Catastrophic health care payment reconsidered: An ethical perspective with application to Nigeria

Presenter: John Ataguba, University of Cape Town

Abstract

Rationale: From an ethical perspective, defining ‘catastrophe’ or ‘catastrophic health care payments’ should incorporate the notion of equity in health financing. If we accept progressivity in health financing as the ethical principle (i.e. the case where those in the upper end of the income distribution spend more as a proportion of their income on health care than those at the lower end), we cannot accept the uniform threshold for categorizing households as making catastrophic payments. Though it is difficult in this regard to suggest an appropriate variation reflecting ethical concern across the entire distribution, a percentile-dependent threshold is proposed.

Objectives: The paper modifies the measurement of catastrophic payments and seeks to implement a new paradigm in assessing catastrophic payments. It questions the uniform threshold based on the paper by Adam Wagstaff and Eddy van Doorslaer (WvD).

Methodology: The paper applies a function that weights a specified fixed threshold by a parameter g (i.e. the density of the lowest income rank in a sample of g randomly selected individuals). This implies that the effective threshold varies according to an individual’s location on the income distribution ladder. Though an initial attempt was considered by WvD, the method rather weights, using the concentration index, those already categorized as making catastrophic payment. It therefore, only accounts for concentration of payments along the distribution.
The proposed method is empirically applied to Nigerian data (Nigerian Living Standard Survey 2003/2004) to assess catastrophic incidence and intensity.

Results: It shows that the catastrophic headcount (positive gap) obtained using a uniform threshold – whether or not it is weighted by the concentration index, is lower (higher) than those predicted by the rank-dependent threshold. Using an initial threshold of 10% of household expenditure, the catastrophic headcount increased from 25.5% (for g=1) to 28.3% (for g=0.6) while the mean positive gap decreased from 23.6% (for g=1) to 21.4% (for g=0.6). With an initial threshold of 40% of household non-food expenditure, the catastrophic headcount increased from 17.2% (for g=1) to 20.2% (for g=0.6) while the mean positive gap decreased from 28.4% (for g=1) to 26.6% (for g=0.6). Also, the concentration of catastrophic payment among those in the upper end of the distribution increased with increasing value of the parameter (i.e. increasing the thresholds of the poor while simultaneously decreasing those of the rich). The results obtained in WvD are easily replicated when the value of the parameter g=1, which makes it a special case of the proposed method.

Conclusions: The rank-dependent measures proposed are motivated by ethical consideration for the poor. In the light of policy making, we are often interested in the welfare condition of the worse-off. The method of using a uniform cut-off is therefore not likely to incorporate the need for the notion of equity in health financing. The extent of catastrophe could be underestimated (overestimated) when higher (lower) uniform threshold is used. Therefore, the proposed percentile-dependent threshold is likely to reflect the reality facing households with regard to extent of catastrophic payments.