Related Experiment Videos
Methods for confidence interval estimation of a ratio parameter with application to location quotients
Joseph Beyene1, Rahim Moineddin
1Department of Public Health Science, University of Toronto, Toronto, Ontario, Canada. joseph@utstat.toronto.edu
BMC Medical Research Methodology
|October 14, 2005
Summary
This study introduces methods to calculate confidence intervals for the location quotient (LQ), a key measure for comparing health outcomes across regions. Reporting LQ with confidence intervals enhances its value for policy decisions.
Area of Science:
- Health outcomes analysis
- Geographic health disparities
- Spatial epidemiology
Background:
- Location quotient (LQ) quantifies activity concentration across areas.
- LQ is applicable to population health for comparing outcomes spatially.
- A limitation of LQ is its use as a point estimate without confidence intervals.
Purpose of the Study:
- To present statistical methods for constructing confidence intervals for location quotients.
- To address the limitation of point estimates for LQ in health outcome comparisons.
Main Methods:
- Utilized delta and Fieller's methods for ratio parameters.
- Employed generalized linear modeling for profile-likelihood confidence intervals.
- Conducted simulation experiments and used health utilization data for illustration.
Main Results:
- Different analytical methods yielded similar confidence limits for LQ.
- Confidence limits were nearly indistinguishable with large sample sizes and non-rare outcomes.
- Generalized linear model approach may be preferable for small sample sizes.
Conclusions:
- LQ is a valuable, simple measure for quantifying and comparing health outcomes across regions.
- Reporting LQ with confidence intervals increases its utility for policymakers.
- The presented methods enhance the interpretability and applicability of LQ.