pith. sign in

arxiv: 2605.13870 · v1 · pith:ZN5GZUPRnew · submitted 2026-05-04 · ⚛️ physics.soc-ph · q-bio.PE

A method for including socio-demographic factors in social contact matrices for compartment-based epidemic models

Vincent X. Lomas , Tim Chambers , Leighton Watson , Michael Plank This is my paper

Pith reviewed 2026-05-15 07:20 UTC · model grok-4.3

classification ⚛️ physics.soc-ph q-bio.PE
keywords social contact matricesepidemic modelingsocio-demographic factorscompartment modelsmixing patternsreproduction numberepidemic size
0
0 comments X

The pith

A method stratifies existing social contact matrices with an extra socio-demographic factor using population structure data and mixing-rate assumptions.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a procedure to add a second socio-demographic dimension to an age-based contact matrix without requiring a new survey that records every factor. The procedure uses known population proportions together with simple assumptions on how people mix inside and across the new groups. When the extended matrix is inserted into a standard compartment model, both the reproduction number and the final epidemic size change markedly. Minority groups exhibit the largest shifts in outcomes when contact or transmission parameters vary. The approach therefore lets models incorporate compounding demographic effects on spread while remaining computationally light.

Core claim

The stratification procedure combines an existing social contact matrix with demographic proportions and assumptions about aggregate mixing rates within and between groups; the resulting matrices produce substantial differences in model reproduction number and final epidemic size, with minority-group outcomes most sensitive to parameter changes.

What carries the argument

The stratification procedure that extends a base contact matrix by weighting entries according to demographic proportions and assumed within-group versus between-group mixing rates.

If this is right

  • The basic reproduction number changes once the extra factor is included.
  • Projected final epidemic sizes differ substantially from those obtained with age-only matrices.
  • Epidemic outcomes for smaller demographic groups respond most strongly to variation in contact or transmission parameters.
  • Compartment models can now account for compounding effects of multiple socio-demographic factors without requiring full multi-factor surveys.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Health authorities could use the method to identify which demographic intersections are most sensitive and to allocate limited interventions accordingly.
  • Validation against existing multi-factor contact surveys would reveal how accurate the mixing-rate assumptions are in real populations.
  • Extending the same procedure to additional factors such as household size or occupation would expose further heterogeneities in outbreak risk.

Load-bearing premise

The approach depends on assumptions about aggregate mixing rates within and between socio-demographic groups.

What would settle it

Direct comparison of contact frequencies predicted by the extended matrix against observed contacts recorded in a survey that collects both age and the additional socio-demographic factor.

Figures

Figures reproduced from arXiv: 2605.13870 by Leighton Watson, Michael Plank, Tim Chambers, Vincent X. Lomas.

Figure 1
Figure 1. Figure 1: Heat-plots of the purely proportionate mixing contact matrices for each scenario; plots [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Heat-plot of the segregated mixing contact matrices for each scenario; plots show scenarios [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Plots of the variation in a) the basic reproductive number; b) the whole population attack [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Plots of the variation in a) the basic reproductive number with a dotted line at [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Heat-maps showing how variation in the socio-demographic factor assortativity and relative [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Ethnicity-aggregated results of a SEIR model run using the POLYMOD contact survey [PITH_FULL_IMAGE:figures/full_fig_p014_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Age-aggregated results of a SEIR model run using the POLYMOD contact survey projected [PITH_FULL_IMAGE:figures/full_fig_p023_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Bar-plot showing the percentage difference (of the population) in final attack rates when [PITH_FULL_IMAGE:figures/full_fig_p023_8.png] view at source ↗
read the original abstract

Socio-demographic factors influence social contact patterns and play a fundamental role in shaping the transmission dynamics of infectious diseases. However, compartment-based models of infectious disease dynamics commonly consider the dependence of contact patterns on age, but ignore other factors that are likely to have compounding effects. Methods that stratify the population by multiple socio-demographic factors are few and require social contact surveys that contain information about all factors of interest. Here we present a method that can stratify an existing social contact matrix with an additional socio-demographic factor using information about the population structure of the socio-demographic factors and assumptions about the aggregate mixing rates within and between groups. We then analyse hypothetical populations and a projection of a social contact survey onto Aotearoa New Zealand's age-ethnic structure to show how these extended social contact matrices can change epidemic dynamics and outcomes. The inclusion of the additional factor has a big impact on the model reproduction number and final epidemic size. We find that minority group epidemic outcomes are most sensitive to variation in model parameter values.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes a method to extend an existing age-stratified social contact matrix to include an additional socio-demographic factor (such as ethnicity) by combining population structure counts with explicit assumptions on aggregate within- and between-group mixing rates. These rates populate the block entries of the extended matrix. The authors illustrate the approach on hypothetical populations and a projection of a contact survey onto Aotearoa New Zealand’s age-ethnic structure, claiming that the added stratification produces large changes in the model reproduction number and final epidemic size, with minority-group outcomes being most sensitive to parameter variation.

