Are survey measures of medical care utilisation misleading? A comparison of self-reported medical care consumption with actual medical care utilisation.

A substantial over-estimate of medical care consumption cost was found when estimates from self-report data from an epidemiological study were compared to actual cost data extracted from administrative records. Even though the few subjects who were actually provided with two or more services in the two-week self-report period substantially under-reported their medical care consumption, a large net over-estimate of medical care consumption was produced by the self-report data. This finding has important implications for use of self-report data from surveys such as the Australian Bureau of Statistics (ABS) National Health Survey for estimating health service consumption. By combining epidemiological survey data from the Australian Vietnam Veterans Health Study (AVVHS), with data on actual medical care for which the Health Insurance Commission (HIC) or the Department of Veterans' Affairs (DVA) paid benefits, we were able to directly compare self-reported medical care consumption with actual medical care utilisation. The comparison revealed that veterans' self-reports were a valid measure of relative medical care consumption because those who reported care over the past two weeks were much more likely to have been recent consumers than those who did not. This relationship became even stronger if the comparison of self-report was extended to data on benefits paid beyond the two-week self-report period. However, the HIC and DVA data confirmed only 51% of veterans self-reporting medical care consumption during the past two weeks actually received a service.