Click to edit Master title style Using Pollutant Release and Transfer Register data for environmental health research Hackett, 1 C. ; Campbell, 1 S. ; Wine, 1 O. ; Osornio-Vargas, 1 A. 1. Faculty of Medicine and Dentistry, University of Alberta Introduction Research Methods Discussion Broad searches of the published literature were performed using “PRTR” and related search terms in multiple databases of scientific literature as well as in Canadian Newsstand. Results were included if they matched one of the three following criteria: In reviewing the publication record around PRTR data, several trends can be observed. Few articles were published until 13 years after the release of the first set of US TRI data. A greater lag was seen for health studies, which used more complex methods. The use of Canada’s NPRI may be hindered by a similar time lag, as the scientific community develops methods to use the data and as reporting requirements and accuracy of reporting stabilize. The challenges to the use of PRTR data in the Canadian context may be better understood by analyzing the use of other nations’ PRTR data. • Click to edit Master text styles Pollutant Release and Transfer Registers (PRTRs) are national databases of chemical releases reported by industry to governments. The importance of pollutants to health is well documented, however the use of PRTR data to study relationships between health outcomes and pollution has not been evaluated. A preliminary literature search found few published studies that used data from the Canadian PRTR. Understanding past uses of PRTR data is important in the evaluation of their applicability to future health studies. – Second level 1. Peer-reviewed studies that investigated the relationship between PRTR data and health outcome data. 2. Peer-reviewed studies that used PRTR data in other ways. 3. Canadian media articles referring to the use of PRTR data. • Third level Objectives: 1. To investigate how PRTR data has been used to date by the scientific community, and 2. To review the challenges to the use of PRTR data for future Canadian environmental health research. – Fourth level » Fifth level Many nations’ PRTRs were represented in the literature search, including Australia, Canada, Japan, Spain, the UK and the US, but the largest volume related to the United States’ Toxics Release Inventory (TRI). The literature search identified 41 health studies, 139 studies using PRTR data in other ways, and 203 media articles referencing use of Canada’s NPRI data, of which few referenced scientific studies. Country United States Canada Name Toxics Release Inventory To inform communities about toxic chemical releases To support informed decision-making by industry, government, non-governmental organizations and the public 1987 682 National Pollutant Release Inventory To improve public understanding First Data Chemicals Tracked (2010) Health 35 Studies Other Studies 95 To identify pollution prevention priorities, the assessment and risk management of chemicals, and air quality modelling 1993 346 0 A variety of limitations, both in the data and in study design are apparent in the studies. 30 25 20 15 10 5 0 Both Location Volume Containment Buffer Decay Lags between exposure and health effects were addressed in some designs by studying fetal and child health. Some studies addressed the lack of a toxicity standard by using a variety of toxicityweighting systems. A limitation of many of the health studies was the use of location or total volume, neither of which is an accurate proxy for toxic exposure or potency. It is difficult to use topography, meteorology and release rates to model chemical dispersion and therefore exposure to pollution. The literature revealed three main methods of studying spatial associations of pollution and health outcomes, each with its own limitations. It may be important to address this in study design by using multiple methods. Conclusion Method of Correlation Figure 2: Spatial correlation methods used in studies of health data, divided into the type of PRTR data used The use of PRTR data in environmental health research hinges on: 17 Table 1: Comparison of US and Canadian Pollutant Release and Transfer Registers (PRTRs) Number of Articles - Health data studied included cancer incidence, population mortality, and birth defects. Most designs used a containment method, which related health data and the number of facilities or the volumes of pollutants released in defined areas. Buffer methods assumed a uniform exposure in rings around facilities, and decay methods modeled exposure inversely related to distance from facilities. 68%, 50% and 100% of the studies using a containment, buffer and distance decay method, respectively, found a positive correlation between pollutant release and negative health outcomes. Number of Studies Results Purpose Specific data were extracted from each article. Only articles written in English were included. One of the greatest challenges in the use of PRTR data is the limitations of the data itself. Though PRTRs often lack data from smaller emitters, and industry may report inaccurately, it is often the most comprehensive national data available. No study considered all identified limitations. Some of the studies addressed some of the limitations through study design, though this increased the complexity of the studies and made conclusions more difficult. Conclusions about the relationship between health and pollution were also made difficult due to potential confounding demographic variables. 90 80 70 60 50 40 30 20 10 0 TRI NPRI Year Figure 1: Total number of articles published using US TRI data Limitations Identified Imputable to PRTR data Lack of non-threshold emissions reporting (35) Change in reporting requirements over time (30) Estimation errors and assumptions in data reporting (28) Does not track all chemicals in use (24) Lack of mobile and/or area sources (19) Underreporting of emissions (19) Limitations Identified Imputable to Study Design Confounding variables (63) No dispersion modelling to estimate exposure (50) Lack of toxicity standard (49) Aggregation of population data and exposure (20) Modified Areal Unit Problem (14) Lag time between exposure and health effects (14) Table 2: Limitations identified in studies using PRTR data 1. Acknowledgement of the limits of the data, 2. Evaluation of exposure to chemicals, and 3. Correlation between chemical exposure and health outcome data. A review of the literature using Pollutant Release and Transfer Register data revealed that the scientific community has found methods to investigate associations between health outcomes and pollution. Though PRTR data is the most complete pollutant data available, the limitations of the data present study design challenges for environmental health research. Associated Press Images, Image ID: 090110020006 Acknowledgements Funding for this project was generously provided by the University of Alberta Office of the Provost and VP (Academic) and from an Emerging Research Team Grant administered jointly by the University of Alberta Faculty of Medicine & Dentistry and Alberta Health Services, Edmonton Region.