SBA-Healt h
Earth Observation and
Environmental Modelling for
the Mitigation of Health Risks
by the EO2HEAVEN Consortium
1
AIP-5, Societal Benefit Area HEALTH
This engineering report provides an overview of the work done as part of the application implementation
pilot phase 5 (AIP-5); thread SBA Health. It introduces three different scenarios that have been implemented to test the current GEOSS architecture, services, information models and knowledge representations in the context of health related scenarios. The report describes all three scenarios in a level of
detail that allows the reader to get a quick overview of the actions performed. For more detailed information, please be referred to http://www.eo2heaven.org.
In addition to the specific experiences made in the three scenarios, a number of general aspects need
to be emphasized that are relevant to the SBA health in general:
•
Data availability: Temporal and spatial resolution is always an issue, as data requirements raise
with available resolution.
•
Data discovery and accessibility: In particular health data is barely available at standardized interfaces. Often the only way of electronic publication is via pdf files.
•
Spatial variability in urban environments requires extremely high spatial data resolution.
•
Satellite data to ground level conversion/mapping is very challenging. Existing literature often
over-simplifies this task, as specific situations, e.g. land/sea situation, close to equator with high
cloud coverage, etc. are often ignored and the mapping has been calculated for ideal conditions
only.
•
Time consuming data cleaning tasks: Earth observation as well as health data is often provided
in formats and maturity levels that require intensive data cleaning, gap filling, interpolations, error
handling etc. Those tasks should be more automated / provided by the services providing access
to the data.
•
Data-to-tools mapping often very poor: Existing Web services often make use of complex hierarchical, often XML-based data structures, whereas the majority of statistical and AI tools requires
flat data structures.
•
Lack of in-situ observation stations is general an issue, as satellite data calibration and evaluation depends on the availability and quality of reference data.
•
Usability of health data: Data privacy policies and regulations make it often impossible to access
the raw data directly. Data aggregation often adds artificial uncertainty to the data.
•
Usability of GEOSS tested and confirmed to be functional. The general architecture with distributed services and data supports all requirements set by SBA health. Only more sophisticated
security has been missed in some cases. In particular OGC Web service interfaces work well
with both EO and health data
•
GEOSS is often unknown to many domain researchers. It is shocking to what low extent GEOSS
is in fact known outside but even inside the earth observation community.
In the following, this reports describes the three scenarios implemented and analysed in the context of
AIP-5. Scenario one on environmental effects on respiratory and cardiovascular diseases in Dresden
and Saxony identifies health and environment agencies as well as researchers as the main users of the
system. The system itself integrates in-situ, remote sensing, and health data and provides all data at a
web-based information system. The goal is to develop a system featuring maximal ease of use for the
different user communities, i.e. it offers rich data sets in an optimal fashion.
The second scenario on environmental challenges to health in the south Durban industrial basin has
requirements very much similar to the first scenario. Just, after intensive testing and experimenting, the
requirement to integrate remotely sensed data was dropped, as the correlation coefficients proofed poor
correlations between available data sets. In order to optimize system performance and ease of use, an
integrated solution was developed.
The third scenario on water-borne infectious disease in Uganda has two dimensions to it. On the one
hand, we identified a number of requirements and processing steps to provide an optimal tool set for
scientists to tackle research questions related to water-borne diseases. On the other hand, those tools
need to be implemented within an improved national health system with more efficient and reliable reporting and communication structures.
2
2
Case Study 1 - Environmental effects on respiratory and cardiovascular diseases in Dresden and Saxony
The Saxon Case Study focuses on the environmental effects of ozone, particulate matter (~10 micrometres or less), sulphur dioxide and nitrogen dioxide on respiratory and cardiovascular diseases in the
German federal state Saxony. Although air quality can be considered to be good on average, the
thresholds (according to the EU-Directive 2008/50/EC on ambient air quality and cleaner air for Europe)
for ground-level ozone, particulate matter and nitrogen dioxide are regularly exceeded at the measuring
stations.
There are two main goals in the Case Study:
1. Study the potential impact of air quality on human health.
2. Develop an information system providing the means for analysing and visualizing environmental and health information.
As health authorities are responsible for health promotion, health reporting as well as social and health
advisory services, they are considered the main users of the information system. Additional users comprise environmental agencies (e.g. the Saxon State Office for the Environment, Agriculture and Geology), medical and environmental scientists and enterprises related to health or environment.
Targeted users
To identify needs and expectations towards a common system for health risk prediction, a number of interviews and a training/stakeholder workshop have been conducted; mostly with local health and environment authorities, but also health insurance employees, scientists and physicians. In general, the authorities are interested in investigating current health risk situations to inform the public accordingly.
From the interviews and workshop, the following conclusions can be drawn:

