Models and algorithms for contaminated area detection based on geospatial lifelines

Abstract

Environmental factors are considered to be one of the elements responsible for the development of certain diseases. Examples of these environmental factors include deficiency of some elements (e.g., certain vitamins) that are necessary for maintaining a person's health or environmental contamination of an area by hazardous chemicals. For people living in a contaminated area, they are prone to get sick. If residential places are fixed for all people, then we can easily identify contaminated areas by locating clusters of patients with similar diseases. Unfortunately, this is not the case especially in modern society since people change their residential places frequently. In this paper we use patients' residential history (also called geospatial lifeline) to locate contaminated areas. In various domains, such as epidemiology and public health research, detection of space-time clusters is an important task. Current cluster analysis methods can only identify general hot spot in a given time period but cannot pinpoint the exact area where environmental factors may be responsible for the development of a disease. We propose a novel method to identify possible relationships between a disease and the locations where environmental factors might be responsible for the development of a disease. This method differs from previous methods in two ways. Firstly, we adopt the concept of geospatial lifeline which is actually a piecewise linear trajectory in three dimensional space (x, y dimensions plus time dimension). Secondly, based on disease principles, we divide a patient's geospatial lifeline into four periods: normal period, the period of being exposed to a contaminated area (exposure period), latent period, and sick period. Therefore, a geospatial lifeline is not only a spatial-temporal trajectory but also has useful semantic information in different parts of the trajectory. Based on patients' geospatial lifelines, this new method helps unearth unknown contaminated areas responsible for the development of a given disease and disclose other useful disease related information.