{"id":"be53c2af-e7e4-43a7-9a69-b784a2659fd8","arxiv_id":"1908.02434","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"An age-structured host and early-life-stage vector dengue model yields Rv and R0 thresholds governing disease-free and endemic stability.","lead":"This paper builds a dengue transmission model that combines an early-life mosquito stage with age structure in the human population, and it derives threshold conditions for disease extinction or persistence. The value is a sharper modeling framework for age-targeted vector control, though no real epidemiological data are used.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.4's global asymptotic stability claim is not valid when multiple stable vector demographic equilibria exist; the disease-free equilibrium is not unique, so no single equilibrium is globally asymptotically stable.","rationale":"The reader's weakest_assumption identifies a genuine gap in the proof of Theorem 3.4: the relation in Eq. (9) is asserted for limit-superior values without adequate justification. However, I regard that gap as secondary to a more fundamental issue: even if the limsup step were repaired, the theorem's conclusion—global asymptotic stability of a single disease-free equilibrium—is not true under the paper's own assumptions. The model deliberately allows multiple vector demographic steady states (Lemma 3.1, Remark 3.2, Section 4.2), and each stable vector state gives a distinct disease-free equilibrium of the full system. With R0 < 1, all such equilibria are locally stable, so no single one can be globally asymptotically stable. This is not a matter of consensus or external criticism; it follows from the paper's own definitions. The authors should either add a uniqueness assumption on the vector demographic steady state or restate the theorem as global attractiveness of the disease-free set. The reader's conditional verdict remains appropriate: the paper is not acceptable as is, but the problem is fixable by a careful restatement and a rigorous proof. Hence I do not change the verdict, but I want to redirect the emphasis from the technical limsup issue to the structural non-uniqueness that makes the headline claim false as written.","tokens_in":10524,"tokens_out":16572,"duration_ms":175143,"concrete_test":"Analytical check: take g(Nv) = r e^(-Nv/c1)(sin(c2 Nv) + 1) as in Eq. (12) with parameters yielding at least two roots of g(Nv) = phi where g'(Nv) < 0 (as in Figure 1). Verify that both corresponding vector equilibria (E*, Nv*) from Eq. (5) are locally asymptotically stable and that each, combined with Iv = 0 and ih(a) = 0, is a locally stable disease-free equilibrium of System (3). Then, with R0 < 1 (for instance by taking beta small), there are two distinct asymptotically stable disease-free equilibria. By definition of global asymptotic stability, neither equilibrium can attract all initial conditions, so Theorem 3.4 as stated is false; the correct conclusion is global attractivity of the disease-free set.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Theorem 3.4) asserts global asymptotic stability of 'the' disease-free solution when R0 < 1. But the paper explicitly admits multiple vector demographic steady states: Lemma 3.1 and Remark 3.2 show that every root Nv* of g(Nv) = phi with g'(Nv*) < 0 yields a locally stable vector equilibrium, and Section 4.2, Eq. (12) with Figure 1 exhibits eight such roots (four stable). For each such Nv*, the full system has a disease-free equilibrium (E*,Sv*,Iv*) = ((mu_v/delta) Nv*, Nv*, 0) with sh* = 1 and ih* = 0. Since the R0 formula in Eq. (7) does not depend on Nv*, all these disease-free equilibria have the same R0. If R0 < 1 and at least two such stable roots exist, then there are at least two locally stable disease-free equilibria. A necessary condition for an equilibrium to be globally asymptotically stable is that it be the unique asymptotically stable equilibrium; with two stable equilibria, neither can attract all solutions. Thus Theorem 3.4 as stated is false. Even if the proof's limsup step leading to Eq. (9) were justified, the argument only attempts to show Iv(t) -> 0; this establishes global attractivity to the disease-free set, not convergence to a single equilibrium. The theorem must be qualified: either assume a unique stable vector demographic state, or restate the result as global attractivity of the set of disease-free equilibria.