Abstract. Employing the theory of belief change, we study the implementation of a simple causal model that can capture how the blood sugar level changes in a diabetes patient. For this purpose we use distance measures between worlds as the underlying mathematical foundation. Using a simple example in the medical domain we investigate how an agent with initially incomplete and/or incorrect knowledge can iteratively develop a simple causal model by interacting with an oracle that represents the complete and correct model of a diabetic patient.