SAHSU Studies
What is the Rapid Inquiry Facility?
The Rapid Inquiry Facility (RIF) has been developed by the Small Area
Health Statistics Unit (SAHSU) at the Department of Epidemiology and
Public Health, Imperial College London.
The RIF is an automated tool that provides an extension to ESRI® ArcGIS functions, and uses both database and GIS technologies. The purpose of this facility is to rapidly address epidemiological and public health questions using routinely collected health and population data.
The RIF can perform risk analysis around putative hazardous sources, and can be used for disease mapping. It generates standardised rates and relative risks for any given health outcome, for specified age and year ranges, for any given geographical area.
This facility was initially designed as a tool for SAHSU staff to analyse routinely collected health data in relation to environmental exposures in the UK, but this UK RIF was subsequently transformed for use by several European countries as part of the European Health and Environment Information System (EUROHEIS) project.
Centers for Disease Control and Prevention (CDC) and SAHSU are collaborating to adapt and enhance the UK RIF software for use in CDC’s National Environmental Public Health Tracking (EPHT) Network.
The goal of this project is to increase the functionality and versatility of the RIF for evaluating temporal and spatial relationships between disease and environmental hazards in the National EPHT Network.
RIF development
The software has been reprogrammed to be database independent.
In addition to the point source ‘hazard analysis' and disease mapping options, it is now also possible to import detailed exposure data, such as output from dispersion modelling.
RIF provides a tool that allows users with skills in epidemiology take advantage of the many of the functions that a GIS offers without requiring an in-depth knowledge of GIS. Since the application is embedded in ArcGIS, those with GIS skills will of course still be able to use all the functionality that ArcGIS offers.
Within the disease mapping tool, the RIF also performs empirical Bayes smoothing of the raw relative risks.
The RIF can also export data for further analysis in other (statistical) software packages such as WinBUGS and SaTScan.
The new RIF system is being tested in several case studies in the UK and USA, including a pilot project studying cancer outcomes associated with living over a plume of trichloroethylene contaminated ground water in Utah.
RIF output
The RIF automatically generates contextual maps showing the area under
study.
A report is generated summarising the study details, and reporting the crude and adjusted rates and risks for each health outcome investigated. Graphs comparing the age, gender and socio-economic (or other covariate) structure of the study and comparison populations are provided to aid interpretation.
As well calculating rates and relative risks (and associated 95% confidence intervals) for each exposure group or distance band, the RIF runs Chi-square tests for homogeneity and linear trend to test the global association between distance/exposure covariate and disease risk.
In disease mapping analyses, maps showing crude, adjusted and smoothed risks by area are also displayed.
Why develop the RIF?
There has been increased interest expressed in developing ‘environmental
public health tracking systems’ in the UK, as well as elsewhere
around the world. The RIF is able to rapidly link environmental and
health databases, and is thus a powerful tool for evaluation of spatial
relationships between disease and environmental hazards in such tracking
systems.
Demos
The demos are large Windows Media Video (.wmv) files. The files must first be unzipped
(use WinZip on Windows), and then
they can be viewed using Windows Media Player.
- Disease Mapping (1.06 MB)
- Risk Analysis (948 KB)
RIF-Related Links
- EPHT at the CDC: http://www.cdc.gov/nceh/tracking/
- SAHSU 'Partner profile' at the CDC: http://www.cdc.gov/nceh/tracking/sahsu.htm
RIF Software
Click here for information
on how to register for RIF software access.
Growing policy concerns about health inequalities highlight the importance of examining associations between deprivation and health outcome more comprehensively than has been done to date. We have therefore extended the work done on the Multiple Deprivation study, aiming at quantifying associations between socio-economic factors, such as income, employment and education, on the one hand, and major causes of mortality and morbidity, on the other. We are considering all cause mortality as well as mortality from specific causes and incidence of specific cancers. The purpose is to quantify the variation in disease rates at small area scale and determine the degree to which that variation can be explained by socio-economic factors.
SAHSU collects, integrates and analyses an increasingly large volume and varied range of information on environmental health in England and Wales. Many of the data have direct relevance to policy as indicators of the state of environmental health. As part of the SAHSU work plan, therefore, we will analyse and present these data in the form of an environmental health atlas of England and Wales. The atlas will comprise two sections: one describing patterns of exposure to environmental hazards across England and Wales (primarily chemicals, via various pathways, but also some physical risk factors, such as radon), and the second describing selected health outcomes on a geographic scale. Each will comprise a series of small-area scale maps, together with an interpretive commentary. Histograms and bar charts will be used to show statistical distributions across the population and trends over time (where suitable temporal data are available). To aid interpretation, contextual information – including population distribution, socio-economic status, urban/rural areas, selected geophysical characteristics (e.g. climate) – will be provided. Indicators to be included in the atlas are not yet finalised, but selection criteria will comprise policy relevance, data availability and quality, and interpretability (i.e. known or suspected links between environment and health). Examples of exposure indicators might include air pollution (e.g. NO2, SO2, PM10, ozone and benzene), contaminants in drinking water (e.g. THMs), noise from airports and roads, and modelled exposures to RF emissions from mobile phone base stations. Examples of health outcomes might include mortality (due to all causes, cardiovascular disease, respiratory disease and selected cancers), cancer incidence (e.g. leukaemia, lung cancer, prostate cancer, breast cancer) and hospital admissions due to cardiovascular and respiratory disease. Computation of these indicators will be carried out using state-of-the-art GIS and statistical techniques. Where appropriate, programs will be developed to automate this process, so that maps can be updated as required. Manual and automatic checking will be carried out during analysis, both to ensure the validity of the indicators and also to provide a means of checking the input data (an added advantage of these analyses will be that they provide a means of routinely mapping and validating SAHSU data).
