Mathematical modeling of vector-borne infectious diseases has demonstrated that the feeding preferences of insect vectors can fundamentally reshape disease transmission dynamics and the effectiveness of threshold-triggered public health interventions. Researchers Ying Wang, Weike Zhou, and Xia Wang investigated these interactions by constructing and analyzing a nonsmooth Filippov dynamical system that incorporates both vector biting preference and a joint host-vector threshold control policy. Their theoretical calculations and numerical simulations show that when vectors preferentially select susceptible hosts over infected hosts, the epidemic system becomes highly sensitive to intervention triggers, generates complex backward bifurcations, and produces multistability regimes where diseases can persist even when the standard basic reproduction number falls below one.
Recent empirical and biological research has confirmed that vectors do not bite hosts uniformly at random across populations. Concurrently, plant pathology and ecological studies have revealed reverse patterns, in which vectors selectively prefer healthy or susceptible hosts depending on physiological condition and environmental context. While continuous and impulsive mathematical models have historically been applied to epidemic control, continuous frameworks assume uninterrupted resource deployment that risks substantial logistical waste, whereas impulsive models enforce predetermined, instantaneous intervention schedules that cannot capture state-dependent switches triggered by real-time outbreak surveillance.
To overcome these limitations, the authors developed a two-dimensional reduced compartmental model derived from a four-dimensional system tracking susceptible hosts, infectious hosts, susceptible vectors, and infectious vectors under constant total population sizes. Heterogeneous feeding was formalized by defining biting probabilities, where vectors bite infected hosts with probability p and susceptible hosts with probability f, yielding normalized preference parameter p̃ defined as p divided by the sum of p and f. When p̃ equals one-half, feeding occurs without preference; values above one-half denote a preference for infected hosts, while values below one-half indicate a preference for susceptible hosts. The system switches discontinuously between two differential subsystems: an unmanaged subsystem operating below an intervention threshold and an active management subsystem that applies host treatment at rate c_I and vector culling at rate c_V once a joint infection pressure function exceeds a critical threshold C_T. This switching boundary constitutes a nonsmooth manifold separating the dynamical space into distinct operational domains.
By conducting geometric and algebraic analyses of the switching surface, the investigators classified the system’s sliding regions, regular equilibria, pseudo-equilibria, boundary equilibria, and tangent points. Lie derivative formulations demonstrated that repelling sliding motion cannot occur within the interior domain, ensuring the forward uniqueness of trajectories. The sliding motion along the discontinuity manifold is governed by a derived one-dimensional cubic differential equation whose real roots determine the existence and stability of admissible pseudo-equilibria. Depending on the sign configuration of the cubic coefficients and the values of the intervention parameters, the system can exhibit zero, one, or up to two coexisting sliding segments, and up to three admissible pseudo-equilibria, providing a formal mathematical mechanism through which populations can stabilize precisely along the predetermined control threshold.
A critical revelation of the study is the dramatic divergence in bifurcation structures driven by vector feeding preferences. When vectors exhibit no preference, both subsystems display forward bifurcations at a basic reproduction number of one. Under conditions of no preference, the disease-free equilibrium remains globally asymptotically stable whenever the reproduction number is below unity. However, when vectors preferentially feed on susceptible hosts such that p̃ falls below a critical threshold dictated by vector mortality, transmission, and removal parameters, both the uncontrolled and controlled subsystems exhibit backward bifurcations. In this backward bifurcation regime, stable endemic equilibria coexist with the disease-free state even when the basic reproduction number is strictly less than one.
The study further mapped the discontinuity-induced bifurcations that occur as public health officials adjust the joint control threshold C_T. In regimes where vectors prefer infected hosts or feed neutrally, raising the threshold triggers at most two boundary equilibrium bifurcations, during which regular equilibria collide with the switching surface and tangent points to transition smoothly into pseudo-equilibria. In stark contrast, when vectors preferentially bite susceptible hosts, the system undergoes up to four successive boundary equilibrium bifurcations. These transitions cause sliding segments to split, merge, and vanish, generating thirteen distinct bistable configurations and a remarkable tristable regime. In this tristable state, an admissible pseudo-equilibrium, three real equilibria, and a disease-free equilibrium coexist simultaneously.
