Background: In many diseases men and women, for no apparent medical reason, are not offered the same investigations and treatment in health care. This may be due to staff's stereotypical preconceptions about men and women, i.e., gender bias. In the clinical situation it is difficult to know whether gender differences in management reflect physicians' gender bias or male and female patients' different needs or different ways of expressing their needs. To shed some light on these possibilities this study investigated to what extent it was possible to identify patients' sex when reading their blinded illness narratives, i.e., do male and female patients express themselves differently enough to be recognised as men and women without being categorised on beforehand?
Methods: Eighty-one authentic letters about being diseased by cancer were blinded regarding sex and read by 130 students of medicine and psychology. For each letter the participants were asked to give the author's sex and to explain their choice. The success rates were analysed statistically. To illuminate the participants' reasoning the explanations of four letters were analysed qualitatively.
Results: The patient's sex was correctly identified in 62% of the cases, with significantly higher rates in male narratives. There were no differences between male and female participants. In the qualitative analysis the choice of a male writer was explained by: a short letter; formal language; a focus on facts and a lack of emotions. In contrast the reasons for the choice of a woman were: a long letter; vivid language; mention of emotions and interpersonal relationships. Furthermore, the same expressions were interpreted differently depending on whether the participant believed the writer to be male or female.
Conclusion: It was possible to detect gender differences in the blinded illness narratives. The students' explanations for their choice of sex agreed with common gender stereotypes implying that such stereotypes correspond, at least on a group level, to differences in male and female patients' illness descriptions. However, it was also obvious that preconceptions about gender obstructed and biased the interpretations, a finding with implications for the understanding of gender bias in clinical practice.