Social contact patterns in Vietnam and implications for the control of infectious diseases

Abstract

Background: The spread of infectious diseases from person to person is determined by the frequency and nature of contacts between infected and susceptible members of the population. Although there is a long history of using mathematical models to understand these transmission dynamics, there are still remarkably little empirical data on contact behaviors with which to parameterize these models. Even starker is the almost complete absence of data from developing countries. We sought to address this knowledge gap by conducting a household based social contact diary in rural Vietnam.

Methods and findings: A diary based survey of social contact patterns was conducted in a household-structured community cohort in North Vietnam in 2007. We used generalized estimating equations to model the number of contacts while taking into account the household sampling design, and used weighting to balance the household size and age distribution towards the Vietnamese population. We recorded 6675 contacts from 865 participants in 264 different households and found that mixing patterns were assortative by age but were more homogenous than observed in a recent European study. We also observed that physical contacts were more concentrated in the home setting in Vietnam than in Europe but the overall level of physical contact was lower. A model of individual versus household vaccination strategies revealed no difference between strategies in the impact on R(0).

Conclusions and significance: This work is the first to estimate contact patterns relevant to the spread of infections transmitted from person to person by non-sexual routes in a developing country setting. The results show interesting similarities and differences from European data and demonstrate the importance of context specific data.

Figure 1. Household sizes (A) and number of reported contacts per person per day (B).

Figure 2. Contacts by location, duration and frequency.

The figures are based on a WGEE with weights based on household size, days of the week and age.

Figure 3. The location, duration and frequency of contacts.

The proportion of contacts that were physical or non-physical by duration (panel A), location (panel B) and frequency of contact (panel C). The duration of contact by frequency of contact (panel D). The figures are based on a WGEE with weights based on household size and days of the week.

Figure 4. Contact intensity matrices for all contacts (A) and for physical contacts only (B).

Yellow indicates high contact rates and blue low contact rates, relative to the mean contact intensity.

Figure 5. The predicted effect on R ₀ of immunizing individuals or households.

The figure shows the predicted effect on R ₀ immunizing a random selection of individuals (solid line) versus a random selection of households (broken line).