Wstępne wyniki zewnętrznej weryfikacji odpwoiedzi respondentów w wywiadzie na temat zachowań zdrowotnych
MetadataShow full item record
The purpose of the study was to check the answers of respondents against the records of a district dispensary of general medical services as well as to analyse the influence of certain traits of respondents (sex, age, occupation and education) upon the accuracy of their answers. The data obtained from interviews conducted in the WHO International Collaborative Study of Medical Care Utilization were used. The checking related only to medical visits which occurred in three time periods: (1) not longer than two weeks before the date of interviewing, (2) from two weeks to one year, (3) more than one year. Data obtained from direct interviews with 524 adults and 111 indirect interviews on children (for whom their parents or foster parents were answering) were used. Two kinds of accordance (“positive” and “negative”), two kinds of discrepancies (overstating, understating) and so called inaccurate answers were distinguished in the analysis. Discrepancies between the answers of respondents and the medical records turned out to be pretty large; the range of differences was, depending on the category of respondents and the period covered, from 8.8 to 47.1 percentage points. Positively less answers out of line with reality (i.e. discordant with medical records) were found when the question related to a short time period (i.e. “two weeks ago”); much more when longer time periods were covered. The range of the gross errors, depending on the time period covered, was from 7.0 to 28.8 percentage points for direct interviews, and from 6.3 to 47.1 for indirect ones. Discrepancies between the data obtained in interviews and the records of medical dispensary are much greater in case of indirect interviews, in which a tendency toward "incomplete” answers (i.e. understating discrepancy) is discernible; a contrary tendency appeared in direct interviews, in which half of the respondents were giving false answers overstating the actual number of visits. Most of the false answers were given by older respondents with elementary – or less than elementary – education level. In these persons there prevailed a tendency to give “overstating” answers. Most of discrepancies of all kinds consisted in the wrong location in time of a visit that actually took place earlier. Time discrepancies ranged from two to six days. It follows from the analysis that least truthful are answers relating to past time periods longer than two weeks before the date of an interview as well as the data obtained from indirect interviews.