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs a dengue transmission model with an early-life vector stage (E) and adult vectors (Sv, Iv) coupled to an age-structured human host population (Sh, Ih, Rh). The early-life recruitment is modeled through a Kolmogorov-type function f(Nv)=Nv g(Nv), allowing multiple vector demographic steady states. The authors define a vector demographic number Rv and a basic reproductive number R0, prove local stability results for vector equilibria, and state global stability results for the infection-free state when R0<1 and existence/stability of an endemic state when R0>1. The paper closes with an upwind finite-difference scheme and numerical examples, including logistic growth, multiple vector steady states, and seasonality.","tokens_in":10832,"tokens_out":6539,"duration_ms":78557,"significance":"If the results were fully rigorous, the model would be a useful contribution to vector-borne epidemic modeling: the R0 formula in Eq. (7) explicitly incorporates host age structure through p∞(a), β(a), and γ(a), and the vector demographic threshold Rv is consistently derived from the early-life-stage submodel. The numerical section explores a genuinely relevant phenomenon, multiple vector equilibria, and the authors are explicit about the dependence of outcomes on initial vector densities. However, the central global-stability theorem is overstated in the presence of multiple stable vector demographic equilibria, and the proofs of Theorems 3.4 and 3.5 contain significant gaps. These issues are fixable but affect the paper's principal claims.","major_comments":[{"comment":"Theorem 3.4 asserts that when R0<1 the disease-free solution of System (3) is globally asymptotically stable. The system, however, does not have a unique disease-free equilibrium. Lemma 3.1 and Remark 3.2 show that every root N_v of g(N_v)=φ with g'(N_v)<0 yields a locally stable vector demographic state, and Section 4.2 with Eq. (12) and Figure 1 exhibits four such stable roots. For each such N_v, the full system has a disease-free equilibrium (E*,S_v*,I_v*,s_h*,i_h*,r_h*)=((µ_v/δ)N_v,N_v,0,1,0,0), and R0 in Eq. (7) is independent of N_v. If R0<1 and two stable vector states coexist, there are at least two asymptotically stable disease-free equilibria, and a necessary condition for global asymptotic stability of an equilibrium is that it be the unique asymptotically stable equilibrium. Thus Theorem 3.4 as stated is false. The theorem should be qualified, for instance by assuming a unique globally stable vector demographic state, or restated as global attractivity of the set of disease-free equilibria.","section":"Theorem 3.4; Section 3.2; Section 4.2"},{"comment":"The limiting argument in the proof of Theorem 3.4 is not justified. The proof defines limiting quantities B*, I_v*, S_v*, N_v*, then asserts Eq. (8) by replacing p(t,a) with p∞(a) and the ratio Iv(τ+t-a)/Nv(τ+t-a) with I_v*/N_v* inside the integral, and Eq. (9) by treating these limits as if they satisfied the equilibrium relation B*=(µ_v/β_v)I_v*/S_v*. A rigorous proof would need a lemma showing that any sequence t_n with B(t_n)→B* also gives convergence of the relevant shifted vector quantities to a steady state, together with a dominated-convergence argument. In addition, the characteristic formula preceding Eq. (8) is stated only for t>a, yet it is integrated over all a≥0; the contribution from the initial-data part (a≥t) is not controlled. Without these steps, the conclusion I_v*=0 is not established by the argument given.","section":"Theorem 3.4 proof, Eq. (8) and Eq. (9)"},{"comment":"Theorem 3.5 claims that if R0>1 and a stable vector demographic state exists, then there exists one endemic non-uniform stable steady state. The proof establishes only that the fixed-point equation H(B)=B has a positive solution: it defines G(B)=H(B)/B, argues that G(0)=R0>1 and G(1)<1, and concludes existence. It does not analyze the eigenvalues of the full linearization at the constructed endemic state, nor does it otherwise show that this state is locally or globally stable. The assertion that G(B) is strictly decreasing is also stated without proof. The stability part of the theorem is therefore unsupported. Moreover, if multiple stable vector demographic states N exist, applying the argument separately to each N yields multiple candidate endemic states, so the uniqueness