OBJECTIVE: To develop and validate the design of a grid that assesses doctor-patient relationship skills. DESIGN: Evaluation study of an assessment instrument. SETTING: Private practices and family practice units. PARTICIPANTS: From a sample of volunteers, 100 family physicians either in private practice or in a family practice unit completed the proposed grid independently. MAIN OUTCOME MEASURES: The Cronbach alpha coefficient was used to analyze internal consistency. Factorial analysis was used to determine whether the grid's anticipated dimensions were in fact present. RESULTS: The Cronbach alpha coefficient had a very high value (0.92), indicating that the items in the grid were highly homogeneous. Two key factors emerged from the factorial analysis; the first factor alone (understanding patients' experience) accounted for almost 42% of the variance. CONCLUSION: The proposed grid presents some interesting metrologic qualities. It is short and relatively simple to use to assess relationship skills of future and practising family physicians. The grid must now be further validated using a variety of cases.