Subjective aspects such as oral health-related quality of life (OHRQoL) and depression are important aspects in the periodontal care. The objectives of the study were to test a predictive model of clinical attachment loss and OHRQoL in a pooled sample of dental patients with periodontitis and mental health patients with depressive symptomatology, and test the invariance of the model across both types of patients. Three self-report scales were applied to assess depression, OHRQoL and oral hygiene habits, saliva samples were collected for three proinflammatory biomarkers, and the clinical attachment loss was measured in 35 patients with periodontitis and 26 patients with depressive symptomatology. Data were analyzed through structural equation modeling. The one-group analysis revealed a psychosomatic complaint model of disagreement between the complaint and the clinically observable. In the multi-group analysis, the model was not invariant. It was necessary to introduce a singularity in relation to depressive symptomatology for each population. Thus, a good and equivalent fit was achieved between the six nested models in constraints, as well as equivalent parameters between both types of patients. The study of a dental population in conjunction with a mental health population with a psychosomatic risk factor reveals interesting and unexpected results.