📊 epidemiology

Epidemiological methods provide target metrics and control parameters for multi-actor violent conflicts

This paper adapts epidemiological methods to model multi-actor violent conflicts in Nigeria as an epidemic process, revealing that effective violence suppression and civilian harm reduction require distinct, actor-specific strategies rather than uniform interventions.

Original authors: Smah, M. L., MacKay, N.

Published 2026-08-10

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Smah, M. L., MacKay, N.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine the world of science as a giant, bustling kitchen where researchers are constantly trying to figure out why things spread. Sometimes, they study how a cold virus jumps from person to person; other times, they look at how a rumor travels through a school or how a fire moves through a forest. This field is called epidemiology, and its superpower is using math to track "contagion." The key idea is simple: for anything to keep spreading, it needs to find new people to infect faster than it can be stopped. If it finds enough new people, it becomes an epidemic; if not, it fizzles out. Scientists have long used these tools to fight diseases, but they are now asking a big, scary question: Can we use the same math to understand something much darker—violent conflict? When armed groups fight over the same people, it's not just a battle of guns; it's a complex web of recruitment, fear, and loyalty that behaves a lot like a virus. Understanding this "infection" is crucial because if we treat a violent conflict like a simple fight between two armies, we might miss the hidden ways it keeps growing, leaving millions of people in danger.

Now, meet two mathematicians who decided to treat a violent conflict like a disease outbreak. Instead of tracking germs, they tracked armed groups and the civilians caught in the middle, using a model inspired by the messy reality of Nigeria. They imagined the conflict as a game of musical chairs where two different "germs" (let's call them Group 1 and Group 2) are trying to infect the same pool of people. Group 1 is like a cult that recruits people by force and ideology, while Group 2 is more like a criminal gang that kidnaps people for ransom money. The researchers built a giant digital simulation to see how these groups interact, compete for "susceptible" civilians, and keep the violence alive.

Their most surprising discovery is that you can't just hit one group with a hammer and expect the whole problem to vanish. In fact, their simulations suggest that trying to stop the violence by only focusing on one group might actually make things worse for the civilians. They found that the conflict has a "reproduction number" (a fancy way of saying how many new fighters a group creates on average). In their model, this number was about 2.5, meaning the violence was self-sustaining and would never stop on its own. To kill the conflict, you'd need to lower that number below 1, but their math showed that no single action—like arresting more people or stopping kidnappings alone—was strong enough to do it. You need a massive, coordinated team effort to tackle recruitment, ideology, and rehabilitation all at once.

Here is the twist that might make your head spin: the things that are best at stopping the spread of the violence aren't always the same things that save the most lives. The researchers created a "Civilian Harm Index" to measure how much suffering the population endures. They found that while arresting fighters is great for reducing the total number of bad guys, it's not the only way to stop the pain. In fact, some common "solutions" like paying ransoms or swapping prisoners actually made the problem worse in the long run. Their simulation showed that paying ransoms is like feeding a fire; it gives the kidnappers money to buy more weapons and recruit more people, leading to more violence later. Similarly, swapping prisoners might seem like a nice humanitarian gesture, but in their model, it just kept the cycle of kidnapping going.

The paper also reveals that these two armed groups are like rival viruses that don't just ignore each other; they actually help each other survive in weird ways. If you suppress one group too hard without helping the other, the suppressed group might just "relapse" and join the other one, or the other group might grow to fill the gap. The model suggests that the violence is so deeply rooted in the population's support and the groups' ability to recruit that it becomes a permanent, "endemic" state—like a cold that never quite goes away—unless you change the entire system.

So, what's the takeaway? The authors aren't saying they have solved the world's conflicts. They are just showing that treating violence like a disease can help us see the hidden rules of the game. Their simulations suggest that if we want to protect civilians, we can't just pick a favorite team to fight against. We have to understand that every action has a ripple effect. Stopping the spread of violence requires a delicate, multi-pronged strategy that tackles the root causes—like why people join these groups in the first place—rather than just trying to shoot the symptoms. It's a reminder that in the chaotic world of conflict, the most obvious solution isn't always the right one, and sometimes, the things we think are helping are actually making the infection worse.

Technical Summary: Epidemiological Methods for Multi-Actor Violent Conflicts

Problem Statement
Contemporary violent conflicts increasingly involve multiple armed actors competing for influence over shared civilian populations, creating complex dynamics that challenge conventional security analysis. Unlike single-insurgency models, these environments feature diverse groups (e.g., ideological insurgents, criminal networks, and militias) with distinct recruitment mechanisms, operational strategies, and objectives. In contexts like Nigeria, these groups interact indirectly by competing for the same susceptible civilian population for recruitment, coercion, and financial extraction. Traditional security approaches often fail to account for these systemic interactions, leading to suboptimal resource allocation where interventions effective against one group may be ineffective or counterproductive when applied uniformly. The core challenge is determining how to allocate limited resources among military, policing, negotiation, and rehabilitation strategies to minimize both violence transmission and humanitarian harm in such multi-actor ecosystems.