claim needs a more precise statement.","section":"Theorem 3.5"}],"minor_comments":[{"comment":"The boundary condition is written as Iv(t,0)=0, but it should be Ih(t,0)=0 for the infected host age class; as printed, the notation is inconsistent.","section":"Section 2, System (1)"},{"comment":"The update for (Sv)^{j+1} is printed as (Sv)^{j+1}=E^j+Δt[...], which is dimensionally inconsistent; it should read (Sv)^{j+1}=(Sv)^j+Δt[...].","section":"Section 4, numerical scheme"},{"comment":"The integral in Eq. (8) is missing the differential dτ in the displayed formula, which makes the expression harder to parse.","section":"Eq. (8)"},{"comment":"Reference [11] lacks complete bibliographic details for the proceedings volume and page range, and reference [9] contains a typo in the word 'Vertical'; these should be corrected in the final version.","section":"References"},{"comment":"The discussion states that 'the local and global asymptotic stability of the disease-free equilibrium was determined when R0<1', but in view of the multiple disease-free equilibria discussed in the major comments, the phrase 'the disease-free equilibrium' should be made precise regarding which equilibrium and under which uniqueness assumption.","section":"Section 5, Discussion"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the pairing of early-life vector dynamics (with a Kolmogorov egg-production function that can give multiple vector steady states) with age-structured hosts, plus the resulting Rv and R0 formulas. That combination is not in the existing literature, and the R0 expression is a clean, self-contained quantity: it depends on host age through the survival/contact/recovery kernels and is independent of the vector equilibrium. The vector demographic threshold analysis is standard and correct. For a modeling-framework paper, this is a solid building block.\n\nThe soft spots are real and, in one case, load-bearing. Theorem 3.4 claims global asymptotic stability of \"the\" disease-free solution when R0 < 1. But the paper itself admits multiple locally stable vector demographic states (Section 4.2, Fig. 1). For each such state there is a disease-free equilibrium with the same R0, since R0 does not depend on Nv*. If two such states exist and are stable, no single disease-free equilibrium can be globally asymptotically stable. The proof only shows that every omega-limit point has Iv* = 0, which is attractivity to the disease-free set, not convergence to one equilibrium. The reader's weakest-assumption flag is exactly right. This needs a qualifier: assume a unique stable vector demographic state, or restate the result as global attractivity of the union of disease-free equilibria.\n\nThe other gaps are smaller. Theorem 3.5's uniqueness claim relies on \"G(B) is strictly decreasing,\" which is asserted without proof; the existence part is fine, but the uniqueness is not established. There is also a visible misprint in the numeric scheme: the update for Sv should start with (Sv)^j, not E^j. And the paper does not fit any data, so its value is as a framework, not a validated public-health tool. The simulations do illustrate the local stability claims and the multiple-vector-state behavior, and they match the math where the math is sound.\n\nOn balance: the central R0 threshold and the demographic-bifurcation analysis are worth keeping, and the combination of age-structure and early-life vector dynamics is worth taking seriously. The global stability theorem needs to be fixed or qualified before the paper's main claims are accepted. I would send this to peer review, but with a major-revision requirement and a referee instructed to focus on Theorem 3.4 and on the missing monotonicity proof for Theorem 3.5.","headline":"Useful modeling combination, but the flagship global stability theorem overreaches; the rest is a competent extension that deserves a revision.","tokens_in":11356,"tokens_out":1681,"would_cite":false,"duration_ms":22672,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["92D30","92D25"],"pacs":[],"model":"deepseek-v4-flash","headline":"A single threshold number, $R_0$, governs whether dengue dies out in an age-structured human population with mosquito egg dynamics.","keywords":["dengue","age-structured host population","early-life vector stage","basic reproductive number","vector demographic threshold","global stability","endemic steady