Complex systems simulations for intervention development: Human trafficking and conflict-related violence

Lead Research Organisation: University College London
Department Name: Institute for Global Health


Overview & Aim.
This project will develop interdisciplinary research methods at the intersection of complex systems, violence, public health interventions, and data visualisation. We aim to evaluate the use of dynamic complex systems modeling to develop interventions for hard-to-reach populations affected by human trafficking and conflict-related violence. The project will address the following questions: What contributions can these novel methods make to the global response to human trafficking and conficlt-related violence? How can evidence-based, accessible and visually powerful complex systems models inform decision-makers working on interventions?

Project Justification.
Crises, population mobility, violence and exploitation cause deleterious health consequences for millions of individuals, particularly for refugees, low-wage migrant workers and victims of human trafficking [4, 8-10]. Currently, there is limited evidence on effective interventions for these marginalised groups. Scholars emphasise the causal complexity of violence and exploitation, however, most analytical approaches have erroneously assumed linear, average effects of exposures on single outcomes. Experimental approaches to test interventions are often unethical or unfeasible to implement among these hard-to-reach populations. Reductionist research methods, such as traditional epidemiological studies, dangerously oversimplify complex problems, which frequently results in wasted resources and potential harm to already vulnerable groups.

We will invite experts in complexity science, violence research, public health and data visualisation to join the Complexity & Violence Research Network. This Network will jointly design a complex intervention case study that will serve as an evaluated 'proof of concept' for these methods. In brief, our methods will include the following steps:

1. Literature reviews on: 1) complex systems modelling for intervention development, 2) visualising complex systems, and 3) theoretical and epistemological underpinnings of complex systems methods.
2. Joint selection and design of an intervention case study using the Network's existent research on violence and other historical data, empirical data, and theory.
3. Calibrate, visualise and iterate a complex intervention model with dynamic multi-level interactions of exposures, associated harms, and simulated counterfactuals (e.g. intervention).
4. Collaboratively validate the model and evaluate the usefulness, feasibility and ethics of using complex systems methods for intervention development. Evaluation will include Network and Stakeholder reflections and critiques during small in-person meetings and a virtual two-wave Delphi Panel of funders and intervention developers to assess overall methods and usefulness.
At every phase, we will critically examine and document the methodological and epistemological contributions, short-comings and risks, and best practice for adopting these methods in interdisciplinary research groups.

To contribute to our future work and advancements in the field, we will produce:

-Academic papers on: 1) the case study findings; and 2) methodological recommendations
-Lay guidance on the opportunities and limitations of these methods, with the aim of supporting intervention development, evaluation and policy decisions targeting populations affected by violence, conflict and human trafficking.
-A web browser accessible version of the final case study model and visualisations to allow a wide user-audience to engage and play with the model inputs to understand how incorporating dynamic and complex characteristics of a system can explain causal mechanisms and potential effects of interventions.
-Articulate future research priorities and identify funding opportunities to pursue longer-term sustainability of the Network and the development of future proposals.


10 25 50