Demand For Healthcare: Two randomized health insurance experiments

When was the RAND Health Insurance Experiment conducted?

Between 1974 and 1982 in the United States.

Why was the RAND HIE groundbreaking?

It was the first large-scale randomized experiment to randomly assign health insurance coverage and cost-sharing levels.

Why was RAND important relative to earlier studies?

Before RAND, most studies were non-randomized, leading to conflicting conclusions about whether health care demand responds to price.

What did the RAND HIE ultimately show?

That the demand curve for health care is downward-sloping, not vertical.

Who participated in the RAND HIE?

Around 2,000 families from six American cities

How were participants assigned in the experiment?

Families were randomly assigned to different health insurance plans for several years.

what varied across the insurance plans?

The generosity of coverage, specifically the copayment rate.

What is a copayment rate?

The fraction of the medical bill paid out-of-pocket by the patient.

Why is varying the copayment rate useful?

Because it directly changes the price of care faced by consumers, allowing researchers to observe how usage responds.

What copayment rates were studied in the RAND HIE?

• 0% (free care)

• 25%

• 50%

• 95%

What is a cost-sharing plan?

A plan with a positive copayment rate, where costs are shared between the insurer and the patient.

Why is the RAND HIE ideal for estimating demand?

Because the plans differed only in price, not in other characteristics, ensuring clean causal identification.

What is a major limitation of the RAND HIE today?

The health care system has changed significantly since the 1980

Why was the Oregon Medicaid Experiment conducted?

To provide a more recent test of health care demand under modern conditions.

What general result did the Oregon Medicaid Experiment find?

Like RAND, it found downward-sloping demand for health care.

How did the Oregon experiment differ from RAND in design?

A lottery that randomly selected individuals who were allowed to apply for Medicaid.

Who were the participants in the Oregon Medicaid Experiment?

Low-income adult residents of Oregon.

What were the two comparison groups?

• Lottery winners (allowed to apply for Medicaid)

• Lottery losers (not allowed to apply)

How did the lottery affect prices faced by participants?

Lottery winners were more likely to obtain Medicaid and therefore faced lower out-of-pocket prices.

Was Medicaid coverage automatic for lottery winners?

No. Winners were only about 25 percentage points more likely to be insured than losers.

What is an advantage of the RAND HIE?

It studied a nationally representative population and directly randomized insurance generosity.

What is an advantage of the Oregon Medicaid Experiment?

It reflects modern health care conditions and includes an uninsured comparison group.

What is a disadvantage of the Oregon Medicaid Experiment?

It focuses only on low-income individuals, limiting generalizability

What key question is the RAND Health Insurance Experiment trying to answer here?

Whether higher prices (cost-sharing) for health care affect health outcomes, not just health care use.

Why is this question important?

Because earlier evidence shows that higher prices reduce health care use. The crucial policy question is whether reduced use actually harms health.

Why do researchers look at mortality rates?

Mortality is an extreme and unambiguous health outcome, making it a natural starting point for analysis.

What does the RAND mortality table show for all participants?

Mortality rates are almost identical across insurance plans, regardless of coinsurance.

How similar are mortality rates between free and copayment plans?

Participants on the free plan were 99% as likely to die during the experiment as those on cost-sharing plans.

What does this imply for the average participant?

For the general population in the experiment, higher prices did not increase the risk of death.

Who are considered “high-risk” participants in the RAND study?

: Individuals with conditions such as:

• High blood pressure

• High cholesterol

• Smoking habits

What does RAND find for high-risk participants?

Differences in mortality between free and copayment plans are larger for high-risk individuals.

Why might mortality be an incomplete measure of health?

Because mortality is rare and extreme; many health effects are subtler and may not affect survival.

What alternative outcomes does RAND examine?

A wide range of clinical and functional health indicators, including:

• Lung function (FEV)

• Blood pressure

• Cholesterol

• Glucose

• Vision

• Joint symptoms

How many health outcomes did Newhouse (1993) examine in total?

23 separate health comparisons.

How many of these showed statistically significant differences?

Only 3 out of 23 outcomes.

What does this pattern suggest?

hat most health outcomes are not strongly affected by price, at least over the time horizon of the experiment

What is the main conclusion of the RAND experiment regarding health outcomes?

Health care prices strongly affect utilization but have limited average effects on health outcomes.

What question does Card et al. (2009) investigate?

They study whether increased access to health care affects mortality rates among individuals aged 65 and above.

Why focus on individuals aged 65 and over?

Because at age 65, individuals in the US become eligible for Medicare, leading to a sharp increase in health care access and usage.

What effect does increased health care usage have on mortality?

Card et al. (2009) find that increased access to health care reduces mortality by about 1.1 percentage points.

What is the baseline (average) 28-day mortality rate?

The average 28-day mortality rate is 4.7%

How large is the mortality reduction relative to the baseline?

A 1.1 percentage point reduction corresponds to approximately a 20% decrease in mortality.

Why is this result economically important?

It provides evidence that reducing the effective price of health care can improve health outcomes, especially for older individual

How does this finding compare to the RAND experiment?

Unlike RAND, which finds little effect of price on average health outcomes, Card et al. (2009) shows substantial health gains for a specific high-risk population.

What key lesson does this suggest for health policy?

Lower prices and expanded insurance coverage can save lives, particularly among vulnerable or high-risk groups.

What is the main takeaway in one sentence?

Card et al. (2009) show that increased access to health care at age 65 significantly reduces mortality, indicating that health care prices can meaningfully affect health outcomes.