Economic effect of angina on households in south Asia: a propensity score matching approach

BACKGROUND: Poor population health causes adverse national economic outcomes. Out-of-pocket medical expenditure, loss of earnings because of absence from work, decreased non-medical consumption expenditure (eg, on food, housing, education), and the subsequent effect on living standards underlie this aggregate economic effect. Worldwide, an estimated 54 million people have angina, of whom 16 million are from the WHO South-East Asia region. Despite the increasing burden of cardiovascular diseases in south Asia, there is no evidence of an economic effect of angina on households in this region. We investigated the economic effect of angina on households in Bangladesh, India, Nepal, and Sri Lanka. METHODS: We used nationally representative World Health Survey data from Bangladesh, India, Nepal, and Sri Lanka that included information for 31 966 households collected in 2003. We assessed the effects of angina on per person out-of-pocket health spending, drug spending, and hospital stays in the 4 weeks before the survey. We also investigated the effect of angina on non-medical consumption spending per person, workforce participation of the angina-affected individual, and household borrowing or selling of assets to finance out-of-pocket health expenses. We used propensity score matching to compare households in which a respondent reported being diagnosed with angina (treated cases) with control households (respondents not diagnosed with angina). We used the nearest-neighbour method to match angina-affected households with control households with similar propensity scores. FINDINGS: The nearest-neighbour method yielded 402 treatment households and 375 controls in Bangladesh, 881 treatment households and 803 controls in India, 431 treatment households and 400 controls in Nepal, and 274 treatment households and 262 controls in Sri Lanka. Angina-affected households had higher out-of-pocket health spending per person in the 4 weeks preceding the survey than did matched controls, ranging from US$0·73 in India to $3·26 in Nepal. This difference was mostly caused by out-of-pocket drug expenditure per person, which was significantly higher in angina-affected households than in matched controls; by $0·63 (p=0·01) in Sri Lanka, $0·69 (p=0·04) in India, $0·77 (p=0·05) in Bangladesh, and $2·32 (p=0·0006) in Nepal. However, per person hospital admission expenditure in angina-affected households did not differ from matched controls. Also, workforce participation of the angina-affected individual and non-medical expenditure per person did not differ between angina-affected households and matched controls. 12% of angina-affected households in India and 8% in Sri Lanka significantly relied on borrowing or selling assets to finance their out-of-pocket health expenses compared with matched controls. INTERPRETATION: Angina has a large effect on households' out-of-pocket health expenses, especially expenditure on drugs. However, households are able to maintain their non-medical consumption spending. Household reliance on borrowing or selling assets might partly explain this finding. Our study suggests the need for policy measures to address the financial risks that south Asian households face from angina. FUNDING: Endeavour International Postgraduate Research Scholarship, Australian Postgraduate Award.