Vishnu Menon
ASU ID:1203320386
REPORT
LAHVA: Linked Animal-Human Health Visual Analytics
By Ross Maciejewski, Benjamin Tyner, Yun Jang, Cheng Zheng,
Rimma V. Nehme, David S. Ebert, William S. Cleveland,
Mourad Ouzzani, Shaun J. Grannis and
Lawrence T. Glickman
This paper proposes to use data sets obtained from both Human hospitals and veterinary hospitals to help visualize the spread of diseases and essentially help in understanding the source of disease outbreaks and whether it can be predicted in advance by monitoring the health of the animals in the locality.Timely and accurate detection of unusual population health trendsis a challenging problem requiring the analysis of data collectedfrom disparate sources over time. These data sources vary widely inaccuracy and reliability, and it is often the case that unusual healthtrends, such as outbreaks or poisonings, often have an incidenceprofile (signal) that is obscured by the statistical noise.Further the authors state that the LAHVA tool eliminates many of the issues faced by similar tools like duplicate records, incomplete data fields, data quality concerns etc.Essentially this information will serve as an early warning system. However combining information from these two fields to create a dataset that is compatible and usable by a visual analytics tool is a tough ask as the parameters recorded and disease spread information vary greatly in humans and animals. The LAHVA tool provides a framework for understanding and analyzing both human emergency room data and animal veterinary data. The data sets used are not random but actual patient records obtained from the Indiana Networkfor Patient Care (INPC) and all general visits to the Banfield PetHospitals. Care has been taken to ensure that patient privacy and confidentiality has not been compromised in any way.
The authors of this paper have documented the use of this tool with two case studies:
1) seasonalinfluenza and its correlation to general companion animal health,
2) the effects of an industrial wastewater release on companionanimals and the correlation to potential human health issues.
In the first case, the visualization by the tool suggested that there may be a relation between the onset of seasonal influenza and dog respiratory symptoms.In the case of the industrial wastewaterrelease, several syndromes for both cats and dogs were analyzed
and preliminary results indicated that the industrial wastewater releasenegatively influenced the health of companion animals in thisregion. In both cases (especially the second) the information is what is to be expected by common sense, but what this shows is that this tool can also be effectively used in a myriad other situations and can prove very useful.
Figure 1:LAHVA screen shots.
Figure 2: Using LAHVA to identify seasonal influenza.
Figure 3: Yearly pattern for human and dog respiratory syndromes.