Previous: Examples

Up: Examples

Next: Incorporating Additional Information Into

Estimating Cause-specific Adult Mortality Fractions

In this section, we give an example of estimating cause-specific mortality fractions based on simulated data constructed from a verbal autopsy survey and related hospital deaths. In these verbal autopsy data, there are 764 registered deaths from hospital and 738 deaths collected from the general population. The total number of symptoms are 49, and there are 19 causes of reported deaths. Below we give the R syntax we use to analyze these data. The results are plotted in Figure 1.

  > data(VAdata)
  > FBA<-VAdata[VAdata$region==1,-1]
  > DSS<-VAdata[VAdata$region==2,-1]
  > res1<-va(formula=cbind(S1+...+S49)~cod,
          data=list(FBA,DSS),  nsymp=16,
          n.subset=300,prob.wt=1,
          printit=TRUE,boot.se=FALSE)

Figure 1: Estimation of Adult Cause-Specific Mortality Fractions. The observed cause-specific mortality is plotted horizontally and our verbal autopsy estimate is plotted vertically.