This study proposes a methodology for analysis of catastrophic spending on health in statistically underdeveloped
countries. A binary logistic regression model, based on data from households with reported non-zero
expenditure on health, is proposed for the estimation of the likelihood of spending on health for all households
irrespective of whether they spent on health or not within the reference period for the survey. “Univariate”
discriminant functions, also based on data from households who spent on health within the reference period of the
survey, were proposed for discriminating households that made catastrophic expenditure on health from those who
did not. An application of this methodology to the data from the Ghana living Standards survey (round V) indicates
that the binary logistic regression model estimates correctly at least 78% of household’s likelihood of spending on
health while correctly discriminating the households as having a catastrophic expenditure.