Almost all stakeholders consider it useful to analyse the correlation between health and environmental data, most likely the influence of air pollution on respiratory and cardiovascular diseases.

The majority of stakeholders feel insufficiently informed about data sets from additional and
possibly unknown sources (e.g. images taken by satellites). Thus, detailed information on
available datasets from different sources is requested.

Most of the stakeholders experience deficits in current data analysis. They wish clear advice on
how to deduce a certain health risk from environmental information. Therefore, a kind of best
practice guidance for environmental health-risk analysis is requested.

Risk and exposure maps were identified as most useful in terms of information provision. As
most of the stakeholders already use the Internet as information medium, an online system is
preferred.

Concerning the spatial resolution, administrative areas are seen as being most applicable. Furthermore, seasonal information is most commonly demanded for temporal resolution. However,
it is also stated that the spatial and temporal fluctuation of the air quality parameter should be
considered.

In addition to maps, further information on descriptive statistics, correlation and trend analysis is
requested. Animated visualizations are preferred, to show spatial and temporal changes.

An exposure and risk prediction system is seen as highly valuable, especially for a long-term
period.
Benefits for the health authorities
The main applications for the project and its result, as seen by the stakeholders, are the provision of reliable information on air quality and health issues, an improvement of public relations as well as the
possibility of a better prevention of air pollution. Following these views, the information system could be
used by authorities to receive more in-depth information on the impact of air quality on human health,
monitor air pollution and corresponding health risks and give recommendations to the public. Furthermore, the provision of a best practice guide could support institutions to run their own analysis. The au-
3
thorities would thus be equipped with the necessary information and tools to identify adverse health effects from air pollution, supporting initiatives for a well-directed and hence more efficient reduction of air
pollution.
Tasks to analyse environmental effects on respiratory and cardiovascular diseases
Figure 1 shows consecutive tasks as part of the information system identified for this Case Study, followed by the respective requirements:
Figure 1: Tasks to analyse environmental effects on respiratory and cardiovascular diseases carried
out in the information system for CS1.
Data acquisition and preparation requirements
Data from the following fields is required for the analysis in the project.
Health datasets: In order to be able to establish a robust association between environmental pollution
and health effects, a test health dataset with valid and reliable diagnoses is required. The ‘AOK Plus’,
one of the German statutory health insurance companies, provides one of their primary care billing datasets as a surrogate for the test health data set. Further datasets on hospital admissions are provided
by the Federal states’ statistical agencies (Forschungsdatenzentrum, FDZ).
In situ datasets: Following the EU-Directive 2008/50/EC, pollutant concentrations and additional meteorological parameters are measured half-hourly by an in situ station network operated by the Saxon
State Office for Environment, Agriculture and Geology (LfULG). The data can be obtained from the
LfULG Website or the European Air quality database (AirBase) as tabular data.
Remote sensing datasets: The major advantage of remote sensing products (e.g. satellite imagery) is
the areal characteristic of data recording. This characteristic is described by means of four different
kinds of resolution: geometric, radiometric, spectral, and temporal. All available datasets have been
checked against those resolution aspects to decide on usability for the EO2HEAVEN information system.
Usually, the observed phenomena are measured between the satellite and the earth’s surface. For the
case study analysis, only the phenomena on the earth surface are relevant. Thus, further datasets
about meteorological conditions near to the earth’s surface are required to eliminate the irrelevant conditions in the atmosphere.
Micro-drone datasets: Compared to satellite techniques, micro-drones have no orbital restrictions, the
equipped sensor and camera systems can be changed and the flight path is easy to manipulate. Thus,
4
for applications at a local scale it seems promising to use data collected by micro-drones that allows a
higher temporal and spatial resolution than offered by satellite systems.
Data analysis requirements
Air quality modelling: The modelling of continuous air quality information from in situ sensor observations is of great interest for environmental and health risk analysis. However, the accuracy and reliability
strongly relies on the number and distribution of in situ sensor stations. Because of the small number of
in situ sensors for the state of Saxony, standard interpolation techniques are difficult to apply especially
in sparsely covered regions. Therefore, interpolation techniques need to be developed taking into account the sensor distribution, certain pollutant characteristics as well as uncertainty of measurements
