The Effect of Prescription Drug Insurance on Utilization and Health among the Elderly

Medicare was first established in 1965, providing near-universal coverage of hospital and physician services for seniors. However, it was not until nearly forty years later that seniors obtained the same widespread access to prescription drug coverage, with the passage of the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 (MMA).

Earlier studies established that prior to the passage of this act, many seniors had difficulty paying for the medications prescribed to them by their doctors. For example, one survey found that more than half of diabetes and heart disease patients without drug coverage skipped doses or failed to fill prescriptions due to cost.

The primary goal of the MMA was to increase access to prescription drugs for the elderly and thereby improve their health. Yet there is limited evidence to support an effect of drug coverage on prescription drug utilization or on health among the elderly.

In "Prescription Drug Insurance and Its Effect on Utilization and Health of the Elderly" (NBER Working Paper 12848), researchers Nasreen Khan, Robert Kaestner, and Swu Jane Lin aim to fill this gap.

The authors' basic approach is to test whether seniors with drug insurance coverage differ from those lacking coverage in their prescription drug use, hospitalizations, and health. The authors use the Medicare Current Beneficiary Data for the years 1992 through 2000 for their analysis. This is a nationally representative survey of Medicare beneficiaries that follows respondents for four years and links self-reported survey data with Medicare claims data.

To account for the possibility that prescription drug coverage is non-random - for example, that those who expect high drug expenditures are more likely to have coverage, or that those who have coverage happen to be a healthier slice of the population than average - the authors estimate models with individual fixed effects. This amounts to asking whether, for a given individual, obtaining or losing drug coverage has any effect on health utilization and health. This approach is valid so long as movements in and out of drug coverage are not related to changes in expected drug utilization or health.

The authors show that such movements in and out of drug coverage are quite frequent - on average, 12 percent of their sample either gained or lost coverage each year. They also examine whether movements into drug coverage are associated with rising drug use or declining health just before the switch is made. They find no evidence that those obtaining drug coverage do so in response to a significant change in drug use or in health.

The authors first look at the effect of drug coverage on prescription drug utilization. They find that having drug coverage is not associated with a significant increase in the probability of having a prescription filled. This is true regardless of the source of the drug coverage - public program, employer-sponsored program, HMO coverage, or Medigap policy. Turning to the annual number of prescriptions dispensed, they find that drug coverage has a significant though moderate effect on utilization. For example, having coverage through a public program increases the number of prescriptions filled by 13 percent, while having coverage through an employer-sponsored or HMO plan did so by 6 percent.

Next, the authors examine hospital admissions, in order to test whether having drug coverage allows beneficiaries to better manage their health problems and avoid costly hospitalizations. There is no evidence of any such effect in the short term, although it remains possible that there is an effect in the longer term, something the authors cannot easily test with their approach.

Finally, the authors explore the effects of drug coverage on health, using both self-reported health status and the ability to perform activities of daily living as measures of health outcomes. They find little evidence that prescription drug coverage is associated with an improvement in these measures of health, although they note that it would also be useful to look at disease-specific outcomes, such as reduction in blood pressure for hypertensive patients.

They authors consider the hypothesis that the effects of drug coverage may be larger for economically disadvantaged individuals, who are more likely to be in poor health and may be less able to purchase drugs in the absence of insurance. However, results for this group are quite similar to those for the entire sample.

In summary, the authors find only a modest effect of prescription drug coverage on drug utilization and no evidence of a beneficial effect of coverage on hospital admissions and health outcomes. The authors suggest several possible reasons for the lack of an effect on health. One is that increased drug utilization could result in a greater number of side effects and adverse events. A second possibility is that some drugs may be prescribed inappropriately or that the beneficiary may not comply with the prescribed therapy. The authors conclude that simply providing drug coverage may not be sufficient to improve health and that "other interventions such as improving prescribing and adherence to medication therapy ... are warranted."


National Bureau of Economic Research, 1050 Massachusetts Ave., Cambridge, MA 02138; 617-868-3900; email:

Contact Us