[Malaysia] Understanding and managing the risk of water related diseases under hydrometeorological extremes

Lead Research Organisation: Imperial College London
Department Name: Civil & Environmental Engineering


Globally, water-related diseases are a major obstacle to sustainable development (WHO, 2018). Many of these diseases, such as Cholera and Hepatitis A, have been successfully phased out in Malaysia. However, leptospirosis and malaria still affect Malaysians every year. The annual incidence rate of Leptospirosis is actually increasing, from 0.97 cases per 100,000 population in 2004 to 12.47 per 100,000 in 2012.

It is well known that leptospirosis and malaria are strongly linked to environmental conditions, and humidity and temperature in particular. Although scientific understanding of this link is advancing at a rapid pace, it is still very difficult to build computational models that make quantitative forecasts of outbreaks. Yet such systems are indispensable for proactive disease management, and to optimise the allocation of resources for medical prevention and interventions.

A major difficulty with predicting outbreaks of water-related diseases is the large number of driving factors, which span the environmental and socio-economic realms. Additionally, many of the processes that link the driving factors with disease outbreaks, are highly non-linear and difficult to represent in computational algorithms.

This proposal therefore sets out to explore the use of artificial intelligence approaches to identify and model the physical and microbiological interactions that lead to conditions favouring disease occurrences, with the goal of developing an early warning system for disease outbreaks. The complexity and non-linearity in the processes makes AI methods such as the neural network approach highly promising as it is inherently suited to problems that are mathematically difficult to describe and highly non-linear.

The scientific field of artificial intelligence is developing at a very rapid pace. This evolution is driven by the exponentially increasing amount of information available online (often referred to as the "big data" era), much of which is highly unstructured and diverse (e.g., data from social media such as twitter feeds and news posts). This has resulted in the development of many novel and powerful algorithms and routines. However, its exploration in the context of water-related diseases is still very limited. Therefore, we propose to leverage these breakthroughs, by testing and adapting these new methodologies to advance predictive modelling of the link between hydrometeorological extremes and water-related diseases. The proposed research combines extensive compilation, synthesis and integration of socio-demographic and infrastructural data alongside data of environmental extremes, with novel computational algorithms to "learn" from the datasets and leverage the outcomes to improve operational forecasting systems.

We have assembled a world-leading consortium of scientists that combines expertise on hydrometeorological extremes, artificial intelligence and community health issues. We will use the Malaysian state of Negeri Sembilan as a case study, and will work in close collaboration with the State Department of Health. This will allow us to access historical records that include patients' demographic information. More recently, risk assessment have been conducted using questionnaires that includes assessment of water supply and drainage infrastructure. The epidemiological data will be complemented by environmental data from the Department of Meteorology and the Department of Irrigation and Drainage (which are either available for academic use for free or a small fee), and monthly water quality monitoring data from local District Offices.


WHO, 2018. http://www.who.int/water_sanitation_health/diseases-risks/diseases/diarrhoea

Planned Impact

The proposed project arose from one of the major research priorities identified during a stakeholder engagement session as part of a workshop in Belum Forest in August 2017. The workshop was funded by a Newton Fund Researcher Links grant awarded to Dr Zulkafli and colleagues. As a result, the proposed research is strongly demand driven.

In order to maximise the impact of the proposed research, we will work closely with the crucial support from the State Department of Health, Negeri Sembilan. To ensure proper end-user buy-in and compatibility with operational procedures, a meeting was held on 8 February 2018 with the Head of Communicable Disease Control Unit. The current proposal is written in consultation with medical doctors in charge of surveillance and response to outbreaks.

Our pathways to impact strategy is based on 4 pillars:

" A strong engagement with our project partners

The Malaysian team will hold frequent follow-up meetings throughout the project with the State Department of Health, and the Communicable Disease Control Unit as well as Vector-borne Disease Control Unit in particular. Through this partner, we will further extend our network of potential end-users, for instance by reaching out to other states within Malaysia, and potentially in the rest of South East Asia.

" Usability and availability of project outcomes

The proposed research is very applied and will create outputs that are usable in an operational setting, in particular the forecasting model. Throughout the development of this model, we will work closely with potential end-users of these outputs to ensure the compatibility with existing systems and maximise the usability. For instance, we refrain from using commercial software solutions, but instead rely upon open-source modelling frameworks, such as the very extensive set of AI toolboxes available in the R statistical modelling environment.

" Engagement of the private sector

The investigator team has a track record of working with the private sector. We aim to engage private companies, and SMEs in particular, from the start of the project to maximise technology transfer. We identify two specific technologies that may be relevant: (1) disease forecasting services; and (2) low-cost sensing. For the former, we aim to work with SMEs in the sector of weather services. For the low-cost sensors we aim to connect to the thriving innovation community in Malaysia. For instance, Buytaert's group already sources components of the sensor platform from small company based in Kuala Lumpur (Rocketscream.com). Specific activities may include training sessions, and active participation in virtual innovation networks such as the Arduino and Internet of Things online forums.

" Engagement with the international disaster risk reduction community

The UK team in particular, has excellent contacts with the international disaster risk reduction community, for instance through their engagement with the UNESCO Sendai Framework for Disaster Risk Reduction (UNISDR) as part of their involvement in the NERC programme on "Science for Humanitarian Emergencies and Resilience" (SHEAR). Relevant activities include attendance to UNESCO Knowledge Fora (most recently in October 2017), production of policy briefs, organisation of conference sessions in policy-oriented events such as the World Water Forum.

Further details on these activities and their context is provided in the "Pathways to Impact" attachment.


10 25 50
Description Ongoing collaboration with project partner university UPM (Universiti Putra Malaysia) 
Organisation Putra Malaysia University
Country Malaysia 
Sector Academic/University 
PI Contribution - we perform hydrological monitoring (using low-cost sensors developed in our group at Imperial College) and modelling - we play a mentoring role for the Malaysian PhD students - we will lead on publications from the project that involve hydrological modelling - we are organising the annual project meeting in the summer of 2020
Collaborator Contribution - UPM have two PhD students working on this project, who are performing spatial and temporal modelling, predictive modelling, and secondary and primary (survey) data collection. - UPM have organised the annual project meeting in the summer of 2019.
Impact The collaboration is fruitful but there are no finished outputs yet. The collaboration is multi-disciplinary, involving human behaviour, hydrology, machine learning, spatial analytics, and epidemiology.
Start Year 2019
Description Machine learning training to JKNSS staff 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Professional Practitioners
Results and Impact The UPM researchers conducted a half-day machine learning training to staff at JKNSS, the health department of Negeri Sembilan state, who are project stakeholders. Negeri Sembilan state is our case study.
Year(s) Of Engagement Activity 2019
Description Progress update presentation to the Malaysian Ministry of Education 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact Comprehensive progress update to the Malaysian Ministry of Education by the UPM researchers.
Year(s) Of Engagement Activity 2020