and models.
Health risk analysis: To derive health risk from air quality data a corresponding health risk analysis covering respiratory and cardiovascular diseases needs to be conducted. This analysis will be based on
prescription data from the AOK public insurance company as well as mortality and morbidity statistics
from the German Research Data Centre. The outcome of this analysis (e.g. a mathematical model) can
be used to create risk maps and to facilitate the development of a Spatial Information System for responsible health authorities.
User interface requirements
Besides providing tools and means to address the above-described requirements for the information
system, the requirements of its user interface are important. The main user interface shall focus on the
animated visualization of spatio-temporal data (including time series). Visualization of merged environmental, health and spatial reference data is required:
1. The change of temperature and air pollutants (including air quality indices) over time and space
in a map based on administrative units.
2. The change of health information over space and time in a map based on administrative units.
3. The change of correlated environmental and health data showing the development of possible
risks over space and time in a map based on administrative units.
4. Time series of environmental and health parameters as a diagram (showing the variation over
time for a single administrative unit)
The results shall be web-accessible with an Internet browser. A sketch of how the pilot implementation
could look like is depicted in Figure 2.
The following elements, even if not all shown on the sketch, should be included:

The risk and exposure map visualization, including an overview map, a map selection and a
legend graphic with corresponding explanations.

Interaction with the map to gain detailed information on administrative units, like the current risk
classification, further descriptive statistics or time series.

An information button leading to more detailed information on the impact of air quality on human health, the used risk classification, the methods used for air quality modelling and prediction as well as starting points and recommendations for analysis and information dissemination.
5
Figure 2: Sketch of the pilot implementation for CS1.
6
3
Case Study 2 - Environmental challenges to health in the south Durban industrial basin
The relationship between industrial pollutant exposure and adverse respiratory outcomes has been well
documented in the scientific literature. Much debate exists around causality of effects, role of specific
pollutants and the populations particularly vulnerable to elevated pollution. Although certain pollutants,
such as oxides of nitrogen, particulate matter and sulphur dioxide are known to result in adverse outcomes, the ability to use this information in developing interventions to improve the life of affected and
vulnerable sub-populations is limited.
The Durban South Industrial Basin (DSIB) is at particularly high risk for exposure to significant levels of
ambient air pollution because of its geographic relationship with certain stationary sources of air pollutants. Specifically, two major petroleum refineries are within the community, together with a pulp and
paper manufacturer, a waste water treatment plant and several small to medium industries. The community is believed to be at risk for intermittent substantial exposure to ambient air pollutants. Available
data on sulphur dioxide indicate that average and/or maximum exposures at sites in South Durban have
frequently exceeded World Health organization (WHO) and the South African National Ambient Air
Quality Standards. Health studies in the area have indicated elevated risk for respiratory outcomes
among those exposed.
The Durban Case Study focuses on the environmental effects of particulate matter (~10 micrometres or
less), sulphur dioxide and nitrogen dioxide on the respiratory health of residents, especially children, in
the South Durban Basin.
The primary objective is to be able to develop a real time spatial-temporal monitoring system that will
provide information to the health authorities about elevated pollution and potential health risks in general, and identify specific areas of elevated pollution within the affected communities.
The main goal in the Case Study is to develop an information system providing the means for analysing
and visualizing environmental and health information.
Users
The health authorities, Environmental Health Services Managers and Environmental Health Practitioners are responsible for health promotion, health reporting as well as social and health advisory services
and as such they are considered the main users of the information system. Additional users comprise
the Pollution Control and Risk Management department and medical and environmental scientists.
Key health authorities, management and health workers, medical and environmental scientists were
consulted via workshops and interviews about their needs and environmental challenges they faced in
the South Durban Basin. In general, the health authorities are interested in receiving real-time air pollution information and corresponding health risk situations. The following needs and challenges were
identified during the workshops and interviews:

