I just finished reading a new analysis of the decades old RAND Health Insurance Experiment by Jonathon Gruber at MIT. The report has the scintillating, no-nonsense title, “The Role of Consumer Copayments for Health Care: Lessons from the RAND Health Insurance Experiment and Beyond.’’
As a newbie to the world of health care policy, I found it extremely helpful in summarizing why we get so many disparate views of what the HIE proved with respect to copayments – and what it actually did prove.
- Coinsurance Leads to declines in overall health spending.
- Usage falls as Co-insurance increases
- Inpatient care, however, shows less fluctuation within co-insurance levels.
- Coinsurance causes equal declines in all kinds of care – effective and ineffective care.
- ER use is higher when co-insurance is required, than with free care.
- Coinsurance means reductions in preventive care
- Most people, who are ‘typically healthy to start’ aren’t negatively affected (healthwise) by co-insurance – whether they are rich or poor.
- If a person is sick, though, coinsurance has a bad effect. Particularly if they are sick and poor.
- Poor children suffer a larger reduction in effective care than in ‘ineffective care’ when co-insurance is imposed.
- Those with bad health to start with were at significantly higher risk of dying if they had a coinsurance plan.
Gruber does note, however that the study doesn’t factor in the more recent phenomenon of uninsurance. The people who were part of this study had maximum out-of-pocket limits – so there was a limit to how much their pocketbooks would bleed. Unfortunately, Gruber says there is no apples-to-apples study that looks at the current health care schema. To the extent that people have studied it, the most recent data doesn’t measure – as the HIE did – whether the types of care that are skipped with these higher deductible plans are “effective’’ or “ineffective.’’
Clinical care changes a lot in 35 years: Ask yourself how different things were in 1935 than in 1971? No polio vaccine; No antibiotics. Is the degree of difference between 1971 and 2006 as great? In terms of major conditions, such as asthma and diabetes, yes. What we do know, though, is that when the RAND study was done in the 1970s – thirty years ago – clinical care was very different. This is an important point. Gruber mentions the changes in clinical care but does not dwell on them. We now know that effective management of asthma, diabetes, high blood pressure, heart disease and some other chronic diseases reduces avoidable and expensive visits to the doctor or emergency room. Let’s also not forget that it also keeps people healthier, longer.
As a starting point, Gruber takes these lessons from the HIE study on coinsurance:
- High Deductible health plans don’t really get anyone to a good co-insurance model because it applies the same deductible to all income types.
- Out-of-pocket limits need to be related to income.
- Co-insurance should be targeted to places where care is least effective so as not to discourage patients from maintaining their chronic diseases.
Health Access’ lessons
For us, Gruber’s last point is particularly salient. Most patients call “disease management’’ of chronic conditions “preventive care’’ – because their doctor told them to observe a regimen that would prevent them from getting sicker.
Maintenance of these diseases requires many steps – prescription drugs, labs, follow-up doctor visits. Insurance plans, which require co-insurance and co-pays at each point, deter patients from properly maintaining their regimens. It interferes with them taking drugs regularly, from getting follow-up tests, from seeing their physicians for follow-up care. Those who can’t afford the asthma meds, diabetes tests, lab co-pays, heart meds or postpone their follow-up visits, and delay their care because of cost have demonstrably worse outcomes and use more expensive care down the road.