state","seasonality"],"falsifier":"Run System (3) numerically with parameters giving $R_0<1$ but a stable positive vector steady state; if the infected-host density $i_h(t,a)$ converges to a positive steady state from some initial condition, Theorem 3.4 is false. A simpler check is to track the ratio $B(t)/(I_v(t)/S_v(t))$ along transients and see whether it approaches $\\mu_v/\\beta_v$ as required by Eq. (9).","tokens_in":10294,"feed_emoji":"🦟","tokens_out":10378,"duration_ms":101185,"temperature":0.7,"pith_summary":"This paper builds a dengue transmission model that tracks humans by age and mosquitoes through a juvenile egg/larvae stage plus adult susceptible and infected classes. It derives a basic reproductive number $R_0$ that depends on age-specific contact, recovery, and host survival, as well as on the mosquito-to-human transmission rate. The main results are threshold theorems: when $R_0<1$, the infection-free state is globally stable, so the disease dies out, and when $R_0>1$, a stable positive mosquito population supports an endemic steady state. The authors also define a vector demographic number $R_v$ and show that early-life-stage control can eliminate mosquitoes and stop transmission even when $R_0$ is above one. This matters because it ties dengue intervention strategy to measurable age structure and vector demography.","feed_headline":"Dengue dies out when R0 stays below 1—even if mosquitoes remain","feed_subtitle":"Mosquito egg-stage control can stop transmission even when the basic reproductive number is high","key_machinery":"The engine is the pair of thresholds $R_0$ and $R_v$. The vector demographic number $R_v=f'(N_v^*)/\\varphi$, with $\\varphi=(\\delta+\\mu_e)\\mu_v/\\delta$ the proportion of eggs that survive to adulthood, determines whether mosquitoes persist and which of possibly many demographic steady states of the birth function $f(N_v)=N_v g(N_v)$ are stable. The basic reproductive number $R_0$ is a double integral over host age that weights transmission by host survival to each age and by the limiting host age distribution $p_\\infty(a)$. The stability proofs chain the two together: $R_v$ fixes the vector background, and on that background $R_0$ decides whether the infection grows or decays.","core_discovery":"The paper's central claim is that in a dengue model with age-structured human hosts and an early-life stage in mosquitoes, the infection threshold is still a single basic reproductive number, now given by\n$$R_0 = \\frac{\\beta_v}{\\mu_v} \\int_0^\\infty \\int_0^a p_\\infty(a)\\$\\beta$(\\tau)$e^{{-\\int_\\tau^a \\gamma(h)\\,dh}}$\\,d\\tau\\,da.$$\nThis $R_0$ incorporates the age-specific contact rate $\\beta(a)$, the age-specific recovery rate $\\gamma(a)$, the limiting host age distribution $p_\\infty(a)$, and the mosquito-to-human transmission efficiency. The paper proves that if $R_0<1$, the disease-free state is globally asymptotically stable, so infection disappears from every initial condition. If $R_0>1$ and the vector population has a stable positive demographic steady state, an endemic steady state exists. It also defines a vector demographic number $R_v=f'(N_v^*)/\\varphi$, with $\\varphi=(\\delta+\\mu_e)\\mu_v/\\delta$ the proportion of eggs that survive to adulthood, and shows that egg-stage control making $\\varphi$ exceed the maximum of the per-capita egg-production function $g(N_v)$ eliminates the vector population regardless of $R_0$.","pith_inferences":["A testable extension is to estimate $R_0$ from age-specific seroprevalence and vector survival data; the model would then predict which age groups dominate incident cases, which can be compared with surveillance records.","The existence of multiple vector demographic steady states suggests possible bistability: for the same $R_0$, whether an outbreak takes off could depend on the initial mosquito density, a phenomenon the paper's simulations display but the theory only treats through local stability.","Introducing periodic (seasonal) parameters in the same framework would likely convert steady endemic states into periodic outbreak orbits; the paper simulates this but does not analyze the stability of such orbits.","A multi-serotype version with temporary cross-immunity or antibody-dependent enhancement could inherit the same age-structured $R_0$ structure, but that extension is not addressed here."],"forward_implications":["If $R_0<1$, dengue