Childhood onset (type-1) insulin-dependant diabetes mellitus (IDDM) shows spatial and temporal trends in incidence. Incidence varies widely across Europe and the world, and there has been a fairly consistent increase in incidence over time throughout most of Europe and in the UK. In addition, seasonal trends (especially in older children), spatial clustering, and time-space clustering have been observed. There is currently no national register of IDDM; however, records of type-1 diabetes in children admitted to hospital for this condition (estimated to be ~2500 per year) are likely to be nearly complete. SAHSU holds national hospital admissions data for the period 1991-2005, which could potentially be used to assess spatial and temporal trends in incidence of childhood diabetes across England and Wales. Diabetes is a chronic condition requiring regular treatment and is associated with many complications; as such it is likely that someone diagnosed with diabetes might be admitted to hospital on more than one occasion. The hospital admissions data would need to be cleaned to remove duplicate records, so that the remaining data represented the first admissions of each child for diabetes. To enable a validation of the cleaned hospital admissions dataset, childhood diabetes registrations by gender, age group and year have been obtained from the Yorkshire Register of Diabetes in Children and Young People for the period 1992-2000. This register covers the former Yorkshire region of West Yorkshire, North Yorkshire and the former county of Humberside (based on 1991 administrative boundaries). Records pertaining to cases of childhood diabetes in this same area of ‘Yorkshire’ were identified from the cleaned hospital admissions dataset to allow direct comparison. If the hospital admissions data prove to be an acceptable source of data on childhood diabetes, work can be carried out to:
- Check whether incidence rates fit with other data from England and Wales;
- Investigate time trends across England and Wales;
- Assess seasonal trends (by age group);
- Map childhood diabetes at an appropriate resolution across England and Wales;
- Assess space-time clustering as part of SAHSU's space-time clustering methods project.
Heathrow is the world’s busiest international airport, operates the country’s busiest bus / coach station and manages rail connections into London. Continuous urban development has resulted in Heathrow today being an integral part of West London, thereby affecting local air quality along with other sources of pollution such as road traffic. The local population is therefore exposed to a considerable amount of pollution coming from noise and aircraft exhaust emissions and from airport associated road traffic.
Past Health Impact Assessments (HIAs) have been carried out historically for all submitted planning applications for airport expansion or construction. They have identified several health impacts possibly related to exposure to airport noise and air pollution and a number of air pollutants have been found to have an effect on cardiovascular as well as respiratory health. Noise exposure has also been linked with cardiovascular disease.
Several studies have investigated the relationship between noise exposure and blood pressure, including the EU-funded HYENA study (2002-2006), which mainly assessed the impact of exposure to noise on blood pressure. See the HYENA website http://www.hyena.eu.com for recently-published papers reporting the main findings.
We will now study health effects in the vicinity of Heathrow airport associated with air pollution and noise using SAHSU data on mortality, cancer incidence and hospital admissions.
Comprehensive SAHSU Data Documentation has been produced. This includes an introduction to the dataset, the tables and logical views, field definitions, annual record counts and column frequencies.
The aim of the SAHSU postcode dataset is to capture the temporal and geographic elements of a postcode in a single value, the SAHSU geo-reference or SGR. This can then be used to perform analysis. To achieve this, a history of all postcodes, including changes, is being created and validated, and includes grid references and accuracy.
Postcodes that refer to PO boxes, Northern Ireland, the Channel Islands and the Isle of Man can also be identified. Postcodes that have changed location can be identified together with the pattern of change and the year(s) (and month) of re-introduction.
The first comprehensive SAHSU Data Documentation has been produced. This includes an introduction to the dataset, the tables and logical views, field definitions, annual record counts and column frequencies.
Work has been carried out to clean postcodes, ICD9 and ICD10 codes; develop improved data verification and extraction tools and to introduce the use of version control software.
All new data sets are now put through the new postcode checking process. This will be reviewed later to improve the documentation.
The postcode sets have been cleaned to produce better matching, thereby improving the ED91 to postcode link. The cleaning involves conversion to the formal Post Office format with the correct number of spaces in the middle and the flagging of invalid characters. Feedback on postcode quality was provided to the National Down’s Syndrome Cytogenetics Register (NDSCR).
As an additional quality check, in future, all datasets will be analysed by region and year for a number of (to be defined) disease rates to ascertain that they have been loaded correctly and to better understand and document any limitations in the dataset. These will be compared to published rates.
Issues with the data extraction and processing carried out for some studies have shown the need for clear coding guidelines and standard processes. These are currently being drawn up. The aim is to ensure that data extracts are reproducible, and fully documented.