random effects also indicated that work in specific units and the year significantly impacted the chances of requesting AF in both groups. The ROC curve analysis also demonstrated good model discrimination, with the TAE group model showing a greater ability to correctly identify those who requested AF (sensitivity), while the MAG group model demonstrated a greater ability to correctly identify those who did not request AF (specificity).