Aim: To characterize the factors associated with the 'doctor' effect in primary care, as reported in randomized controlled trials (RCTs).
Method: A systematic search of Medline (1964-2004) sought to identify all original reports of RCTs, as well as those reported in reviews and metaanalyses. We used the following key words: RCT, doctor-patient relationship, doctor-patient communication, knowledge, skill, attitude, non-pharmacologic effectiveness, primary care.
Results: Ten RCTs and one metaanalysis provided evidence that a combination of emotional and cognitive care has a consistently positive effect on health outcomes. This effect relies on specific attitudes and skills: empathy, reassurance, explanation, counseling, influencing patients' 'health beliefs and expectations, promoting change in behavior, thoughts or emotions. These can be integrated into a specific patient-centered approach to general practice. Medical education must provide training in the relational skills needed for effective treatment.
Conclusion: Qualitative and quantitative research, including RCTs, are necessary; they should be designed to deal with the heterogenous situations and specific characteristics of general practice.