Перейти к основному содержанию

Mapping WASH-related disease risk: A review of risk concepts and methods

Страны
Мир
Источники
UNU
Дата публикации
Происхождение
Просмотреть оригинал

Daudelin, F., Taing, L., Chen, L., Lopes, C., Adeniyi, F.A., Mehmood, H., 2021. Mapping WASH-related disease risk: A review of risk concepts and methods, Issue 12. United Nations University Institute for Water, Environment and Health, Hamilton, Canada.

The report provides a review of how risk is conceived of, modelled, and mapped in studies of infectious water, sanitation, and hygiene (WASH) related diseases. It focuses on spatial epidemiology of cholera, malaria and dengue to offer recommendations for the field of WASH-related disease risk mapping.

The report notes a lack of consensus on the definition of disease risk in the literature, which limits the interpretability of the resulting analyses and could affect the quality of the design and direction of public health interventions. In addition, existing risk frameworks that consider disease incidence separately from community vulnerability have conceptual overlap in their components and conflate the probability and severity of disease risk into a single component.

The report identifies four methods used to develop risk maps, i) observational, ii) index-based, iii) associative modelling and iv) mechanistic modelling. Observational methods are limited by a lack of historical data sets and their assumption that historical outcomes are representative of current and future risks.

The more general index-based methods offer a highly flexible approach based on observed and modelled risks and can be used for partially qualitative or difficult-to-measure indicators, such as socioeconomic vulnerability. For multidimensional risk measures, indices representing different dimensions can be aggregated to form a composite index or be considered jointly without aggregation. The latter approach can distinguish between different types of disease risk such as outbreaks of high frequency/low intensity and low frequency/high intensity.

Associative models, including machine learning and artificial intelligence (AI), are commonly used to measure current risk, future risk (short-term for early warning systems) or risk in areas with low data availability, but concerns about bias, privacy, trust, and accountability in algorithms can limit their application. In addition, they typically do not account for gender and demographic variables that allow risk analyses for different vulnerable groups.

As an alternative, mechanistic models can be used for similar purposes as well as to create spatial measures of disease transmission efficiency or to model risk outcomes from hypothetical scenarios. Mechanistic models, however, are limited by their inability to capture locally specific transmission dynamics.

The report recommends that future WASH-related disease risk mapping research:

  • Conceptualise risk as a function of the probability and severity of a disease risk event. Probability and severity can be disaggregated into sub-components. For outbreak-prone diseases, probability can be represented by a likelihood component while severity can be disaggregated into transmission and sensitivity sub-components, where sensitivity represents factors affecting health and socioeconomic outcomes of infection.
  • Employ jointly considered unaggregated indices to map multidimensional risk. Individual indices representing multiple dimensions of risk should be developed using a range of methods to take advantage of their
    relative strengths.
  • Develop and apply collaborative approaches with public health officials, development organizations and relevant stakeholders to identify appropriate interventions and priority levels for different types of risk, while ensuring the needs and values of users are met in an ethical and socially responsible manner.
  • Enhance identification of vulnerable populations by further disaggregating risk estimates and accounting for demographic and behavioural variables and using novel data sources such as big data and citizen science.

This review is the first to focus solely on WASH-related disease risk mapping and modelling. The recommendations can be used as a guide for developing spatial epidemiology models in tandem with public health officials and to help detect and develop tailored responses to WASH-related disease outbreaks that meet the needs of vulnerable populations. The report’s main target audience is modellers, public health authorities and partners responsible for co-designing and implementing multi-sectoral health interventions, with a particular emphasis on facilitating the integration of health and WASH services delivery contributing to Sustainable Development Goals (SDG) 3 (good health and well-being) and 6 (clean water and sanitation).