Significance. If the mixing-rate assumptions can be shown to be consistent with empirical contact data, the method would supply a practical route to multi-factor stratification in compartment models without requiring new surveys that record every factor simultaneously. This could improve realism in heterogeneous populations. The current manuscript, however, leaves the quantitative claims dependent on unvalidated assumptions whose effect size is not benchmarked against observed contact patterns, so the practical significance remains conditional on further validation.

major comments (2)
  1. [Methods (extension procedure)] The central quantitative claims rest on postulated aggregate mixing rates within and between groups that determine the block entries of the extended contact matrix and therefore the eigenvalues of the next-generation matrix. These rates are free parameters not derived from the epidemic data being modeled; no direct comparison to any survey that records contacts for both age and the additional factor is provided.
  2. [Results] In the hypothetical-population and New Zealand projection analyses, the reported changes in reproduction number and final epidemic size are driven by the chosen mixing rates. The manuscript does not quantify how sensitive these outcomes are to plausible variation in the rates or benchmark the effect sizes against empirical contact matrices that already include multiple factors.
minor comments (2)
  1. [Abstract] The abstract states that minority-group outcomes are “most sensitive” without specifying the range of parameter values explored or the metric used to establish relative sensitivity.
  2. [Methods] Notation for the extended contact matrix and the aggregate mixing rates should be introduced with explicit symbols and a small worked example to improve readability for readers unfamiliar with the construction.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for their constructive and detailed review. We agree that the mixing-rate assumptions require clearer justification and that sensitivity should be quantified. Below we respond point-by-point to the major comments and indicate the revisions we will make.

read point-by-point responses

  1. Referee: The central quantitative claims rest on postulated aggregate mixing rates within and between groups that determine the block entries of the extended contact matrix and therefore the eigenvalues of the next-generation matrix. These rates are free parameters not derived from the epidemic data being modeled; no direct comparison to any survey that records contacts for both age and the additional factor is provided.

    Authors: We agree that the within- and between-group mixing rates are modeling assumptions rather than parameters estimated from epidemic incidence data. The method is explicitly designed for settings in which no survey simultaneously records all factors of interest; it combines an existing age matrix with demographic counts and aggregate mixing assumptions to produce a usable multi-factor matrix. We will revise the Methods section to state these assumptions more explicitly, justify their ranges with reference to the assortative-mixing literature, and add a dedicated limitations paragraph noting the absence of a joint age-ethnicity contact survey for New Zealand. revision: partial

  2. Referee: In the hypothetical-population and New Zealand projection analyses, the reported changes in reproduction number and final epidemic size are driven by the chosen mixing rates. The manuscript does not quantify how sensitive these outcomes are to plausible variation in the rates or benchmark the effect sizes against empirical contact matrices that already include multiple factors.

    Authors: We accept the need for sensitivity quantification. In the revised manuscript we will add a systematic sensitivity analysis that varies the within- and between-group mixing rates over a plausible range (from strongly assortative to disassortative) and reports the resulting variation in reproduction numbers and final epidemic sizes for both the hypothetical populations and the New Zealand projection. Regarding benchmarking, comprehensive empirical matrices stratified by both age and ethnicity are not publicly available; we will therefore compare the projected matrices against any partial ethnic-mixing data in the literature and discuss the magnitude of the observed effects relative to single-factor matrices. revision: yes

standing simulated objections not resolved
  • Direct empirical comparison to a survey that records contacts stratified simultaneously by age and ethnicity, because no such dataset exists for the New Zealand population studied.

Circularity Check

0 steps flagged

No circularity: explicit assumptions on mixing rates are inputs, not derived outputs

full rationale

The paper's method constructs an extended contact matrix from an existing age-stratified matrix, external population-structure counts, and postulated aggregate within- and between-group mixing rates. These rates are stated as modeling assumptions and directly determine the block entries; the subsequent computation of next-generation-matrix eigenvalues and final epidemic sizes follows standard compartment-model arithmetic applied to the constructed matrix. No equation reduces its claimed result to the inputs by construction, no parameter is fitted to epidemic data and then relabeled as a prediction, and no self-citation chain or uniqueness theorem is invoked to justify the central construction. The reported sensitivity of minority-group outcomes is a direct numerical consequence of varying the explicit mixing-rate inputs, not a tautological re-expression of those inputs.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The method rests on postulated aggregate mixing rates within and between socio-demographic groups; these rates are introduced to close the model and are not derived from independent contact data.

