Implementing an animated geographic information system to investigate factors associated with nosocomial infections: a novel approach
Kho, A, Johnston, KG, Wilson JS, & Wilson, SJ. Implementing an animated geographic information system to investigate factors associated with nosocomial infections: A novel approach. American Journal of Infection Control; 34(9):578–582.
Question
Are hospital outbreaks of transmitted infections caused by direct human contact?
Introduction
Abel Kho, an Indianapolis MD and researcher studied the relationship between methicillin-resistant Staphylococcus aureus (MRSA) and the bacterium’s transmission to hospital patients. Kho hypothesized that MRSA infection was transmitted through environmental factors such as direct human contact (e.g., from the hands of a health-care provider to a patient).
With the assistance of geography-graduate student, Kelly Johnston, they used a geographic information system (GIS) to model both the spatial and temporal movements of patients and staff at the Wishard Memorial Hospital in Indianapolis.
A GIS can analyze and model spatial and associated attribute information (e.g., a nurse’s location in a hospital relative to the location of infected patients). The output of a GIS can be maps, tabular reports, and/or graphs, all of which are designed to reveal relationships, patterns and trends from the model inputs.
Methods
For Kho’s study, model inputs for the GIS included the following:
- microbiology and laboratory data (used to identify infected patients)
- patient admission/discharge/transfer (ADT) data
- login data, as inputted by hospital staff, into bedside, automated vital signs monitoring equipment (this information located a health-care provider relative to an infected patient)
Data were collected over a three-month period and included 44,485 logins. A primary assumption of the study was that the average time for a health-care provider to check a patient’s vital signs and wash hands before moving on to the next patient and “vital signs” login was at least three minutes. Otherwise, if login times between patient visits occurred faster than three minutes, it was assumed that the health-care provider was not sanitizing his or her hands.
Results
By modeling both changes in space and time, the GIS found a high correlation between staff-patient contact patterns and the transmission of MRSA. The output from the GIS visually demonstrated that, over a three-month period, 6,248 logins occurred in less than three minutes per patient visit, and that 5,585 (89%) of those logins were made by nursing aides.
The GIS model also revealed an additional correlation of increased MRSA transmission where 8 out of 24 MRSA-infected patients shared a room with other patients.
Conclusion
The use of GIS was successful in modeling the transmission of MRSA infection between health-care providers and patients. Interpretation of the GIS’s output showed a high correlation between infection transmission and direct human contact.
The results from this study initiated changes to the Wishard Hospital’s hand-sanitation procedures. By educating hospital staff on proper hand hygiene and instituting a “computerized reminder” for physicians to isolate patients with MRSA, the Wilshard Hospital has seen, over a two-year period, a nearly 50% decrease in patients infected with MRSA.
Commentary
The study by Kho is a good example of the many success stories of GIS’s role as an analytical tool in modeling epidemiological data in order to reveal relationships, patterns, and trends between diseases and their environment.
