Healthcare Financing Challenges and Opportunities to Achieving Universal Health Coverage in the Low- and Middle-Income Country Context

Abstract

Background: In Bangladesh, on an average 62% of total healthcare spending was borne by households through out-of-pocket (OOP) payments annually during 2000- 2015. Because of such high OOP payments, a sizable proportion of households (15.7%) faced catastrophic health expenditure (CHE) and a number of them fell into poverty in 2010. Protecting households from such payments and consequently, the risk of impoverishment are desirable objectives of health systems worldwide. The Sustainable Development Goals (SDGs) resolution emphasized ensuring quality and affordable essential health services through Universal Health Coverage (UHC) by 2030. In order to achieve UHC, the World Health Organization (WHO) recommends to ensure the protection against the risk of large healthcare payments or CHE by spreading the risk among the population through pre-payments e.g., tax, social security contribution, insurance premium. Informal workers in the agricultural and non-agricultural sectors including readymade garments (RMG) workers constitute a large proportion of the total labor force (88%), who contribute to 64% of the total Gross Domestic Products of Bangladesh. Efforts should, therefore, be made to ensure sustainable quality healthcare for this group of workers by bringing them under pre-payment health schemes. Community-Based health insurance (CBHI) and employer-sponsored health insurance (ESHI) schemes were thus piloted among selected informal workers with an aim to increase utilization of medically trained healthcare providers (MTPs) at an affordable price. Objectives: The main objective of this dissertation is twofold: firstly, to study the effect of the current healthcare financing system on the financial risk of households and secondly, to explore potential solutions through pre-payments schemes (CBHI and ESHI) for mitigating such challenges. Methods: Based on both primary and/or secondary data, five studies were conducted. In study I, nationally representative Household Income and Expenditure Survey, 2016 has been used which provide data on household consumption expenditure including health expenses. We calculated the incidence of CHE, which was later predicted by demographic and socio-economic characteristics of the households using multiple regression analysis. The incidence of CHE was defined as the proportion of households having healthcare expenditure of more than a threshold level such as 10% of their total consumption expenditure or 40% of their non-food consumption expenditure. We estimated the impoverishment effect of OOP payments using both the national (cost of basic need approach) and the international (1.90 International dollar per person per day) poverty line. For study II, 557 informal workers were surveyed during 2010-11 in three geographic locations (a metropolitan city, a district town and a sub-district area) to estimate the willingness-to-pay (WTP) for CBHI, using the contingent valuation method. The association between WTP and demographic characteristics was measured by employing the log-normal regression model. Study III adopted a case-control design to estimate the effect of the CBHI scheme on healthcare utilization from MTPs. We, therefore, surveyed 1,292 (646 insured and 646 uninsured) households after 1 year of implementation of the scheme. In order to minimise the unobserved baseline differences between the insured and uninsured groups, a propensity score matching was performed. A multilevel logistic regression model was applied to measure the association between MTP healthcare use and CBHI membership, in comparison to uninsured. Using the same design in study IV, a two-part regression model was applied to assess the relationship between CBHI membership and the OOP expenditure (probability and magnitude) when adjusted for other confounding factors (demographic and socio-economic). Study V utilized a case-control design with cross-sectional pre-and post-intervention surveys among workers from 7 purposely selected RMG factories (6 intervention and 1 comparison factories) in Safipur of Gazipur, Bangladesh. Randomly selected RMG workers were interviewed in pre-(October 2013) and post-intervention phases (April 2015) from insured and uninsured RMG factories. In total, 1,924 workers were interviewed (480 from the insured group and 482 from the uninsured group in pre- and post-intervention periods). We estimated the difference-in-difference (DiD) of the utilization of healthcare and OOP expenditure. The DiD is a counterfactual estimate derived by measuring the change in outcomes in the intervention group, which is deducted from the change in outcomes in the comparison group between the pre- and post-intervention periods. Beside DiD estimation, we used a two-part regression model to measure the association between OOP payments and membership of the ESHI scheme while controlling for workers’ demographic and socio-economic characteristics. Results: Study I found that CHE were faced by 24.6% of households at the 10% threshold level, the incidence was 25.3% and 22.0% among the poorest and the richest households, respectively. The poverty rate rose by 5.5% (9.0 million individuals) due to OOP payments. In study II, we observed that approximately 87% of the informal workers were willing to pay for the CBHI. The average weekly WTP was 22.8 BDT [95% confidence interval (CI): 20.9–24.8] or 0.32 USD. Monthly income, occupation, geographic location and educational level were the main determinants of WTP. Study III suggested that the insured of CBHI were 2.111 (95% CI: 1.458- 3.079) times more likely than uninsured to use MTP for healthcare. Applying the two-part regression model in study IV, we found that in comparison with the uninsured, the average OOP payment was 6.4% (p<0.001) smaller among the insured for such healthcare utilization. Nonetheless, no significant difference was observed in OOP payments for the health service utilization from all types of providers, i.e., both MTPs and non-trained providers though the latter one was not included in the benefit package of the scheme. Study V showed that the ESHI scheme has resulted in a significant 26.1% escalation in the utilization of healthcare (DiD=26.1; p<0.01) from MTPs among the insured relative to uninsured. When accounting for covariates, such utilization fell to 18.4% (p<0.05). The DiD calculation showed that OOP spending for insured group decreased by -3,700 BDT and -1,100 BDT in comparison to uninsured group while utilized MTPs or all types of providers respectively, although not statistically significant. Conclusions: Reliance on OOP payments for healthcare leads to financial hardship and a challenge for securing financial protection to achieve UHC in low- and middle-income country settings with a large informal sector, like in Bangladesh. To mitigate the challenge of healthcare utilization at lower OOP payments, preppayment schemes such as CBHI and ESHI, are useful for increasing utilization of healthcare from MTPs by both informal and RMG workers. These schemes are in considerable demand that was supported by the WTP findings. However, the insured of the CBHI scheme had a significantly lower OOP payment, while worker insured by ESHI did not experience such reduction. Broader healthcare provider networks of ESHI schemes would reduce dependency on external providers (not contracted by ESHI) and consequently reduce OOP payments while increasing utilization of services.

Type
Publication
Karolinska Institutet