Infant Mortality

GIS-based methods for reducing Infant Mortality Rates

The African American infant mortality rate in East Baton Rouge Parish, Louisiana, is relatively stable at 14 per thousand. This rate is approximately three to four times higher than the white infant mortality rate. However, when the geographic area is changed, to that of the program area for Baton Rouge Healthy Start, this rate rises above 15 per thousand, and even hits consistent highs of between 40 to 70 per thousand for the worst neighborhoods. GIS has been used to identify the program area for the Healthy Start program, and identify those neighborhoods most “at-risk”. The Baton Rouge Healthy Start program designed the Eliminating Disparities in Perinatal Health program, which currently serves approximately 300 women from indigent areas within the city. These women receive a full range of prenatal and postnatal care from a qualified staff of nurses and social workers. Program caseworkers benefit from continuing GIS analysis that has revealed hotpots of poor prenatal care, low birth weight deliveries, and tobacco and alcohol use during pregnancy. The most exciting aspect of the Baton Rouge Healthy Start program is that a GIS sits at the heart of the data collection and analysis. All women entering the program are initially evaluated by a caseworker. During the three years of their involvement with the program, other data is collected on a regular basis

These data, uploaded by the caseworkers into a secure GIS, allow for standard program reports to be generated, as well as site-specific investigations. Data is entered into a specially constructed database with drop-down menus, hierarchical data entry (for example once a risk is identified it leads to further data collection fields), and automatic referral generation. Additional data collected include detailed individual, social and neighborhood information that are suitable for GIS analysis. For example, spatial analyses have been performed on neighborhood infrastructure (the quality of housing), neighborhood perception (the fear of crime), and individual stresses (mental health screening). In this way a holistic approach is being developed to understand the problems that coincide to contribute to a poor birth outcome. In addition, specific spatial questions are captured, such as how mobile the program participants are (do they move during pregnancy?). Apart from hotspot analyses, the LSU research team has also used this collaboration to publish on preserving confidential information within a GIS environment, and in general produce GIS guidelines for similar health programs across the United States.

Bus stop shelter advertisement boards are another way of communicating the program around the city. Shown here is Dr. Andrew Curtis (Lab director and Project PI) in front of a bus stop billboard in Baton Rouge.

Analytical outputs from IMR GIS

The GIS can be used to map a variety of parameters to investigate the spatial patterns of infant mortality. One important factor in minimizing risk is to identify areas where little or no prenatal care are being received. This can then direct future health worker visits to ensure that areas of the city lacking services will be targeted.

Kernel density analysis provides a first look at the distribution of hot spots for IMR throughout Baton Rouge. Here a three year period is mapped to show those regions which remain consistently high through time and the dynamic characteristics of IMR.

One of the major goals of the IMR project was to develop a near real-time data entry system so that results from public health worker surveys and interventions could be accessible to both GIS personnel and other project team members. An interactive Access® database was developed to facilitate data entry and management. Priorities for data forms were placed on both the needs of health care workers and the demands for spatial data for analysis in the GIS environment. The database is entirely self contained and health workers need only understand the front end GUI forms for entering data. Future database enhancements may include secure online data entry to take advantage of near real-time surveillance techniques.

(Click HERE to return to ongoing research)