From a societal perspective long-term clinical trials or follow-up studies should preferably not only include an evaluation of the health effect for the patient, but also an economic evaluation. In order to yield comprehensive medical and nonmedical resource use data, we at least partly depend on respondents' recall for collecting these costing data. A patient cost diary was developed in order to estimate total resource use, expenses, and lost production due to illness and treatment. We applied the cost diary in two randomized clinical trials evaluating the cost-effectiveness of behavioral rehabilitation in 205 fibromyalgia and chronic low back pain patients. The use of the diary was evaluated, studying the feasibility, the influence of the period of data collection on the results, and some aspects of validity. Eighty-five percent of the patients completed at least one diary and in total 68% of the diaries were returned. Although the results for the three alternative periods of data collection (keeping the diary 1 week every month, 2 weeks every 2 months, or a full year) were not significantly different, they were only moderately correlated. Finally, self-reported specialist care contacts were generally in agreement with data from an insurance company. However, for physiotherapy contacts there were differences between the self-reported and insurance data. This study shows how the cost diary might be used successfully in cost-effectiveness studies.