Almost all stakeholders consider it useful to analyse the correlation between health and environmental data, most likely the influence of air pollution on respiratory diseases.

The majority of stakeholders felt that there was a lack of summarised real-time air quality information from the current data acquisition system. Thus, summarised and easy to interpret information on air quality and health risk need to be provided.

The stakeholders expressed the need for information air quality and health risk while working
out in the field so that they are able to act immediately in the event of an air pollution episode. A
web application that is available on a smart phone and tablet would be useful to accommodate
health workers in the field.

Concerning the spatial resolution, primary health clinic (PHC) areas are seen as most applicable. Furthermore, real-time information is required every ten minutes. However, it is also stated
that the spatial and temporal fluctuation of the air quality parameters should be considered.

Requirements for visual presentation in the form of maps, graphs and pollution roses have
been identified. Descriptive statistics, correlation and trend analysis are also requested.
7
Benefits for the health authorities
The main applications for the project and its result, as seen by the stakeholders, are the provision of reliable real-time information on air quality and health risk to enable the health authorities to make informed decisions in the event of increased air pollution events.
The Pollution Control and Risk Management (PCRM) section: This section is part of the eThekwini
(Durban) municipality. They are responsible for monitoring pollutant levels, and determining the sources
of such pollutants, enforcing industrial practice regulations and identifying mechanisms to reduce exposure. This agency currently manages the Air Quality Monitoring System (AQMS) for the city. This is a
real time ground level monitoring system located in fixed “hotspot” locations. This agency will be among
the key users of the proposed system. The system aims to incorporate and provide access to data and
information from the current AQMS. By using this system, this agency is able to determine conditions
likely to result in elevated pollution, and engage with the other stakeholders to either reduce pollutant
emission, or advise on responses to the anticipated increase in adverse respiratory effects.
Despite having a state of the art in-situ monitoring system, resources prevent the implementation of a
more extensive system in this major industrial region of Southern Africa. This results in many key zones
in the city that are not adequately monitored for key pollutants that are likely to produce adverse health
outcomes. The EO2HEAVEN system shall enable the PCRM section to make decisions based on the
air quality and health risk information in other areas of Durban.
The Clinical Services department (Health Unit, local government): This agency is responsible for the
management of the community-based clinics throughout the city. The system shall provide this agency
with information about the likelihood of increased patient load with adverse respiratory outcomes at the
local clinics when environmental conditions result in elevated ambient pollution.
The District Environmental Health Managers (Health Unit, local government): In each district within the
city, environmental health services managers (EHSMs) take responsibility for reviewing the daily pollution levels, determining sources and implementing corrective or regulatory action. Receiving support
from Pollution Control and Risk Management, EHSMs are required to make management decisions
generally with limited data and time. The system being developed shall provide these managers with
more complete data, together with additional meteorological parameters. This will allow for recognising
adverse conditions, and identify sources of pollutants more rapidly.
Researchers: Researchers at University of KwaZulu Natal, who may wish to better understand the environment and pollution interactions, may wish to run their own additional analysis based on the data provided.
Following these views, the information system could be used by authorities to receive more in-depth information on the impact of air quality on human health, monitor air pollution and corresponding health
risks and give recommendations to the public. Furthermore, the provision of a best practice guide could
support institutions to run their own analysis. The authorities would thus be equipped with the necessary information and tools to identify adverse health effects from air pollution, supporting initiatives for a
