Objective:To assess the use of patient‐held medication record cards and their acceptability to patients and doctors.Design:Prospective 12‐month study with data collection at baseline and on three subsequent occasions at four‐monthly intervals.Patients and setting: 187 patients with a mean age of 78.4 years (range, 60–101) were taking a mean of 5.8 medications each (range, 1–18). They lived on Sydney's lower north shore and were able to care for themselves.Main outcome measures: Availability of card on request, frequency of use, status of recorders and accuracy of records (checked by inspection of medications at home).Results:Most patients retained their cards, but the proportion who presented it to their doctor fell from 61% to 23% over the 12 months (P<0.0001), and the proportion with accurately recorded drug regimens ranged from 20% down to 16%. Of the 75 regimens written exclusively by general practitioners in the 12 months, only 19 (25%) were consistent with what the patients were actually taking.Conclusion:Medication record cards introduced into the doctor‐patient relationship by a “third‐party” are unlikely to result in better quality use of medicines.