dies out from every starting condition, even when adult mosquitoes remain abundant and the demographic vector number $R_v$ is above one.","If $R_0>1$, an endemic steady state exists only when the mosquito population has a stable positive demographic steady state, so vector persistence is a precondition for sustained transmission.","Early-life-stage control that pushes $\\varphi$ above the maximum of $g(N_v)$ eliminates mosquitoes and stops dengue independently of the value of $R_0$.","Because $R_0$ depends on age-specific contact and recovery, control measures aimed at the highest-contact age groups can lower the threshold directly.","Seasonal variation in vector parameters, as simulated in the paper, changes the age profile of infected hosts, so the timing of vector control may affect different age groups differently."],"supporting_citations":[{"why":"Supplies the baseline dengue transmission model whose threshold structure the present model extends to age-structured hosts.","marker":"[3]"},{"why":"Provides the variable-human-population dengue model that motivates the host demographic structure used here.","marker":"[4]"},{"why":"Contributes the vector-host competition framework for dengue that this age-structured, single-serotype model builds on.","marker":"[5]"},{"why":"Introduces the early-life-stage vector dynamics and the multi-density egg-production function adopted in the model.","marker":"[11]"}],"fun_headline_variants":["Dengue vanishes when R0<1 despite living mosquitoes","Egg-stage control eradicates dengue vector irrespective of R0","Global dengue elimination proven for R0 below unity","Single R0 threshold decides dengue fate in age-structured model","Mosquito egg-stage intervention defeats dengue regardless of R0"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof that $R_0<1$ forces infection to die out relies on an asserted limiting balance $B^*=(\\mu_v/\\beta_v)I_v^*/S_v^*$ between the force of infection and the infected-vector fraction; the paper does not derive this balance for limit inferior or limit superior values, and if it fails, the global stability conclusion is not established.","fun_headline_variants_meta":{"raw":{"variants":["Dengue vanishes when R0<1 despite living mosquitoes","Egg-stage control eradicates dengue vector irrespective of R0","Global dengue elimination proven for R0 below unity","Single R0 threshold decides dengue fate in age-structured model","Mosquito egg-stage intervention defeats dengue regardless of R0"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000767,"raw_usage":{"total_tokens":3363,"prompt_tokens":871,"completion_tokens":2492,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":487,"completion_tokens_details":{"reasoning_tokens":2411}},"tokens_in":487,"tokens_out":2492,"duration_ms":21259,"temperature":1.0,"reasoning_tokens":2411,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:44:00.942777+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run System (3) numerically with parameters giving $R_0<1$ but a stable positive vector steady state; if the infected-host density $i_h(t,a)$ converges to a positive steady state from some initial condition, Theorem 3.4 is false. A simpler check is to track the ratio $B(t)/(I_v(t)/S_v(t))$ along transients and see whether it approaches $\\mu_v/\\beta_v$ as required by Eq. (9).","supporting_citations":[{"cited_title":"(1997) ”Analysis of a dengue disea se transmission model”, Math Biosci, 150:131–151","cited_arxiv_id":null,"evidence_quote":"Supplies the baseline dengue transmission model whose threshold structure the present model extends to age-structured hosts."},{"cited_title":"(1999) ”A model for dengue disease w ith variable human population”, J Math Biol , 38:220–240","cited_arxiv_id":null,"evidence_quote":"Provides the variable-human-population dengue model that motivates the host demographic structure used here."},{"cited_title":"(1997) ”Competitive e xclusion in a vector-host model for the dengue fever”, J Math Biol, 35:423–544","cited_arxiv_id":null,"evidence_quote":"Contributes the vector-host competition framework for dengue that this age-structured, single-serotype model builds on."},{"cited_title":"Models for Dengue Transmission and Control","cited_arxiv_id":null,"evidence_quote":"Introduces the early-life-stage vector dynamics and the multi-density egg-production function adopted in the model."}],"review_version":1}