well-directed and thus, more efficient reduction of air pollution.
Tasks to analyse environmental effects on respiratory diseases
Figure 3 shows consecutive tasks identified for this Case Study, followed by the respective requirements:
8
SOS Server
Ftp, csv files
WD
WS
AMP1
AMP2
DELTA
RH
BP
ASDM Definition
Static param.
Grid 250x250 meters
shapefile format
last 3 days/10 min
Population
Land Cover
Topography
Traffic
Emission Index
Grid of Points 250x250 meters
CSV format
last 3 days/10 min
Grid 250x250 meters
shapefile format
last 3 days/10 min
The input of the model is only one station with a priority systems
Ftp, csv files
R
Conversor
Conversor
(point to area)
Met. Parameters
Municipally Server
Tranform
ADSM Model
R
Env. Param.
Met. Param.
(5 min)
Pollutant Param.
Pollutant
(10 Param.
min)
(10 min)
Met. Parameters
Met. Parameters
(10 min)
(10 min)
Wd
Ws
PM10
NOx
SO2
Gather Rose Model
Data per point
Temporal
aggregation
(1 day)
Wd
Ws
AMP1
BP
Health
parameters
Spatial
Aggregation
PM10
NOx
SO2
PHC
shapfile format
last 3 days/10 min
Rose diagrams
(jpg)
Pollutant. Param.
(daily)
Met. Parameters
(daily)
Pollutant Param. per Area
PM10
NOx
SO2
Rose model data
Rose Model
Health Model
definition
Pollutant Param.
PM10
NOx
SO2
Temporal
aggregation
(1 day)
Spatial
aggregation
HRRM per Area
Health Relative
Risk Map
Health Hazard
Index Model
Pred
Class
RSE
EO2HEAVEN System
Pred Max
Pred Median
Class Max
Class Median
shapfile format
last 3 days/daily
geotiff format,
Last 3 days/daily
Acquisition of meteorological raw data
Execution of the health risk model
Pre-processing of the meteorological raw data
Spatial aggregation per PHC area of pollutants concentration and health risk map
Execution of the ADMS model
Conversion of met data and pollutants concentration into shapefile
Temporal aggregation of met. data and pollutants concentration
Figure 3: Sketch of the workflow in CS2.
Data acquisition and preparation requirements
In situ datasets: The meteorological parameters from the in-situ monitoring stations will be transferred
from the eThekwini Municipality servers every 10 minutes.
Health datasets: The Durban Health Study conducted by the University of KwaZulu Natal (UKZN) will
provide the health datasets.
Data analysis requirements
Air quality modelling: The modelling of continuous air quality information from in situ sensor observations is needed for environmental and health risk analysis. However, the accuracy and reliability strongly relies on the number and distribution of in situ sensor stations. Because of the small number of in situ
sensors for the Durban area modelling techniques need to be applied to determine air quality concentrations away from the in-situ stations. Sensor distribution, certain pollutant characteristics as well as
uncertainty of measurements and models must be taken into account when modelling.
Two models will be used. The first is the Atmospheric Dispersion Modelling System (ADMS), which is
an advanced atmospheric pollution dispersion model for calculating concentrations of atmospheric pollutants emitted both continuously and intermittently. The modelling system needed is ADMS-Urban
which is a comprehensive tool for tackling air pollution problems in cities and towns. For more information, please see: www.cerc.co.uk/environmental-software/ADMS-model.html.
The second is the Pollution Rose Model (PRM) which produces a pollution rose for the last 24 hours using the pollution concentrations from the ADMS and meteorological data from the eThekwini Municipality server. The PRM is executed on-the-fly.
Health Risk Modelling: Another model written in R language is required to determine the health relative
risk pollution based on the concentrations from the ADMS and meteorological data from the eThekwini
Municipality server. The outcome of this model is a health risk map and a graph indicting the pollution
concentration on the primary axis and health relative risk on the secondary axis.
9
The output of Case Study 2
1. A Pollution concentration map for each priority pollutant (SO 2, PM10 and NOX) created every ten
minutes.
2. Pollution rose for the last 24 hours (see Figure 4Error! Reference source not found.).
3. Summary screen/report detailing animated, graphical and descriptive information on air quality
per PHC area (see Figure 4Error! Reference source not found.).
4. Summary screen/ report detailing animated, graphical and descriptive information on health relative risk per PHC area.
Figure 4: Summary display screen in CS2.
10
4
Case Study 3 – Water-borne infectious disease in Uganda
