Table 1.

Hypothetical studies of patients, presenting the possibilities of using machine learning methods

Hypothetical patient	Health context	Actual behaviour of the patient	Input data “seen” by the application	Possible actions of the application
1.	The patient has been diagnosed with social phobia. She is undergoing pharmacotherapy and psychotherapy.	The patient and her partner left their place of residence and went to a public space. While travelling on the subway, she witnessed a situation in which her partner was ridiculed for her appearance. The patient’s heart rate increased significantly, she might began to think about self-harm. After returning home, she searched the Internet for information related to suicide.	- Changing the location from “Home” to “Other”. - Registration of a significantly higher heart rate than the average for a given user, with no registration of physical activity. - Registration of searching for terms marked as “dangerous”.	Sending a notification to the user about the possibility of obtaining psychiatric and psychological assistance. In a wider variant, after prior consent of the patient – sending an alert to the person conducting the treatment.
2.	The patient was diagnosed with a depressive disorder. She is undergoing pharmacotherapy and psychotherapy.	For the last two weeks, the patient led a relatively active lifestyle, and the disease did not significantly affect her functioning. Recently, her sister has been feeling worse and has more health problems, which is why the patient’s mood is clearly depressed. The patient leaves home less often, searches for new information and uses a mobile phone less often. She spends most of the day in bed. She doesn’t listen to music, does not browse the Internet, and if she does, she looks for information about death.	- Changing the location from “Home” to “Other” is made much less frequently than usual. - Physical activity is much less frequent than usual. - Registration of searches for terms marked as “dangerous”.	Sending a notification to the user with a suggestion to connect to the person conducting their treatment.