Valerie Pavilonis

When it comes to geography, traditional sources of human information have limits. Census information is useful for static measures like race and ethnicity, but researchers have found it difficult to associate places with disease outcomes.

At Yale, research teams have taken advantage of modern technologies to better understand how people interact with their locations. In Brazil, a recent public health study used GPS to determine the movement patterns of individuals in urban slums and to assess the resulting health risks. Their study was published in PLOS Neglected Tropical Diseases on Aug. 31.

Led by Katharine Owers SPH ’18, the team used individual GPS devices to better document residents’ exposures to risk factors of leptospirosis, an infectious disease common in Latin America. It took place in the Pau da Lima urban slum of Salvador, Brazil, and used the location devices to determine the movements of residents in high-risk environments.

“We think GPS is a useful tool for epidemiology that demonstrated its potential nicely in our study,” Owers said.

Over the course of four months, Owers and her team distributed GPS devices to over 100 residents in a leptospirosis-rich area of Pau da Lima. The research group proceeded to track the individuals’ activities over 24 hours. Afterwards, their movements were analyzed to see how often participants came into contact with leptospirosis-prone markers like sewers and trash deposits.

Four of the study’s participants had confirmed cases of leptospirosis infections. Their GPS devices revealed that they had not only come into more contact with areas associated with leptospirosis than the average resident but also had interacted in a larger environment than most of the community. The small number of infection cases limited conclusions made by Owers’ team, but they were still able to recognize a possible association between frequent activity and high rates of leptospirosis.

“Katie’s study raises interesting concepts about urban slum health and emerging diseases,” said senior author and School of Public Health professor Albert Ko.

The study noted that previous inquiries had identified leptospirosis risk factors based on where residents live. Since slums are generally more clustered environments, activities among residents were often hard to measure precisely.

Through the use of positioning devices, however, the research team saw specific movements that increased the likelihood of contracting leptospirosis. Owers’ team was able to identify patterns of disperse wanderings, which weren’t visible with traditional forms of data collection, as potential disease indicators that the team believes can guide effective health interventions.

Other research areas have used GPS technology to describe health-related movement patterns. Stephen Matthews, an anthropology professor at Penn State University, specializes in geographical studies and underscored positioning devices’ value as data collectors in his research.

“GPS traces or footprints, following an individual across a day — or extended sequence of days — reveals that for most of us, we move around and beyond our residential neighborhood,” Matthews said. “Some trips are regular and structured activities — e.g. commutes, shopping trips, school, church — but even among structured journeys the temporal frequency varies across days and weeks.”

While GPS technology has apparent value in public health research, some studies have found potential drawbacks to its use. Basile Chaix, an epidemiology professor at Sorbonne University, linked the use of positioning systems to a phenomenon called the selective daily mobility bias. This problem can occur when the locations to which a person travels have some bearing on their health based on voluntary action — the gym, for example.

According to Chaix, the bias generally does not apply to studies tracking exposures to disease. However, the problem can appear when using GPS devices to study behavior.

Matthews also warned of the limitations of a GPS study. He said its information should be paired with traditional data collection “to gather data on individual health and demographic characteristics.”

There are 1.03 million cases of leptospirosis worldwide, according to a 2015 study based on World Health Organization estimates.

Allen Siegler |