Methodology
The authors propose a mathematical framework adapting epidemiological methods to model multi-actor violent conflict as an epidemic process. The study utilizes a multi-compartment dynamical system involving two distinct armed actor groups interacting through competition for a shared civilian population.

Model Structure: The population is divided into compartments including Susceptible civilians ( $S$ ), populations sympathetic to Group 1 ( $P_1$ ) and Group 2 ( $P_2$ ), abducted individuals ( $A_1, A_2$ ), active violent members ( $I_1, I_2$ ), captured/arrested members ( $C_1, C_2$ ), and deradicalized individuals ( $D$ ).
Group Dynamics:
- Group 1 (Ideological): Relies on coercive and ideological recruitment, including the radicalization of abducted individuals. It does not engage in ransom negotiations.
- Group 2 (Financial/Criminal): Engages in ransom negotiations and prisoner exchanges. Recruitment is driven by financial vulnerability and criminal opportunism. Ransom success is modeled to amplify sympathy for this group.
Key Metrics:
- Basic Insecurity Reproduction Number ( $R_0$ ): Derived using the next-generation matrix framework to determine the threshold for epidemic (exponentially growing) violence.
- Civilian Harm Index (CHI): A novel metric quantifying the instantaneous and cumulative humanitarian impact, combining mortality, abduction, forced radicalization, and ransom costs.
Analysis Techniques: The study employs equilibrium analysis (identifying Violence-Free and Persistent-Violence equilibria), threshold analysis, numerical simulation (using R and the deSolve package), and sensitivity analysis (Latin Hypercube Sampling and Partial Rank Correlation Coefficients) to evaluate intervention efficacy.

Key Contributions

Framework Adaptation: The paper establishes a "proof of concept" for applying standard epidemiological compartmental models to multi-actor conflict, moving beyond single-actor or homogeneous conflict models.
Novel Metrics: The introduction of the Civilian Harm Index (CHI) allows for the quantification of humanitarian impact distinct from transmission dynamics, enabling the evaluation of trade-offs between suppressing violence and protecting civilians.
Decomposition of $R_0$ : The model decomposes the reproduction number into within-group and cross-group transmission pathways, revealing that within-group reproduction (driven by sympathizers) is the dominant factor in sustaining violence.
Intervention Trade-offs: The framework demonstrates that interventions optimizing for transmission control (reducing $R_0$ ) do not necessarily align with those minimizing civilian harm (CHI).

Results

Persistence of Violence: Under baseline parameters, the system converges to a Persistent Violence Equilibrium (PVE) with $R_0 \approx 2.5$ . The PVE is locally stable, characterized by damped oscillatory behavior, indicating that temporary reductions in violence are insufficient to eliminate insecurity without coordinated structural changes.
Single-Parameter Limitations: Systematic single-parameter screening revealed that no single intervention (e.g., increasing arrests, reducing recruitment, or enhancing deradicalization) is sufficient to reduce $R_0$ below 1. Coordinated intervention across multiple mechanisms is required to break the persistence of violence.
Sensitivity Analysis:
- Drivers of $R_0$ : The most influential parameters for sustaining violence are voluntary recruitment rates ( $\beta_1, \beta_2$ ) and the generation of sympathizers ( $\nu_1$ ).
- Drivers of Civilian Harm: Abduction rates ( $\alpha_1$ ) and the radicalization of abducted individuals are the primary drivers of cumulative civilian harm.
- Divergent Efficacy: Parameters that most effectively reduce $R_0$ (e.g., arrests and military elimination) are not identical to those that most effectively reduce CHI.
Counterintuitive Findings:
- Negotiation and Ransom: Increasing ransom payments and prisoner exchange rates was found to increase both violence persistence and civilian harm. Specifically, prisoner exchange was shown to have the highest negative impact on reducing harm (i.e., it increases harm the most), while ransom payments create financial incentives that sustain future kidnapping operations.
- Group Specificity: Interventions effective against one group (e.g., targeting Group 1's recruitment) may be ineffective or counterproductive if applied uniformly, as they fail to address the distinct drivers of the other group (e.g., Group 2's financial incentives).

Significance and Claims
The authors claim that epidemiological methods provide a quantitative framework for evaluating intervention priorities and trade-offs in complex multi-actor conflicts. The study argues that treating multi-actor violence as an endemic disease allows for the identification of system-level thresholds and the optimization of distinct criteria for epidemic control versus humanitarian protection.

Crucially, the paper positions itself as an illustrative framework rather than a predictive model. The authors explicitly state that the results are not quantitative forecasts of specific conflicts (such as the situation in Nigeria) but serve to demonstrate the utility of the modeling approach. The significance lies in shifting policy discussions from actor-specific suppression to system-level intervention frameworks that account for cross-group dynamics, recruitment mechanisms, and the unintended consequences of negotiation strategies. The model suggests that eliminating insecurity requires coordinated interventions across the entire "violence ecosystem" rather than isolated efforts against single actors.

Technical Summary: Epidemiological Methods for Multi-Actor Violent Conflicts

More like this