This is a position paper for illustrating the potential of mediating diagnosis and treatments of mental health disorders by promoting computational models, which support clinical decision-making in psychiatry. The proposal juxtaposes mental health symptoms, deviations in brain biology, social and personal environment of an individual, and current attempts to alleviate individual’s mental health problems, including treatment which may have already existed, and mediates a diagnosis. Mediation would be prone to changes and diagnosis can be delayed and changed, if the environment and factors affect the mediation change. It does not mean that diagnosis would never be firmly decided, but we leave the clinical decision-making open and constantly evaluate diagnosis through the mediation. The proposal uses first-order logic for reasoning about the context in which we perform the mediation. Generative AI tools may add to the semantic richness of the data and relationships between them and possibly enhance the computational part of the proposal. The purpose of the paper is twofold. It could address current problems of diagnosing mental health patients and might move computational psychiatry toward software solutions which do not solely depend on predicting algorithms.