Updates

Some updates on progress.

In June 2018, The DiSARM team attended an Indoor residual spraying (IRS) tool review and...

June 2018

Gearing up for 2018 IRS season in Botswana

In June 2018, The DiSARM team attended an Indoor residual spraying (IRS) tool review and training in Botswana. There were a range of users represented, including the Chobe District Health Management Team (DHMT) leader who is the main end user of Monitor module, Chobe Spray Team Leaders (the main end users of Collect) and Tutume DHMT head. The tool review was lead by a representative of CHAI with support from four colleagues.

Read more
Building on previous work to identify sprayable structures, the DiSARM team is developing a new...

December 2017

Using machine learning to predict roof type

Building on previous work to identify sprayable structures, the DiSARM team is developing a new ensemble machine learning algorithm to predict a structure’s roof type based on satellite imagery. Roof type is related to both building type (e.g. buildings with tile or thatch roofs are more likely to be residential) and wall type (e.g. buildings with thatched roofs are more likely to have walls made of natural materials like mud); further, wall type strongly influences the insecticide used for indoor residual spraying (IRS). This has important applications for the work that the DiSARM team is already doing and for new capabilities that we hope to bring to the platform.

Adding the roof type prediction probabilities to the building type algorithm improved its accuracy by a few percentage points – while this doesn’t seem like much of an improvement, it translates to thousands more structures being classified correctly, which in turn can lead to savings in time and money for IRS teams.

Accurate prediction of wall type from roof type would potentially result in further gains in efficiency for IRS teams. If we can predict wall type using satellite imagery of roofing materials, IRS teams could better plan how much and which insecticides are needed for a given campaign. This prediction capability could also be potentially expanded to predict other features of a building that could help in IRS planning, like number of rooms and surface area.

Read more
With support from the Clinton Health Access Initiative, DiSARM is being used to support the...

October 2017

DiSARM being used in 4 countries in southern Africa

With support from the Clinton Health Access Initiative, DiSARM is being used to support the planning, implementation and monitoring of indoor residual insecticide spraying campaigns in Namibia, Botswana and Zimbabwe and is being piloted in South Africa. Programs are able to combine risk maps with population and building data to prioritize areas for targeted spraying. In the field, teams can collect data offline which once synced can be visualized in near-real time using maps, charts and tables allowing operations to be refined if need be.

Read more
Link to How We Get To Next and Places Journal

September 2017

DiSARM featured in How We Get To Next and Places journal

Link to How We Get To Next and Places Journal

Read more
While incredibly detailed maps showing roads, buildings and rivers are available for most well resourced...

June 2017

Predicting sprayable structures using machine learning

While incredibly detailed maps showing roads, buildings and rivers are available for most well resourced countries, the same is not true for the less well resourced countries affected by malaria. OpenStreetMap, via the Humanitarian OpenStreetMap (www.hotosm.org) and Missing Maps projects (www.missingmaps.org), is helping to map some of the poorest places on Earth. In many countries, such as Swaziland, Botswana and Zimbabwe, there is nearly complete information on roads and buildings, providing a fantastic set of data to help plan malaria household intervention campaigns such as indoor residual spraying with insecticide, distribution of mosquito nets and mass drug administration. However, complete information relating to the building type is rare, making it impossible to distinguish residential from non-residential structures. Yet, this information is crucial for helping to plan and execute intervention campaigns targeted at households.

Read more
The intervention targeting module is one of several modules currently under development. It is comprised...

April 2017

Intervention targeting

The intervention targeting module is one of several modules currently under development. It is comprised of 4 components designed to help guide planning, execution and monitoring of intervention campaigns for malaria elimination. Parts of the targeting module were used by NMCP in Swaziland to define priority localities by risk and historic data and produce printed maps for 2016 IRS season. Responding to feedback from users and partners, the module has since been modified to further enhance current process flows.

The Monitoring component is a dashboard which provides an overview of the current state of the intervention campaign. It allows users to visualize progress of targeting teams in near real time, as opposed to having to wait until the end of an intervention campaign for this information.

Read more
Link to Google blog and Fortune Magazine

April 2017

DiSARM featured in a Google blog and Fortune Magazine

Link to Google blog and Fortune Magazine

Read more
Foci investigation was identified earlier on as one of the potential use cases for DiSARM....

November 2016

Developing foci mapper

Foci investigation was identified earlier on as one of the potential use cases for DiSARM. After our initial focus on IRS, we were invited to attend a National Foci Planning Meeting in Zimbabwe, Matabeleland South, together with other provinces focusing on the elimination stage of malaria. Where the Malaria program is focused on the control of malaria, the foci approach is not appropriate.

Read more
While log-Gaussian Cox process models are one way to model individual case data, implementing them...

November 2016

Changed the model from LGCP to GAM

While log-Gaussian Cox process models are one way to model individual case data, implementing them in R-INLA can be somewhat fiddly, restricting the ability to automate. Parallel work looking at the use of generalized additive models (GAMs) for this problem showed very good results. Furthermore, GAMs are straightforward to implement and run quickly. We therefore made the decision to swap LGCPs for GAMs. As the DiSARM pipeline has been built in a modular way, swapping out LGCPs for alternative modelling approaches was straightforward.

Read more
The point behind the clicker up was to demonstrate two things; one of them was...

November 2016

Clicker + Dashboard

The point behind the clicker up was to demonstrate two things; one of them was to start with a very simple way to collect coordinates of structures which had been sprayed, as well as demonstrate the real time nature of the mobile app.

Read more
Using log-Gaussian Cox process (LGCP) models, implemented using R-INLA, we showed it was possible to...

November 2016

Automated risk modelling is working

Using log-Gaussian Cox process (LGCP) models, implemented using R-INLA, we showed it was possible to model passive surveillance case data (geolocated back to household and classified as locally acquired) from Swaziland to produce monthly smoothed maps of predicted incidence.

Read more
In Swaziland, the surveillance and decision support platform was utilised to create an automated intervention...

August 2016

Automated creation of risk and structures-to-spray maps for Swaziland IRS 2016

In Swaziland, the surveillance and decision support platform was utilised to create an automated intervention mapping process, which was done by hand in previous years.

Read more
We want to use DiSARM to remove as many of the technical barriers as possible,...

June 2016

Built first version of the 'pipeline'

We want to use DiSARM to remove as many of the technical barriers as possible, making it easy to use data and information that’s already been collected. The automated risk modelling and mapping approach was a really important start, but we needed a way to make that process run regularly on its own, gathering the inputs it needs, and creating the full range of outputs needed, and avoiding the need to have R expertise to run the models.

Read more
The Core interface is designed to let users start to view and explore the outputs...

June 2016

Building the 'Core interface'

The Core interface is designed to let users start to view and explore the outputs from DiSARM’s processing. It initially focuses on the risk maps, and includes some examples of simple analysis.

Read more
After we built some of the initial kind of risk maps for each country. We...

May 2016

Built prototype of mobile, offline risk maps

After we built some of the initial kind of risk maps for each country. We took one static risk map for both Swaziland and Zimbabwe and built a tiny little demonstration app that loads up through a website.

Read more
Link to PBS NewsHour segment

April 2015

DiSARM featured in PBS Newshour segment

Link to PBS NewsHour segment

Read more
Link to The New York Times

January 2015

DiSARM featured in The New York Times

Link to The New York Times

Read more