synergystic factors



If you had a mass of data, is there any way of extracting/identifying factors which only produce a desired result if they exist together in that particular instance. eg. you may have a lot of data on cancer patients... and hidden in that data may be two medicines which have proven to have had no real effect on recovery.. that is some patients have had one medicine... others have had the other medicine.. but hidden in that data may be a few situations where patients were given both medicines and this combination produced 100 percent recoveries. Such a situation would not involve a hypothesis... it would be more like data mining... but do statistical programs exist that can identify such nuggets of gold?