free parameters (1)
  • aggregate mixing rates within and between groups
    Chosen to satisfy the assumed overall contact totals; no empirical calibration is reported in the abstract.
axioms (2)
  • domain assumption Existing age-only contact matrix remains valid when re-weighted by the new factor
    Invoked when the paper projects the matrix onto the age-ethnic structure.
  • domain assumption Population structure counts are known and accurate
    Used directly to stratify the matrix for the New Zealand case.

pith-pipeline@v0.9.0 · 5488 in / 1300 out tokens · 42354 ms · 2026-05-15T07:20:08.502537+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

25 extracted references · 25 canonical work pages

  1. [1]

    A mathematical model reveals the influence of population heterogeneity on herd immunity to SARS-CoV-2

    Britton T, Ball F, and Trapman P. A mathematical model reveals the influence of population heterogeneity on herd immunity to SARS-CoV-2. Science 2020 Aug; 369:846–9.doi:10.112 6/science.abc6810. Available from:https://www.science.org/doi/10.1126/science.ab c6810[Accessed on: 2025 Jan 28]

    work page doi:10.1126/science.ab 2020

  2. [2]

    Incorpo- rating social vulnerability in infectious disease mathematical modelling: a scoping review

    Naidoo M, Shephard W, Kambewe I, Mtshali N, Cope S, Rubio FA, and Rasella D. Incorpo- rating social vulnerability in infectious disease mathematical modelling: a scoping review. en. BMC Medicine 2024 Mar; 22:125.doi:10 . 1186 / s12916 - 024 - 03333 - y. Available from: https://doi.org/10.1186/s12916-024-03333-y[Accessed on: 2026 Apr 2]

    work page doi:10.1186/s12916-024-03333-y 2024

  3. [3]

    Age Differences in Daily Social Activities

    Marcum CS. Age Differences in Daily Social Activities. EN. Research on Aging 2013 Sep; 35:612–40.doi:10.1177/0164027512453468. Available from:https://doi.org/10.1177/01 64027512453468[Accessed on: 2026 Mar 19]

    work page doi:10.1177/0164027512453468 2013

  4. [4]

    Using Data on Social Contacts to Estimate Age- specific Transmission Parameters for Respiratory-spread Infectious Agents

    Wallinga J, Teunis P, and Kretzschmar M. Using Data on Social Contacts to Estimate Age- specific Transmission Parameters for Respiratory-spread Infectious Agents. American Journal of Epidemiology 2006 Nov; 164:936–44.doi:10.1093/aje/kwj317. Available from:https: //doi.org/10.1093/aje/kwj317[Accessed on: 2025 Oct 2]

    work page doi:10.1093/aje/kwj317 2006

  5. [5]

    Improving the Use of Social Contact Studies in Epidemic Modeling

    Britton T and Ball F. Improving the Use of Social Contact Studies in Epidemic Modeling. en-US. Epidemiology 2025 Sep; 36:660.doi:10 . 1097 / EDE . 0000000000001876. Available from:https://journals.lww.com/epidem/fulltext/2025/09000/improving_the_use_of _social_contact_studies_in.9.aspx[Accessed on: 2026 Feb 20]

    work page 2025

  6. [6]

    Workshop Report—Heterogeneity and Successful Aging Part I: Heterogeneity in Aging—Challenges and Opportunities

    Kuchel GA, Hevener AL, Ruby JG, Sebastiani P, and Kumar V. Workshop Report—Heterogeneity and Successful Aging Part I: Heterogeneity in Aging—Challenges and Opportunities. The jour- nals of gerontology. Series A, Biological sciences and medical sciences 2025 Mar; 80:glaf023. doi:10.1093/gerona/glaf023. Available from:https://pmc.ncbi.nlm.nih.gov/articles /...

    work page doi:10.1093/gerona/glaf023 2025

  7. [7]

    Bedson, L

    Bedson J, Skrip LA, Pedi D, Abramowitz S, Carter S, Jalloh MF, Funk S, Gobat N, Giles- Vernick T, Chowell G, Almeida JR de, Elessawi R, Scarpino SV, Hammond RA, Briand S, Epstein JM, H´ ebert-Dufresne L, and Althouse BM. A review and agenda for integrated disease models including social and behavioural factors. en. Nature Human Behaviour 2021 Jul; 5:834–4...