The cholera case study focuses on two problem dimensions: Firstly the links between environmental,
including climatic, variables and the outbreak of cholera in Africa, targeting more the research community. Secondly the handling of cholera cases including appropriate predictive and reactive actions on a
various scales within the health system, targeting more the health system participants and decision and
policy makers.
4.1
Links between environmental variables and cholera
Environmental and human related data are analysed together with health data in order to characterize
and explain the spatial and temporal patterns in cholera outbreaks in Uganda. The Case Study includes
the use of remotely sensed data, in-situ data from weather stations and ex-situ data. Ex-situ data are
data collected as part of a field sampling campaign.
The main purpose of the Cholera Case Study is to provide insights into the dynamics of cholera at the
pathogen level and as a disease and to use these insights for decision-making purposes. In order to
accomplish this it is necessary to develop tools to (1) support researchers to collect, handle, analyse,
integrate and model environmental, human related and health data from different sensors and sources;
and (2) visualize data and results for research purposes as well as decision making at different levels.
Historical and current data are used.
The outputs and outcomes of the Cholera Case Study can be used to seed the development of an Early
Warning System.
User groups
The work conducted as part of the case study is aimed at several user groups, including researchers
such as epidemiologists and disease modellers; policy and decision makers at country and global level;
health practitioners at the clinic and hospital level; health officials at different levels in government; and
community representatives who have access to communication and/or electronic tools to receive regular updates on the potential of an outbreak.
Study area
The process of extracting meaningful, actionable information from the combination of health, social and
earth observation/environmental data is difficult and complex. Difficulties are magnified in less developed countries (where the burden of disease is felt), for access to datasets is limited by technical factors such as ICT (Information and Communication Technology) connectivity and poor data as well as
human factors like a lack of training in the complexities of working with earth observation data. The
cholera case study is being conducted in the area around Kasese, southwestern Uganda, interesting for
its proximity to Lakes George and Edward and the equator.
Data analysis of environmental variables, field samples and case data
Data analysis processes are the mechanism by which the potential role of environmental and human related factors in cholera dynamics in Kasese is determined. There are, broadly speaking, three steps.
First, data about different environmental variables, typically from satellite remote sensing sources, are
analysed together with cholera case data. The data are aggregated at different spatial and temporal
resolutions and validated against in-situ observational data, before being run through correlation checks
(based on statistical, signal processing and visualization techniques) to determine similarities in underlying patterns and if any relationships exist between the case data and the individual environmental variables. Further analyses of individual variables (using the same or similar techniques) is undertaken to
tease out which ones are important in relation to each other. The case data and local environmental
variables are then related to global environmental variables, such as the ENSO (El Niño – Southern
Oscillation) signal, to test if these large scale variables can be linked to cholera outbreak.
Second, data collected in field sampling campaigns are analysed (using microbiological techniques for
example). The presence/absence of cholera is checked for, and if present, whether the toxigenic strain
is visible. These results are compared with the individual environmental variables and are also analysed
in context of seasonality. Essentially, the analyses are searching for environmental and climatic drivers
11
of the presence or absence of the bacteria. An aim is to identify thresholds for important environmental
variables, above or below which changes in the cholera pathogen and the disease itself can be expected.
Third, the results from the analyses of case data, environmental variables and sampling data are put into context of social and geographic data – population density, land use practices, demographics, topography, water use practices etc. - and then tied to epidemiological data.
