2007 Progress Report: Database and Tools for Investigation of Climate-Mediated Human DiseaseEPA Grant Number: R832753
Title: Database and Tools for Investigation of Climate-Mediated Human Disease
Investigators: Smith, Mark S. , Feied, Craig , Gillam, Michael
Current Investigators: Smith, Mark S. , Feied, Craig , Gillam, Michael , Handler, Jonathan
Institution: Washington Hospital Center
EPA Project Officer: Chung, Serena
Project Period: August 1, 2006 through July 31, 2010
Project Period Covered by this Report: August 1, 2006 through July 31, 2007
Project Amount: $443,420
RFA: Decision Support Systems Involving Climate Change and Public Health (2005) RFA Text | Recipients Lists
Research Category: Global Climate Change , Health Effects , Health , Climate Change
Create a database to support research and decision making related to human disease that is caused, triggered, modified or predicted by changes in climatologic conditions. This project will enhance understanding of the relation between climate change and human health.
Accomplishments over the past year have met the stated goals and objectives outlined in the grant proposal. In June 2006, an expert panel of physicians and researchers in public health and climatology was convened to identify areas of climate-mediated human disease of highest importance for study. Criteria for importance included assessing: likelihood of climate correlate; impact on human morbidity and mortality; and likelihood of actionable knowledge for mitigation. The panel concluded that the areas of greatest interest included: (1) gastrointestinal disease due to the high likelihood of climate correlation (2) cardiovascular disease – due to the high frequency of cardiovascular disease, even small climate mediated correlates will have large effects on human morbidity and mortality; (3) vector borne infectious disease (e.g. West Nile Virus). Importantly, the expert panel recommended the creation of a standard toolset that could be utilized to statistically evaluate a large variety of climate mediated effects. The first of those tools identified was the ability to correlate a continuous numerical variable (such as wind speed, temperature or barometric pressure) to a categorical set (such as patient diagnosis or chief complaint). By creating such a set of tools, a variety of climate-mediated studies could be facilitated.
Weather related data fields and sources of climate data have been identified from the National Climatic Data Center (NCDC). Only quality controlled data from the NCDC are being used in this project. Interfaces into the Microsoft Azyxxi ™ system have been instantiated by full-time Microsoft programmers at the Institute for Medical Informatics (NIMI) at MedStar Health in Washington D.C. These Azyxxi interfaces FTP data in a secure method from the NCDC and store them inside the hospital’s firewall on internal servers for processing. A complete list of the climate data fields being imported into Azyxxi is included as Appendix I. Initial software agent architecture to assist with data analysis is also complete. At the core of the system are Perl programming language modules that perform a variety of data retrieval and processing steps. To date, these software modules can pass data over web services for analysis to MATLAB or SAS. Processed data from these systems in XML (preferred) or other formats can then be presented to the end user.
Below is a screenshot from the live hospital system showing a list emergency department patient visits from Washington Hospital Center with presenting clinical problem (chief complaint)along with a sampling of some of the climate data now in the system. The current view below represents a sampling of from over 70,000 patient visits this year to the emergency department at Washington Hospital Center. Every patient visit is stamped with current and retrospective weather conditions for the closest weather station. Project Climate Query now has the ability to add climate data to any patient visit in the system for analysis.
As emergency department visits from other hospitals are added to the National Emergency Encounter Registry (NEER) through efforts of the American College of Emergency Physicians (ACEP), the Project Climate Query data core will automatically be able to add climate variables to those clinical patient data fields. ACEP’s goals are to create a patient registry spanning all emergency departments across the country. The combination of clinical and climate data makes possible the ability to study climate-mediated human disease at a national level at a scope that has not previously been possible. Through the tools being built within Project Climate Query, as more hospitals come on-line in the NEER, , any study that had been completed with the Project Climate Query data core previously will be easily duplicated and extended to the new hospitals thus increasing the breadth and sample size of the original studies; this will facilitate regional analysis as well. Project Climate Query will help the EPA fulfill its mission to protect human health by understanding the impact and role played by the environment.
Over the next twelve months, Project Climate Query data will be made available to researchers via Internet2 and via VPN access. Infrastructure needed for enabling Internet2 access has been created previously for the Department of Health in Washington D.C. and for Georgetown University Hospital through a previous National Library of Medicine grant. Initial facilitated data-mining software agents will be completed for the first chosen journal article whose study design will be replicated using Project Climate Query Data. A combination of Perl and C# .NET along with SAS and Matlab are currently being used. Analysis of a continuous numerical climate variable against a categorical patient data entity (i.e. correlation of rainfall to ED diagnoses) will be completed. Bi-directional communication to public health agencies will be instantiated through a partnership with EM System and utilizing web services.