Using probability and virtual patients to aid in diagnostic reasoning

Sian Claire Owen

One of the problems facing medical students is gaining access to a wide variety of patients within a patient population. Conditions such as diabetes mellitus have a wide spectrum of symptoms, and in order to develop their diagnostic skills the student needs experience of dealing with as many varied examples as possible. Dr Jeroen Donkers, Assistant Professor in e-learning at Maastricht University is currently applying artificial intelligence principles to virtual patients in order to help students develop their diagnostic reasoning skills.

As Jeroen explains: “The problem with doctors is that when they do diagnostic reasoning they have to deal with a lot of uncertainties. They see the symptom, but what is the probability that the diagnosis belongs to that symptom? It’s an important ability that students have to develop.”

However, as probabilities depend largely on the patient population, limited exposure to a wide range of patient cases within that population can hamper diagnostic skills.

“For instance” says Jeroen, “If you see a lot of patients who have a headache caused by drinking too much, the next time you see a patient with a headache you’ll think that he drinks too much. On the other hand, if you only see patients with headaches because of brain tumours, the next time you see one [patient with headache] you immediately think of brain tumour.”

To address this issue, Jeroen and colleagues developed a system where information about a given disease state – in this case, diabetes mellitus – is encoded into a probabilistic network. This includes adding demographic and epidemiological data that produces a large number of brief virtual patients, giving an accurate picture of that patient population.

Students were exposed to a set of approximately 100 virtual patients that represented the diabetic population. They chose diabetes because medical students at Maastricht University are unlikely to treat these patients on their internships as the patients are looked after by specialist nurses.

“We hope to find differences between the students who had treatment with the specialized model with those students who hadn’t,” he says. “If this works, it will be a great advantage,” he adds. “Because we will be able to present students with a population that will enhance their abilities to reason with probabilities.”

Jeroen and colleagues will present their findings at the International Conference on Virtual Patients 2010 and the MedBiquitous Annual Meeting.

Listen to Jeroen talk about his research here:

Comments are closed.