NBER Reporter: Spring 2000

Health Care

The NBER's Program on Health Care held its spring meeting in Cambridge on March 10. Alan M. Garber, NBER and Stanford University, organized this program:

Frank R. Lichtenberg, NBER and Columbia University, and Tomas J. Philipson, NBER and University of Chicago, "Can the Public Sector Fine-tune Private R and D? Theory and Evidence from the Pharmaceutical Industry"

Angus S. Deaton, NBER and Princeton University, "Relative Deprivation, Inequality, and Mortality"

Jeffrey Geppert, NBER; Mark B. McClellan, NBER and Stanford University; and Douglas O. Staiger, NBER and Dartmouth College, "Hospital Quality, Treatment Effects, and Health Outcomes: Analyzing Differences in Medical Practices"

Philip J. Cook and Michael J. Moore, NBER and Duke University,

"Environment and Persistence in Youthful Drinking Patterns"

Krista M. Perreira, Duke University, and Frank A. Sloan, NBER and Duke University, "Living Healthy and Living Long: Valuing the Nonpecuniary Loss from Disability and Death"

When the government stimulates R and D, it affects not only current innovators but also their future competitors, whose innovations will make current patents obsolete. Using evidence on all innovations made in the U.S. pharmaceutical industry from 1950-93, Lichtenberg and Philipson demonstrate that the magnitude of the creative destruction of patent-protected profits is significant. They find on average that seven new drugs in the same drug class enter the market before a typical drug's patent expires. Each additional standard-review entrant reduces the sales of the incumbent drug in the same class by about 8 percent. The reduction in the (present discounted) value of the innovator's sales because of the entry of new and similar yet distinct patented products appears to be at least as large as the reduction in sales caused by the entry of generics after a patent has expired. This suggests that creative destruction by new patents is significant and quantitatively important, and limits the innovative returns to a given patent. Hence, the offsetting effects of government R and D policy may be significant.

Deaton presents a model that attempts to integrate the relationship between income and mortality with the relationship between inequality and mortality. He postulates that individual health is affected negatively by relative deprivation within a reference group, defined as the total "weight" of incomes of group members who are better off than the individual. Deaton argues that this model is consistent with what we know about how social status affects health. The theory predicts that within reference groups, which may be as large as whole populations, mortality declines with income, but at a decreasing rate. Whatever the distribution of income, a mean-preserving increase in the spread of incomes raises mortality for everyone. Between reference groups, as between states or countries, mortality is independent of the level of average income, but it depends on income inequality.

Geppert, McClellan, and Staiger devise a method of summarizing between-hospital variation in patterns of medical practice and estimating how these variations are related to patient outcomes. Using this methodology, they analyze the relationship between treatment and mortality among elderly heart attack patients. The authors find that the most important factor associated with lower patient mortality is higher use of aspirin and beta blockers, as well as other drugs shown to reduce heart attack mortality in randomized controlled trials. Significantly lower mortality is also associated with higher use of invasive procedures (for example, angioplasty and bypass). The results are similar when the analysis is based on zipcode-level variation in treatment patterns, which is arguably less biased by differences in patient selection across hospitals. These estimates of treatment effects also are quite similar to McClellan's 1994 results, but are considerably more precise.

Cook and Moore use panel data from the National Longitudinal Survey of Youth to assess the extent to which alcohol control measures influence how much adolescents drink later in life. They confirm that alcohol control measures in the individual's state of residence have a contemporaneous effect on the likelihood of drinking and bingeing. Also, it turns out that the control measures that were in place years earlier, when the individual was only 14, have a discernible effect on current drinking. Finally, Cook and Moore estimate the effect of state per capita consumption of alcohol on drinking and bingeing.

Perreira and Sloan use three approaches to estimate the nonpecuniary cost associated with disability late in life. They also estimate the value of life using a paired risk-dollar comparison. Their data come from interviews with 548 persons; the respondents reported a median value of life of $12 million. They said they would be willing to pay $0.9 million to avoid disability in late life, or approximately $60,400 for each year of disability over age 62. The results stood up to whatever valuation technique was employed.