In this paper we describe a graph, tree and forest model of psychosocial factors dependencies of chronically ill patients, called graphical models. Foundation of the study was the theory of meaningfulness of suffering by V. E. Frankl. 181 patients with either arterial hypertension or neoplasms with bad prognosis were examined thrice: 0-10 days from the time of diagnosis (stage I), about 5 weeks from the diagnosis (stage II) and at a follow-up about 5 months since stage II (stage III). 75 factors were available for consideration: 17 in stage I, 28 in stage II, 27 in stage III and 4 sets of data that describe populations: age, gender, education, number of stages executed. For both diseases graphs and trees are built under assumption that factors are vertices and significant correlations are edges, leading to model of dependencies between factors. Usefulness of this approach to analysis of difference between diseases is discussed.