Systems and methods for predicting a person's likelihood of a situational mental health condition (such as post-partum depression (PPD)) diagnosis using machine learning models are described. In one example, a mental health risk assessment system receives data at recurring intervals for a person during their pregnancy and in the year after their pregnancy. Based on both the recurring data for the person as well as the specific time period at which the data is obtained, a machine learning model can generate highly accurate mental health predictions. The mental health risk assessment system can further implement a software application that manages the flow of information from patients as well as presentation of information to the patients. Furthermore, the application can present targeted information for a patient to the patient's doctor or other medical personnel/facility.