    work page doi:10.1038/s41562-021-01136-2 2021

  8. [8]

    There are no equal opportunity infectors: Epidemiological modelers must rethink our approach to inequality in infection risk

    Zelner J, Masters NB, Naraharisetti R, Mojola SA, Chowkwanyun M, and Malosh R. There are no equal opportunity infectors: Epidemiological modelers must rethink our approach to inequality in infection risk. en. PLOS Computational Biology 2022 Feb; 18:e1009795.doi: 10.1371/journal.pcbi.1009795. Available from:https://journals.plos.org/ploscompb iol/article?i...

    work page doi:10.1371/journal.pcbi.1009795 2022

  9. [9]

    Addressing the socioe- conomic divide in computational modeling for infectious diseases

    Tizzoni M, Nsoesie EO, Gauvin L, Karsai M, Perra N, and Bansal S. Addressing the socioe- conomic divide in computational modeling for infectious diseases. en. Nature Communications 2022 May; 13:2897.doi:10.1038/s41467-022-30688-8. Available from:https://www.natu re.com/articles/s41467-022-30688-8[Accessed on: 2025 Dec 2]

    work page doi:10.1038/s41467-022-30688-8 2022

  10. [10]

    Pandemic inequalities: emerging infectious diseases and health equity

    Bambra C. Pandemic inequalities: emerging infectious diseases and health equity. en. Interna- tional Journal for Equity in Health 2022 Jan; 21:6.doi:10.1186/s12939- 021- 01611- 2. Available from:https://doi.org/10.1186/s12939-021-01611-2[Accessed on: 2026 Apr 2]

    work page doi:10.1186/s12939- 2022

  11. [11]

    Social Justice in Pandemic Preparedness

    DeBruin D, Liaschenko J, and Marshall MF. Social Justice in Pandemic Preparedness. Amer- ican Journal of Public Health 2012 Apr; 102:586–91.doi:10 . 2105 / AJPH . 2011 . 300483. Available from:https://ajph.aphapublications.org/doi/full/10.2105/AJPH.2011.300 483[Accessed on: 2026 Feb 17]

    work page doi:10.2105/ajph.2011.300 2012

  12. [12]

    Health inequities in influenza transmission and surveillance

    Zipfel CM, Colizza V, and Bansal S. Health inequities in influenza transmission and surveillance. en. PLOS Computational Biology 2021 Mar; 17:e1008642.doi:10.1371/journal.pcbi.100

    work page doi:10.1371/journal.pcbi.100 2021

  13. [13]

    Available from:https://journals.plos.org/ploscompbiol/article?id=10.1371/j ournal.pcbi.1008642[Accessed on: 2026 Apr 2]

    work page doi:10.1371/j 2026

  14. [14]

    COVID-19 inequalities in England: a mathemat- ical modelling study of transmission risk and clinical vulnerability by socioeconomic status

    Goodfellow L, Leeuwen E van, and Eggo RM. COVID-19 inequalities in England: a mathemat- ical modelling study of transmission risk and clinical vulnerability by socioeconomic status. BMC Medicine 2024 Apr; 22:162.doi:10 . 1186 / s12916 - 024 - 03387 - y. Available from: https://doi.org/10.1186/s12916-024-03387-y[Accessed on: 2024 Dec 3]

    work page doi:10.1186/s12916-024-03387-y 2024

  15. [15]

    Importance of social inequalities to contact patterns, vaccine uptake, and epidemic dynamics

    Manna A, Koltai J, and Karsai M. Importance of social inequalities to contact patterns, vaccine uptake, and epidemic dynamics. en. Nature Communications 2024 May; 15:4137.doi:10.10 38/s41467-024-48332-y. Available from:https://www.nature.com/articles/s41467-024 -48332-y[Accessed on: 2025 Apr 30]

    work page 2024

  16. [16]

    Generalized contact matrices allow integrating socioeconomic variables into epidemic models

    Manna A, Dall’Amico L, Tizzoni M, Karsai M, and Perra N. Generalized contact matrices allow integrating socioeconomic variables into epidemic models. Science Advances 2024 Oct; 10:eadk4606.doi:10.1126/sciadv.adk4606. Available from:https://www.science.org/do i/10.1126/sciadv.adk4606[Accessed on: 2025 May 6]

    work page doi:10.1126/sciadv.adk4606 2024

  17. [17]

    Individual-based and neighbourhood-based socio-economic factors relevant for contact behaviour and epidemic control

