Their paper, published yesterday in The Archives of Internal Medicine, involved 100 primary-care doctors in the Rochester area. As part of a study on patient care and outcomes, the doctors agreed to allow two people trained to act as patients come to their offices sometime over the course of a year. The test patients would surreptitiously make an audio recording of the encounter. The investigators analyzed recordings of 113 of those office visits, excluding situations when the doctors figured out that the patient was fake.
To their surprise, the researchers discovered that doctors talked about themselves in a third of the audio recordings and that there was no evidence that any of the doctors' disclosures about themselves helped patients or established rapport.
Nor, in the vast majority of cases, did the doctors circle back to the personal conversation or try to build upon it.