Objectives: Diagnosis of the wide variety of lesions that occur in the oral cavity is an essential part of dental practice. The objective of this study was to explore the demographic and behavioural profiles of patients affected by oral cavity diseases on a wide spectrum.Methods: Demographic and behavioural data of 3154 patients with at least one oral cavity disease among those consecutively visited between 2004 and 2009 at the Unit of Oral Medicine, were recorded: gender, age, smoking status, alcohol consumption, drugs use. Analysis was restricted to the eleven more frequent oral cavity diseases. Firstly, Multiple Correspondence Analysis (MCA) was applied as a descriptive tool to geometrically characterize which diseases are more or less similar in terms of their demographical and behavioural profiles in a sub-space of low-dimensionality (Benzecri, 1992; Greenacre, 1993); Secondly, the simultaneous marginal homogeneity hypothesis (SMH) for the multivariate distributions associated to the detected profiles was tested (Agresti and Klingenberg, 2005; Lang, 2004).Results: On the first principal MCA axis, similar profiles were found for the couples Carcinoma-Leucoplachia, Alitosi-SAR and for the couple BMS-ONM. The score-type test of SMH and the goodness-of-fit tests of SMH models confirmed that profiles of some oral cavity disorders could not be considered significantly different at a significance level of 5%. Conclusion: Simultaneously modelling joint and marginal distributions of multivariate binary response contributed a very important issue to preventive oral medicine, in other words some oral mucosal lesions can be considered similar in terms of demographical and behavioural profiles.
|Numero di pagine||0|
|Stato di pubblicazione||Published - 2010|