Background
Despite progress made in the last decades, malaria persists as a pressing health issue in sub-Saharan Africa. Pregnant women are particularly vulnerable to infection and serious health outcomes for themselves and their unborn child. Risk can be mitigated through appropriate use of control measures such as insecticide-treated bed nets. Although social networks can influence uptake of preventive strategies, the role of social influence on bed net ownership has not been explored. During an evaluation of a bed net distribution programme, the influence of non-health care advisors on ownership and use of bed nets by pregnant women in Kumasi, Ghana was examined.

Methods
Data were collected through in-person interviews with 300 pregnant women seeking antenatal care in an urban hospital in Kumasi, Ghana. Participants were asked about their bed net ownership, bed net use, and information about three personal contacts that they go to for pregnancy advice. Information about these advisors was combined into an influence score. Logistic regression models were used to determine the association between the score and bed net ownership. Those who owned a bed net were further assessed to determine if interpersonal influence was associated with self-reported sleeping under the bed net the previous night.

Results
Of the 294 women in the analysis, 229 (78%) reported owning bed nets. Of these bed net owners, 139 (61%) reported using a bed net the previous night. A dose response relationship was observed between the interpersonal influence score and bed net ownership and use. Compared to the lowest influence score, those with the highest influence score (>1 SD above the mean) were marginally more likely to own a bed net [OR = 2.37, 95% CI (0.87, 6.39)] and much more likely to use their bed net [5.38, 95% CI (1.89, 15.25)] after adjusting for other factors.

Conclusions
Interpersonal influence appears to have modest impact on ownership and use of bed nets by pregnant women in an urban area of Ghana. Further investigations would need to be conducted to determine if the relationship is causal or if individuals who associate are simply more likely to have similar practices.