Change-Point Detection Using Directional Derivatives

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.

Allan B. Clark, A. Lawson
University of South Carolina School of Public Health, Columbia, SC

Corresponding author: Andrew B. Lawson, Department of Epidemiology and Biostatistics, Norman J. Arnold School of Public Health, 800 Sumter St., Columbia, SC 29208. Telephone: 803-777-6647; Fax: 803-777-2524; E-mail: alawson@gwm.sc.edu.

Abstract

Introduction: Individual-level disease maps, which estimate the risk for disease across a geographic region, usually are based on observing a set of spatial locations of cases and controls. This study examined the extension of where cases and controls form a space-time point process (locations and dates) and focused on assessing whether the intensity of cases changed across time.

Objectives: The objectives of the study were to develop a method for detecting changes in individual-level disease maps. The method was applied to a data set of birth abnormalities in the United Kingdom, which included locations and times of all the live (singleton) births and abnormalities during a 5-year period.

Methods: The change in intensity across time was measured through the directional derivative, with respect to time, of the geotemporal surface. This can be estimated nonparametrically and is used to check for both sudden and gradual changes over time.

Results: The results do not demonstrate the descriptive ability of an approach that relies on a map of changes in risk. The directional derivative was computed at 10 update points, corresponding to when data were available, and summary statistics were produced (Table). Isolated departures from constant risk were indicated, but the measure aggregated over a map did not demonstrate any change.

Conclusions: A directional derivative approach might yield optimal answers and is worthy of further research. The example data (Table) demonstrate that isolated changes are occurring, but data aggregated over a map did not indicate any change. Therefore, the geography of the problem should be considered, but an analysis that is aggregated over geography should not be performed.

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.