| Summary: | Thesis (M.Sc.)--Memorial University of Newfoundland, 1992. Medicine Bibliography: leaves 138-141. The objective of this thesis was to study the validity of self-reported medical care utilization. Hospitalization and physician visit data for a twelve month recall period were obtained from both an interview and official records. The self-reported information was collected by a telephone survey applied to all adults over 20 years of age in a probabilistic sample of households in metropolitan St. John's (3,300 subjects, 85% response rate). Verification data were later obtained for 2,994 subjects (91%) from the provincial hospitalization database and health insurance plan. -- The utilization data were used to categorize the subjects into those in agreement, the underreporters and the overreporters. A variable denoting level of accuracy was derived. Socio-demographic variables (sex, age and education) and health status variables (self-assessed health status, number of chronic conditions, satisfaction with physical health, and emotional status) were used in a descriptive analysis to compare those in agreement, the under- and overreporters. -- Logistic regression was utilized to investigate the probabilities of being in disagreement on utilization and to compare those who were accurate in self-recall with those who were not. -- The analyses showed that most subjects (97.3%) were in agreement on hospitalizations in the recall period whereas 84.1% agreed on physician visits. The observed rates of overreporting were 16.2% for hospitalizations and 9.7% for physicians, and for underreporting, 7.3% for hospitalizations and 10.0% for physicians. -- For hospitalizations, 1) subjects in disagreement had, usually, less education and more chronic conditions than those in agreement, 2) underreporters tended to be older than either overreporters or those in agreement, and 3) the less accurate at reporting were generally more likely to report several chronic conditions, fewer years of education, lower emotional and self-assessed health status and were older. -- For physician visits, 1) subjects found to be in disagreement were in general male, had better emotional and self-assessed health status, and reported fewer chronic conditions, 2) underreporters tended towards higher self-assessed health status, lower education and fewer chronic conditions than those in agreement, and 3) those more prone to be inaccurate in reporting the number of visits were generally female, had lower self-assessed health status, were less educated and had more chronic conditions.
|
|---|