    Domenico LD, Reichmuth ML, and Althaus CL. Individual-based and neighbourhood-based socio-economic factors relevant for contact behaviour and epidemic control. en. Pages: 2025.03.24.25324502. 2025 Mar.doi:10.1101/2025.03.24.25324502. Available from:https://www.medrxiv.org /content/10.1101/2025.03.24.25324502v1[Accessed on: 2025 Apr 11]

    work page doi:10.1101/2025.03.24.25324502 2025

  18. [18]

    Simulating the impact of perception bias on social contact surveys for infectious disease modelling

    Harris TJ, Alexander PC, Pham ABD, Tuccillo J, Geard N, and Zachreson C. Simulating the impact of perception bias on social contact surveys for infectious disease modelling. en. 2025 Nov. Available from:https://arxiv.org/abs/2511.03897v1[Accessed on: 2026 Apr 2]

    work page arXiv 2025

  19. [19]

    Social Contacts and Mixing Patterns Relevant to the Spread of Infectious Diseases

    Mossong J, Hens N, Jit M, Beutels P, Auranen K, Mikolajczyk R, Massari M, Salmaso S, Tomba GS, Wallinga J, Heijne J, Sadkowska-Todys M, Rosinska M, and Edmunds WJ. Social Contacts and Mixing Patterns Relevant to the Spread of Infectious Diseases. en. PLOS Medicine 2008 Mar; 5:e74.doi:10.1371/journal.pmed.0050074. Available from:https://journals.plos .org/...

    work page doi:10.1371/journal.pmed.0050074 2008

  20. [20]

    conmat: generate synthetic contact matrices for a given age-stratified population

    Tierney N, Saraswati C, Babu A, Lydeamore M, and Golding N. conmat: generate synthetic contact matrices for a given age-stratified population. en. Journal of Open Source Software 2026 Feb; 11:8326.doi:10.21105/joss.08326. Available from:https://joss.theoj.org/p apers/10.21105/joss.08326[Accessed on: 2026 Mar 5] 18

    work page doi:10.21105/joss.08326 2026

  21. [21]

    2023 Census population counts (by ethnic group, age, and M¯ aori descent) and dwelling counts — Stats NZ

    StatsNZ. 2023 Census population counts (by ethnic group, age, and M¯ aori descent) and dwelling counts — Stats NZ. 2023. Available from:https://www.stats.govt.nz/information-rele ases/2023-census-population-counts-by-ethnic-group-age-and-maori-descent-and- dwelling-counts/[Accessed on: 2026 Mar 5]

    work page 2023

  22. [22]

    Modelling the interaction between ethnicity and infectious disease transmission dynamics in Aotearoa New Zealand during the first Omicron wave of the COVID-19 pandemic

    Lomas VX, Chambers T, and Plank MJ. Modelling the interaction between ethnicity and infectious disease transmission dynamics in Aotearoa New Zealand during the first Omicron wave of the COVID-19 pandemic. Mathematics in Medical and Life Sciences 2025 Dec; 2. eprint: https://doi.org/10.1080/29937574.2025.2591407:2591407.doi:10.1080/29937574.20 25.2591407. ...

    work page doi:10.1080/29937574.2025.2591407:2591407.doi:10.1080/29937574.20 2025

  23. [23]

    Modeling the impact of racial and ethnic disparities on COVID-19 epidemic dynamics

    Ma KC, Menkir TF, Kissler S, Grad YH, and Lipsitch M. Modeling the impact of racial and ethnic disparities on COVID-19 epidemic dynamics. eLife 2021 May; 10. Ed. by Schiffer JT, Davenport MP, Schiffer JT, Zelner J, and Moore M:e66601.doi:10 . 7554 / eLife . 66601. Available from:https://doi.org/10.7554/eLife.66601[Accessed on: 2024 Dec 3]

    work page doi:10.7554/elife.66601 2021

  24. [24]

    Place and ethnic group summaries - Stats NZ

    StatsNZ. Place and ethnic group summaries - Stats NZ. 2023. Available from:https://tool s.summaries.stats.govt.nz/[Accessed on: 2026 Mar 6]

    work page 2023

  25. [25]

    30 30 30 30 30 30 30 30 30 30 #T h 1 1 1 1 1 i 2

    Harris T, Richter M, Alexander P, Kitson J, Tuccillo J, Parikh N, Germann T, and Del Valle SY. Why population heterogeneity matters for modelling infectious diseases. Interface Focus 2025 Sep; 15:20250006.doi:10.1098/rsfs.2025.0006. Available from:https://doi.org/1 0.1098/rsfs.2025.0006[Accessed on: 2026 Apr 2] 19 Appendix A: Satisfaction of conditions He...

    work page doi:10.1098/rsfs.2025.0006 2025