The challenge with health insurance model to universal health coverage

In the context of PPPs and the recently unveiled health insurance program in India, Karthik asks the questions (in the comments here),

If the decision of public vs. insurance model of health care is a matter of comparison between the capacity to run public hospitals vs. capacity to regulate private players... are we stuck in a trap where we don't have either of these capacities? In this context, which capacity is easier to build from the current state?

It is a good opportunity to clarify my own thoughts. I naively started out as a strong believer in the insurance model to deliver universal health coverage - see this and this. I no longer subscribe to this view. 

Here is the challenge. Any insurance model is inherently suited to address secondary and especially tertiary care. It cannot be tailored to the effective delivery of non-curative preventive and primary care, especially given the public health challenges that we face. And the cost-effectiveness of insurance by its very nature depends on the ability to limit incidence of the insured events. But the incidence of secondary and tertiary treatment episodes depends on the strength of preventive and primary care. In other words, you need to build the insurance system on a very good preventive and primary care system so that disease incidence itself is minimised. But our preventive and primary care is broken. Worse still, we have a situation where the insured are the poor who are also those with the weakest preventive and primary care and therefore the population category with the highest likelihood of incidence of the insured events. In the circumstances, the insurer faces the maximum disease incidence likelihoods. 

In fact, there is an even greater practical problem with the focus on insurance approach. It takes the attention and resources away from the more important and difficult task of fixing the broken preventive and primary care system. This is not amenable to electorally popular announcements or administrative actions like empanelling insurers, Third Party Administrators (TPAs), and hospitals. It demands persistent and painstaking work which is diffuse and beyond the abilities of systems with weak state capacity. We fall into an even worse equilibrium.

Finally, there is the impossible fiscal challenge. I have written with Lant Pritchett about the problem of doubly-universal coverage - population coverage and conditions coverage - which is an inevitable slippery slope with insurance models. The move from the limited RSBY to Ayushman Bharat is only the latest example. Andhra Pradesh's Aarogyasri Program started with a token amount for just child heart surgeries and in less than a decade (by 2014-15) came to progressively cover nearly 1100 conditions and over Rs 5000 Cr for the combined state. The only reason it did not reach Rs 10,000 Cr was because the supply-side to deliver secondary and tertiary care could grow only so fast. 

It is a only a matter of time before the fiscal limits start to bind as supply expands inexorably to meet the latent demand. The squeeze on preventive and primary care was more or less proportionate to the ballooning of the insurance costs. Even a very very modest health insurance scheme that can reasonably address the secondary and tertiary care needs would easily cost 3-4% of the GDP (if not much more), a tripling or quadrupling of current expenditures. This is a clearly unrealistic expectation for a country grappling with a 11-12% tax-to-GDP ratio. 

So here is the verdict on the insurance model from one of the most authoritative sources, N Srikant, the most outstanding CEO of Aarogyasri who tried to salvage the bloated and expensive corporate give-away that Aarogyasri had become,

We found that adding an additional layer of insurance intermediary between the trust and hospitals reduced the benefit cost ratio under the scheme by 12.2 % (p-value = 0.06). Every addition of 100 beds under the scheme increases the scheme payments by US$ 0.75 million (p-value < 0.001). The gap in claim denial ratio between insurance and trust modes narrowed down from 2.84 % in government hospitals to 0.41 % in private hospitals (p-value < 0.001)... The scheme is a classic case of Roemer's principle in operation. Introduction of insurance intermediary has the twin effects of reduction in benefit payments to beneficiaries, and chocking fund flow to government hospitals. The idea of engaging insurance intermediary should be abandoned.

There is no country in the world which has developed an effective universal health coverage by focusing predominantly on secondary and tertiary coverage based insurance. 

One approach which has the potential to be effective is the capitation model followed in countries like Thailand. But this in turn too requires a strong primary health care system. 

So back to Karthik's questions. The insurance model is simply unsustainable BIG TIME for a country like India. In fact, it is perhaps the worst scenario. So we are left with no option but focus on preventive and primary care and improve public facilities. Even with the weak state capacity, this is just the only long-term way forward. This is my theoretical assessment. 

But since the insurance genie is out of the bottle and given its powerful electoral attractions, some form of insurance cannot be avoided. In fact, in Andhra Pradesh, it is widely believed that Aarogyasri was one of the main reasons for Mr Y S Rajasekhara Reddy retaining power in the 2009 elections. In the circumstances, faced with electoral battles, no government can be faulted for thinking along these lines. So a prudent compromise may be to have an insurance model which covers a basic package of catastrophic illnesses. It should be complemented with some of the following 

1. This insurance should be operated by a Trust with its dedicated TPAs and not insurers. 

2. A high quality IT system that can manage the logistics of screening, pre-authorisation, treatment, payments, and follow-up, which was a feature of Aarogyasri, should be replicated. It should be accompanied by analytics and vigilance to monitor the problem of over-diagnosis and over-treatment. 

3. The public hospitals and government doctors should be incentivized, even with positive discrimination, to attract patients, so that the insurance program does not end up being a give-away to private hospitals.  

4. The rates should be fixed on very objective considerations, free from political interference, and through transparent process of price discovery involving strategic purchasing at an appropriate administrative level. Maybe this should vary across States. 

5. There should be a mechanism to limit the creeping expansion of conditions coverage. An institutionalised arrangement like the National Institute for Health and Care Excellence (NICE) in the UK should be in control of such decisions. 

6. It is useful to demand some limited co-payment or small premium from all but the poorest. 

7. Finally, the expenditure on this insurance cannot come at the expense of primary and preventive care as well as investments in public secondary and tertiary facilities.

If it is decided to purchase from insurers (instead of using the Trust model), there is perhaps some logic in moving away from the current one-premium-for-all model to one which uses different premia for different age-groups, albeit the same for all members in an age-group irrespective of pre-existing medical conditions (community-rating).

All easier said than done! And even if done there is nothing to stop them being changed or reversed. Therefore, the best that can be done for bureaucrats to try incorporate these elements to the extent possible to the insurance program design.