# Estimating the global burden of disease (GBD)

## Julie Legler

This study for the World Health Organisation investigates the estimation of global burden of disease attributable to rotovirus. That is, counting the number of deaths globally due to rotovirus.

### Video

This video requires a web browser capable of HTML5 or Flash video.

The file rotovirus data.xls contains information for 176 countries on various health issues such as the proportion of diarrhoea cases due to rotovirus, the percentages of deaths for children under 5 due to diarrhoea, the population number under 5, female life expectancy, infant mortality, expenditure on health and fertility rate. The following summarises some results in the DVD and suggests several tasks which can be carried out. The proportions of diarrhoea cases due to rotovirus are given for 39 countries. When an asterisk appears the country estimate is the average obtained from several surveys. The average proportion of 0.36 is the average of the 39 proportions. Numbers in the population under 5 are reported. Percentages of diarrhoeal deaths for children under 5 are reported. There are four potential strata choices. The one used in the talk is based on Life Expectancy at birth for females. For each stratum a separate proportion is calculated. Then these modified proportions used.
• Task 1: Obtain the 95% confidence interval for the overall mean proportion using the normal distribution.
• Task 2: Obtain the bootstrap 95% confidence interval using the Excel bootstrap procedure bootstrapper.xls
• Task 3: Use female life expectancy to define the five strata. Calculate the proportion average in each stratum. Then calculate the numbers overall using these values.
• Task 4: Try predicting all the proportions based on female life expectancy as a predictor in a simple regression. (39 countries have a proportion). Compare the result from this procedure with the result obtained by using strata based on this variable.
• Task 5: Investigate individually the four predictors in the Excel file and decide which may be the best individual predictor by looking at the scatter diagrams and the correlation coefficients between these predictors and the known proportions for the 39 countries.
• Task 6: Find other WHO country health indicators from the WHO website for further calculations based on new strata.

Video content recorded and edited by Robert van der Vyver[1] and John Harraway[2].
Web site developed and maintained by Greg Trounson[2] and John Harraway[2]