Through rigorous application of the Bendixson-Dulac criterion and topological trajectory tracking across the switching boundary, the authors proved the complete nonexistence of limit cycles within the model. Neither regular limit cycles entirely contained within either subsystem nor discontinuous cycles traversing the sliding segments or surrounding the switching manifold can arise. Because periodic oscillations are mathematically excluded, every system trajectory must asymptotically converge toward one of the identified stable equilibria or pseudo-equilibria. Numerical simulations parameterized with baseline epidemiological values demonstrated that the global phase portraits follow predictable attractor sequences as C_T increases. For neutral preferences, raising C_T shifts the global attractor from a stable controlled node to an attractive pseudo-equilibrium on the threshold boundary, and finally to an uncontrolled endemic node. When susceptible host preference is introduced, the sliding region expands substantially, dramatically increasing the parameter window over which the system enters sliding motion but making the terminal state extraordinarily sensitive to microscopic shifts in threshold values.
From an applied public health standpoint, these analytical findings deliver essential insights for designing targeted vector-control campaigns. Furthermore, the joint threshold formulation demonstrates that tracking composite risk indicators combining host illness and vector infection rates provides a more robust trigger for intervention than monitoring human clinical cases alone. The authors conclude that incorporating vector behavioral ecology into nonsmooth dynamical frameworks is indispensable for calibrating disease management, emphasizing that future modeling must integrate demographic recruitment and environmental seasonality to refine practical threshold policies.
Beyond general qualitative shifts, the authors mathematically defined the exact threshold governing the bifurcation structure within each subsystem. According to the team's derivations, backward bifurcation occurs in a subsystem when the normalized biting preference parameter falls below the vector mortality and culling rate divided by the sum of three times this combined rate and the transmission rate to vectors. While the algebraic sliding equations theoretically allow up to three pseudo-equilibria and sixteen distinct sliding-region configurations, the researchers found that a coexisting triple pseudo-equilibrium never materializes within the biologically feasible unit domain. Instead, the third root consistently falls outside the interval of valid host infection proportions, leaving only single or paired admissible pseudo-equilibria active during practical sliding motion.
Geometric tracking of the discontinuity boundary further revealed how sliding segments interact with tangent points. In each subsystem, two tangent points establish the boundaries where trajectories graze the switching surface. As the joint intervention threshold shifts, crossing windows bounded by these tangent points open or vanish, prompting sliding segments to split or coalesce through double tangency bifurcations. The team verified these transitions using baseline simulation parameters, setting the vector mortality rate to 13.825, host recovery rate to 2.0278, natural host mortality to 0.1429, vector-to-host density ratio to 1.03, and vector-host transmission rates to 8 and 18.
The authors noted several structural limitations, emphasizing that their parameters were investigated through systematic numerical sweeps rather than direct empirical fitting to field datasets. The formulation assumes constant population totals without disease-induced death, periodic seasonal forcing, or spatial migration, factors that could underestimate transmission intensity across interconnected regions. Nonetheless, the findings point toward novel intervention strategies.
Subject of Research: Mathematical epidemiology and nonsmooth dynamical modeling of vector-borne disease control
Article Title: Bifurcations and dynamics of a Filippov epidemic model with joint threshold control policy and vector preference
Article References: Wang, Y., Zhou, W., & Wang, X. (2026). Bifurcations and dynamics of a Filippov epidemic model with joint threshold control policy and vector preference. Advances in Continuous and Discrete Models, 2026(1), Article 75. https://doi.org/10.1186/s13662-026-04114-y
Image Credits: AI Generated
DOI: 10.1186/s13662-026-04114-y
Keywords: epidemiology, vector-borne disease, Filippov system, bifurcation analysis, mathematical biology, disease control, vector preference, threshold policy, Bifurcations, dynamics, Filippov, epidemic
Cite Scienmag News
Kristina Jarvis. (November 9, 2026). Vector Feeding Preferences Dramatically Alter Epidemic Outbreak Dynamics and Control Outcomes. Scienmag. https://scienmag.com/vector-feeding-preferences-dramatically-alter-epidemic-outbreak-dynamics-and-control-outcomes/
Kristina Jarvis. "Vector Feeding Preferences Dramatically Alter Epidemic Outbreak Dynamics and Control Outcomes." Scienmag, 9 November 2026, https://scienmag.com/vector-feeding-preferences-dramatically-alter-epidemic-outbreak-dynamics-and-control-outcomes/. Accessed 21 November 2026.
Kristina Jarvis. "Vector Feeding Preferences Dramatically Alter Epidemic Outbreak Dynamics and Control Outcomes." Scienmag. November 9, 2026. https://scienmag.com/vector-feeding-preferences-dramatically-alter-epidemic-outbreak-dynamics-and-control-outcomes/








