Health Accounts presents an overview of sources of financing healthcare, how these finances are paid to the providers, for which services and for whom (population characteristics) within the health system at national or sub-national level for a financial year (FY). This information is presented in the Global Health Expenditure Database (GHED) for almost all countries based on health accounts. Out of Pocket expenditure (OOPE) is a major source of financing healthcare in most Low-and Middle-Income Countries (LMICs). OOPE is defined as any payment made by the household directly at the point of receiving a health service. OOPE is a significant indicator to measure financial risk protection offered by the health system and to monitor progress towards universal health coverage (SDG 3.8). When OOPE is a major source of financing, households tend to face financial hardship when seeking health services. Households that cannot afford to pay at point of service, may also forego care. OOPE estimates for India are presented in the National Health Accounts (NHA). These are available for FY 2001-02, FY 2004-05 & consecutively for FY 2013-14 to FY 2018-19. OOPE for NHA India is estimated from household health or consumption surveys conducted by the National Sample Survey Office, Ministry of Statistics and Program Implementation, Government of India (NSSO-MoSPI) and market data available on sale of medical goods. Routine health related administrative data captured by government, insurers, regulators, and providers which is commonly used in OECD countries is not the usual source in India due to limited data availability.

India's NHA estimates show a declining trend of OOPE as a proportion of its total health expenditures (THE). OOPE as % of THE was 70 in FY 2004-05, which declined to 64 in FY 2013-14 and to 48.8 in FY 2017-18 and 48.2 in FY 2018-19. The steep decline observed between FY 2013-14 and FY 2017-18 is debatable, which motivated us to conduct an exploratory analysis of the OOPE trends and methods of estimation in India from survey data. It is understood that there is no change in the NHA-OOPE estimation method during this period. OOPE has been derived from the major data source-the National Sample Survey on Social ConsumptionHealth (NSS-Health) for two data periods (NSS 2014 and NSS 17-18). Items for which OOPE is derived from the consumption expenditure survey data (CES), data was extrapolated from NSS 68th Round (2011-12), as latest data is not available. Market data available on sale of medical goods for both years 2014 & 2017-18 was used. This directed a comparative analysis of the survey methods and instruments used in NSS-Health from the last four rounds (1995-96, 2004, 2014 & 2017-18).

Further NSS-Health 2017-18 data is compared to other surveys conducted during the same period with similar objectives-Longitudinal Ageing Study in India (LASI), 2017-18 and Consumer Pyramids Household Survey-Centre for Monitoring Indian Economy (CPHS-CMIE), January 2014 to April 2018. Between the consecutive NSS-Health rounds, there is no change in survey methods that can significantly explain a decline in the NHA-OOPE estimates. Between the NSS 2014 and NSS 2017-18, there is a decline in the reported proportion of ailing persons (PAP) and hospitalization rate which may explain the decline in OOPE. However, when these indicators are compared for various disease categories, this decline is also noticed for non-communicable diseases. NSS-Health 2017-18 hospitalization rates for persons aged 45 years or more, are lower than those reported in LASI 2017- 18, indicating certain limitations with the NSS 2017-18 in capturing hospitalization rates. A comparative analysis of data from NSS-Health 2017-18 and CPHS-CMIE 2018, indicates that both surveys show majority health spending is incurred by households from the top deciles (richer sections).

However, NSS-Health underrepresents the richer section and CPHS-CMIE might be more equipped to capture the low frequency and high magnitude spending incurred by the top deciles. We understand that the NSS-Health 2017-18 data has limitations with regards to reporting hospitalization rates and representing the health spending by the top deciles, which may have underestimated OOPE. 6 However, to be able to ascertain these findings and exactly suggest where NSS-Health methods are to be strengthened, further study of longitudinal trends of CPHS-CMIE and truncated NSS-Health and LASI data adjusted to age specific morbidity and hospitalization rates is required. Further, exploratory analysis is required to understand the impact of demonetisation and other economic policies implemented during this period, which many researchers believe influenced access to healthcare and utilisation of services in the country.

For accurate NHA-OOPE estimates, in parallel to strengthening the NSS-Health methods, it is suggested to explore pathways to derive estimates using an integrated approach that allows for triangulating NSS-Health data with administrative data from insurers and providers, as recommended by SHA 2011. For estimating OOPE from health facilities, a periodic provider survey can be conducted, as is a practice in most countries reporting health accounts.

To arrive at a sample of providers for the survey, database of healthcare providers available from the National Health Resource Repository (NHRR) and the database of empanelled network hospitals available with Insurance Regulation & Development Authority of India (IRDAI) and the National Health Authority can be used.