Projecting social contact matrices in 152 countries using contact surveys and demographic data

Abstract

Heterogeneities in contact networks have a major effect in determining whether a pathogen can become epidemic or persist at endemic levels. Epidemic models that determine which interventions can successfully prevent an outbreak need to account for social structure and mixing patterns. Contact patterns vary across age and locations (e.g. home, work, and school), and including them as predictors in transmission dynamic models of pathogens that spread socially will improve the models' realism. Data from population-based contact diaries in eight European countries from the POLYMOD study were projected to 144 other countries using a Bayesian hierarchical model that estimated the proclivity of age-and-location-specific contact patterns for the countries, using Markov chain Monte Carlo simulation. Household level data from the Demographic and Health Surveys for nine lower-income countries and socio-demographic factors from several on-line databases for 152 countries were used to quantify similarity of countries to estimate contact patterns in the home, work, school and other locations for countries for which no contact data are available, accounting for demographic structure, household structure where known, and a variety of metrics including workforce participation and school enrolment. Contacts are highly assortative with age across all countries considered, but pronounced regional differences in the age-specific contacts at home were noticeable, with more inter-generational contacts in Asian countries than in other settings. Moreover, there were variations in contact patterns by location, with work-place contacts being least assortative. These variations led to differences in the effect of social distancing measures in an age structured epidemic model. Contacts have an important role in transmission dynamic models that use contact rates to characterize the spread of contact-transmissible diseases. This study provides estimates of mixing patterns for societies for which contact data such as POLYMOD are not yet available.

Fig 1. Methodology and data.

Overview of the data sources and model framework in the manuscript is presented in this flow chart. The categories of the 152 countries are depicted on the world map (i.e. POLYMOD, Demographic and Health Survey (DHS), and Rest of the World (ROW) countries) and their data sources are listed in the table. A summary of the methodology is represented by the model framework: (A) POLYMOD model, (B) construction age-structured populations at home, work, and school in the 152 countries, and (C) projection of global estimates.

Fig 2. Contacts at home stratified by household size.

Panel a shows the number of contacts made by individuals at home (grey dots) in the POLYMOD study stratified by household sizes. The estimated mean number of contacts (red dots) made by an individual and the 95% confidence interval (orange lines) are shown. By observation, a near linear relationship exists between the household size and the number of contacts. The blue line represents a fitted linear model. The panels b–f show the age-specific contact patterns at home of individuals stratified by household sizes. Darker color intensities indicate more contacts were made.

Fig 3. Population and household age distribution, and age-specific contacts at home.

The population pyramids by age and gender (panels a–c), household age matrices (panels d–f) and age-specific contact patterns (panels g–i) are presented for Germany (first column, as a representative of the POLYMOD countries), Bolivia (second column, as a representative of DHS) and South Africa (third column, as a representative of ROW). The population pyramids, panels a–c, and household age matrices (for only POLYMOD and DHS), panels d–e, are observed data. The age-specific contacts at home for Germany (g) is estimated from our hierarchical model. The household age matrix for South Africa (f) and the age-specific contacts at home for Bolivia (h) and South Africa (i) were projected using the described methods. Darker color intensities indicate more likely events i.e. greater tendency of having a household member of that age, higher proclivity of making the age-specific contact.

Fig 4. Inferred regional contact patterns at home.

Countries of the world were group into 7 regions (East Asia & Pacific, Europe & Central Asia, Latin America & Caribbean, Middle East & North Africa, North America, South Asia and Sub-Saharan Africa). The regional mean age-specific contact patterns at home (inferred) of individuals aged 5–10 (first column), 25–30 (second column) and 55–60 (third column) years were represented as bars.

Fig 5. Age-specific contact patterns by location.

The age-specific contact patterns at home (panels a–c), at the workplace (panels d–f), in school (panels g–i) and at other locations (panels j–l) are projected from the model. The contact pattern at all locations (panels m–o) is the sum across the four locations (home, work, school and others). Contact matrices for Bolivia (DHS country; in panels b,e,h,k) and South Africa (ROW country; in panels c,f,i,l) were projected and the age-specific mean contact rates for Germany (part of the POLYMOD; in panels a,d,g,j) were estimated from the German contact data. A comparison between the German empirical and modelled estimates can be found in the S1 Text. Darker color intensities indicate higher proclivity of making the age-specific contact.

Fig 6. Age-specific final epidemic size and percentage reduction.

The age-specific final epidemic size and percentage reduction of infection for Germany (first column), Bolivia (second column) and South Africa (third column) are shown for the three interventions: No intervention (sum of orange and pink/blue bars), School closure and social distancing of younger individuals (blue bars) and Workplace distancing (pink bars) for two epidemics with R₀ of 1.2 and 1.5. The percentage reduction of infection for the various intervention and R₀ values are represented by the black lines.