Persons using assistive technology might not be able to fully access information in this file. For assistance, please send e-mail to: mmwrq@cdc.gov. Type 508 Accommodation and the title of the report in the subject line of e-mail.
Empirical Evaluation of Space-Time Models
for Surveillance of Disease Maps
Carmen L. Vidal Rodeiro, A. Lawson University of South Carolina, Columbia, South Carolina
Corresponding author: Carmen L. Vidal Rodeiro, Department of Epidemiology and Biostatistics, Norman J. Arnold School of Public Health, 800 Sumter St., Columbia, SC 29208. Telephone: 803-777-1232; Fax: 803-777-2524; E-mail:
cvidal@sc.edu.
Abstract
Introduction: Interest in statistical methods for public health surveillance has increased in recent years.
Objectives: Different space-time models for counts of disease were compared to assess their ability to detect changes in risk patterns across space and time.
Methods: Space-time models for estimating disease risk should be able to describe the overall space-time behavior of a disease and should also be sensitive to changes in its spatio-temporal structure. For this study, the observed count of disease cases in a region was assumed to be a Poisson variable. Logarithms of relative risk parameters
were assumed to follow normal distributions with mean that incorporated potential risk factors and variance matrix
that incorporated the possibility of spatial dependence (e.g., correlation induced by unmeasured variables).
Space-time models in different scenarios representing possible changes in risk patterns over space and time were fitted.
Results: As a goodness of fit measure, the deviance information criterion was used. It demonstrated
statistically significant increases in the years in which changes in risk were generated. Analysis of the p-value surface, residuals, and surveillance residuals (difference between observed data for 1 year and data expected under a model when fitted for previous years) proved that an unusual event happened in the counties and years with changes; therefore,
those
data were not representative of what was expected under the model. Where no changes in risk were generated, the p-values indicated that the model produced an optimal fit.
Conclusions: Although existing methods can be used for disease surveillance, additional methods that are more sensitive to the sequential nature of the surveillance task are needed.
Use of trade names and commercial sources is for identification only and does not imply endorsement by the U.S. Department of
Health and Human Services.References to non-CDC sites on the Internet are
provided as a service to MMWR readers and do not constitute or imply
endorsement of these organizations or their programs by CDC or the U.S.
Department of Health and Human Services. CDC is not responsible for the content
of pages found at these sites. URL addresses listed in MMWR were current as of
the date of publication.
Disclaimer
All MMWR HTML versions of articles are electronic conversions from ASCII text
into HTML. This conversion may have resulted in character translation or format errors in the HTML version.
Users should not rely on this HTML document, but are referred to the electronic PDF version and/or
the original MMWR paper copy for the official text, figures, and tables.
An original paper copy of this issue can be obtained from the Superintendent of Documents,
U.S. Government Printing Office (GPO), Washington, DC 20402-9371; telephone: (202) 512-1800.
Contact GPO for current prices.
**Questions or messages regarding errors in formatting should be addressed to
mmwrq@cdc.gov.
