224 Hospital registrations and health survey data – do they agree?

BackgroundIn many countries health interview survey data are used for indicators for injury incidence. However, the validity of self-reported injury incidence may be questioned due to e.g. recall bias and low response rate in groups at high injury risk. In the first European Health Interview Survey the incidence of home and leisure injuries varied as much as from 1.3% to 8.2%. The purpose of the present study is to compare survey response and hospital registration at the individual level with focus on reporting bias.MethodsThis study was carried out using the Danish health interview survey data with information on injury the past year and the treatment. These data were linked at the individual level to the hospital registration of both in- and outpatient data for the period up to two year before the interview, for all hospitals in Denmark.ResultsIn total 368 reported injuries being hospital treated as outpatient, of these 234 were actually hospital treated within the last 12 months (64%). Ninety-six reported being admitted to hospital, of these 51 were actually admitted the past year (53%) and 59 the past two years (61%). Conversely, only about half of the hospital treated injuries were reported in the survey.ConclusionsThere is considerable disagreement at the individual level between self-reported hospital treated injuries and actual hospital treatment. Hospital admissions in particular seem to be over reported.