Indirect standardization was used to produce estimates of the prevalence of selected risk factors and health status indicators for all health units in Ontario, based on their socio-demographic characteristics and on the relationships between these socio-demographic variables and the health variables of interest, as measured by the Ontario Health Survey (OHS). Socio-demographic variables were drawn from both the OHS and the 1986 Census of Canada. In addition, the validity of the estimates was assessed. A total of 13 combinations of predictor variables were used to estimate 18 selected health status indicators and risk factors. Results showed that the choice of predictor variables used for estimation can have a substantial impact on the error associated with a set of estimates. As well, it was found that the dispersion of the estimates was reduced in comparison to the true values. The ranking of the estimates was found to vary from poor to good. In addition, estimates were shown to be better than using the provincial value as the estimate for each health unit value. Comparison by region showed that estimates for northern regions had somewhat more error than estimates for other regions and that addition of a geographic identifier reduced the amount of error among the estimates. Given all these results, it was concluded that this type of methodology does not appear to be adequate to meet the information needs for health units but that other methods such as regression might produce better results.
| Identifer | oai:union.ndltd.org:uottawa.ca/oai:ruor.uottawa.ca:10393/9684 |
| Date | January 1994 |
| Creators | Strike, Carol Janice. |
| Contributors | Spasoff, Robert A. |
| Publisher | University of Ottawa (Canada) |
| Source Sets | Université d’Ottawa |
| Detected Language | English |
| Type | Thesis |
| Format | 99 p. |
Page generated in 0.0134 seconds