4.2
Cholera case surveillance and health system enhancement
Using a predefined framework of essential environmental health services (core functions), adapted from
the Center for Disease Control and Prevention (CDC) Environmental Health Services (EHS), to identify
gaps in the current (environmental) health system in Uganda showed that the effectiveness of information provision is limited, compromising the effectiveness of responses and proper decision support
for policy makers. In Kasese, cholera case detection is currently realized by a passive surveillance system, only proceeding to active surveillance after an outbreak has been confirmed. Yet, the vertical
stratification of the health information system over multiple administrative levels hampers the prompt
dissemination of information. This also introduces errors in the information flow, potentially delaying the
signalling of an outbreak when it occurs. Even though alternative procedures exist to flag initial cases
for follow up investigations, the protocols are fuzzy and the system is prone to human error. Moreover
the lack of resources limits the efforts for efficient active surveillance. It appears that the flow of information from the peripheral health facilities to the ministry of health is hampered by lack of tools to collect, store and disseminate this information. Also a considerable amount of resources is invested in collecting information and summarizing data in district reports, which are then again used to report to the
national government.
Based on the gaps identified, the project tackles the requirements and corresponding mitigating responses concerning the aspects data collection as well as data processing and visualization.
User groups
The Cholera Case Study is expected to realize benefits to different stakeholders and user groups.
Health workers would be supported by tools that enable timeous and accurate capture of case data in
the field and tools that ease the recording and storage of field study samples and sample data. Officials
from government or other agencies would be supported with tools to report case data in various required formats. Tools to visualize case data, sample data, social data and analysis results in a compact,
digestible fashion will be useful to them too. The insights gained into cholera disease drivers and science behind the insights will support enhanced decision-making.
Training events and materials will provide a long-term mechanism to sustain these benefits.
Data collection
There is an urgent need for the development of a standardized data collection protocol and associated
data collection tool facilitating data collection and access to health information across all levels of the
health infrastructure. Such tool would be of great benefit to both health practitioners working in the field
as well as health authorities working at the national ministry of health, as it: 1) reduces the number of
human errors induced by the current reporting system, 2) reduces the amount of resources spent on
data management and dissemination, 3) increases response time, 4) increases the effectiveness of
surveillance by facilitating the exchange and interpretation of information at a sub-district level and 5)
facilitates communication across all levels of the health system. Among other means, there is a clear
requirement for using a robust and simple hand held data processing tool for monitoring new cholera
patients, for example as an app on a smart phone or tablet computer.
Data processing and visualization
To maximize the benefit of the limited resources available the utilization of the generated information
should be optimized. The information generated can be used together with a robust and simple data
processing protocol to be used in various applications: 1) Various representations of the data by means
of mapping and charting epidemiological data allow for a more ridged understanding of the scale and
severity of a cholera outbreak (or equally of any other environmentally driven disease). 2) The early diagnosis of new patients and identification of the hazards posed by the environment are further en-
12
hanced by the rapid identification of areas reporting new patients under conditions favourable to cholera. 3) Tools to visually assess and access cholera data will facilitate the use of this information to monitor the progression of an outbreak and to evaluate the effectiveness of the response.
5
Acknowledgement
This report was supported by the EC co-funded project EO2HEAVEN (project #244100 - EO2HEAVEN
CP-IP).
13