Approaches presented herein enable treating a medical condition via individual customized cognitive guidance. More specifically, a current interactive and customized diagnosis object (ICDO) comprising a treatment schedule for a patient is generated by combining a set of previous ICDOs from the patient, wherein each ICDO has a structure which defines elements that are used to diagnose a symptom and match the patient with a treatment. A status of the patient is monitored in real time via an internet of things network to detect an actual event of the patient differing from the treatment schedule. In real time, the treatment schedule is modified responsive to the differing event by adjusting the current ICDO by exchanging one element of the treatment plan for a new element according to a learned set of rules to compensate for the differing event. This modified treatment schedule is provided to the patient in real time.