Comparison of alternative sources of data on health service encounters.

OBJECTIVES: Due to the paucity and incompleteness of routine databases, many areas of health services research rely on the accuracy of individuals' recollections of health service encounters. The objective of this study is to explore differences between patient and physician recall of health care resource use by employing data from two randomized controlled trials.
METHODS: Data on the frequency of use of community-based health services were reported by both patients and their general practitioners (GPs) at three-month intervals for the first year after randomization. Analysis compared the completeness of data collected, the level of agreement between GPs and patients and differences in the estimated cost of community services. Further analysis assessed the impact of patient characteristics on reporting.
RESULTS: Data provided by GPs were less likely to be complete than patient data. There were significant differences between GPs and patients in their reported use of certain community services, particularly in relation to GP-patient contacts. However, this did not lead to significant differences in estimates of costs. Patient characteristics influenced the level of agreement with GPs, but not the proportion of forms completed.
CONCLUSIONS: In research in which the use of community resources is a major component of costs, differential estimates of resource use may influence study results. Further research is required to identify optimal data collection methods for health service encounters. Where possible, studies should incorporate estimates of resource use from a variety of sources and conduct sensitivity analyses to assess the robustness of the results.