A health insurance model for India

I have written earlier about why health assurance for India that revolves around any insurance model may be fiscally unsustainable. But given the political popularity of health insurance, it may be difficult to put the genie back in the bottle. So what is the least damaging health insurance model? 

I'll propose a five-pronged approach. First, given its fiscal unsustainability, a more realistic compromise may be to ration coverage by limiting insurance to a set of catastrophic medical conditions, ones that contribute the largest share to the country’s high OOP spending. This standard benefits package can be offered as a basic plan by all insurers. Second, it may be useful to consolidate all the public insurance schemes offered to different categories of people under one umbrella, with a basic insurance plan and different types of top-ups, including those which provide premium care. They would include central government schemes like those offered by the Ministry of Labor and Employment, various public sector units for their employees, departments for population categories like weavers, the flagship RSBY, and the various state government schemes. Like in the continental Europe, the basic plan should be community-rated (same premium for everyone in an age cohort, irrespective of pre-existing medical conditions) cover a very basic set of high-incidence catastrophic medical conditions and no more, and have the same premium across insurers within a region.  

Third, instead of unrestrained consumer choice, there should be rigorous enough gate-keeping to ensure both the referral chain and also that it does not become a channel to enrich private multi-specialty hospitals. Patients should not be entertained at tertiary care facilities for simple secondary care treatments and instead be directed to the revamped first referral units. For patients under publicly financed insurance, private facilities should not become the default first choice. Instead, some form of public facility default should be designed, which, while strongly encouraging the use of public facilities should not end up completely eliminating choice. Service standards and cleanliness in public secondary and tertiary care facilities should be improved and their value proposition enhanced. But all this assumes efforts to improve service standards at public secondary and tertiary care facilities. 

Four, such an insurance architecture will provide enormous monopsony power, at the least to the extent of those conditions covered in the basic plan. Strategic buying centered around this plan, done possibly at regional or state levels, can keep costs down. Finally, foster competition among different insurers through transparent purchasing through exchanges. While they would have to offer the basic plan at the same (regional) rates, private insurers should be allowed to differentiate based on the top-ups and the quality of their services. 

But before designing any such plan, it would be necessary to have the requisite administrative and regulatory framework in place for its effective implementation. An independent body like the NICE in the UK, instead of the Ministry of Health and Family Welfare, should be entrusted the responsibility to decide on the basic plan conditions subject to the fiscal constraints. The quality of care delivered should be strictly monitored, with clearly defined treatment protocols and outcomes, through a mix of occupational self-regulation and public regulation. Information asymmetry regarding care and insurance choice and their respective pricing should be bridged through information dissemination efforts revolving around mobile phone apps etc so as to allow informed purchasing. Finally, this opportunity should be used to start developing a robust IT network that can support electronic medical records. Given the concerns about privacy and portability, public policy intervention may be necessary to formulate enabling data management rules, and even support some form of basic IT system that underpins medical records.