Doctors face a fundamental hurdle when seeking to reassure patients with medically-unexplained symptoms that they are physically well. Apparently such patients have a problem remembering information about how likely it is that they or other people have certain illnesses.
Winfried Rief and colleagues played 33 patients with medically-unexplained symptoms a tape-recording of a doctor’s report on a patient with abdominal pain. In his report, the doctor rules out certain illnesses – for example, that the pain is “certainly not stomach flu”. Yet later, when asked to indicate on a scale from 0 to 100 how likely the doctor said these illnesses were, the participants with medically-unexplained symptoms said the possibility was 20 per cent. By contrast, a group of depressed patients and healthy participants rated the possibility as significantly lower – as 13 and 3 per cent, respectively.
This memory bias shown by participants with medically-unexplained symptoms was specific to medical probabilities. They could hold as many numbers in short-term memory as the depressed and healthy participants. And when played an audio-tape about a case of social rejection (a person not being invited to a neighbourhood barbeque), or about a car breakdown, they remembered probabilistic information contained in these reports just as accurately as the depressed and healthy participants.
“Our results show that medical reassurance and the presentation of negative test results can lead to patients remembering overestimated probabilities for medical explanations, especially in patients with unclear somatic complaints”, the researchers said. “Check-back questions on what patients have understood from doctors’ reports, and asking patients for summaries about the provided information, could help detect this memory bias…”, they advised.
Rief, W., Heitmuller, A.M., Reisberg, K. & Ruddel, H. (2006). Why reassurance fails in patients with unexplained symptoms – An experimental investigation of remembered probabilities. PLOS Medicine, 3, e269. Open access.