Cutler, Ghosh, and Landrum 2013 — Evidence for Significant Compression of Morbidity in the Elderly U.S. Population

Summary

Using the Medicare Current Beneficiary Survey (MCBS) linked to death records (1991–2009, N=251,872), Cutler, Ghosh, and Landrum find strong evidence for compression of morbidity in the U.S. elderly population: disability-free life expectancy (DFLE) at age 65 rose by 1.6 years from 1992 to 2005, while total life expectancy (LE) rose by only 0.7 years and years spent disabled fell by 0.9 years. The key mechanism is disease-disability decoupling — disease prevalence was roughly flat or rising, but the disabling impact of each disease fell substantially, particularly for cardiovascular disease and vision problems. This is an NBER working paper (WP 19268); the published version appeared as a book chapter in Wise (ed.), Discoveries in the Economics of Aging (University of Chicago Press, 2014).

Key Claims

• Strong compression confirmed (Table 8): At age 65, 1992–2005: LE +0.7 years, DFLE +1.6 years, disabled LE −0.9 years. The ratio DFLE/LE rose, and years with disability fell absolutely — meeting the strong version of Fries's compression hypothesis.
• Disease-disability decoupling: Conditions became less disabling per prevalence unit. An Oaxaca decomposition attributes 63% of the DFLE gain to cardiovascular disease (2.5 of 7.4 percentage points (pp) disability decline) and vision problems (1.7 pp), predominantly through conditions becoming less disabling rather than becoming less prevalent.
• Time-until-death concentration: Disability is heavily concentrated in the period just before death. Within 12 months of death, ~80% of individuals are disabled; more than 36 months from death, disability rates are much lower and fell over the period. The compression occurred among people not in the final year of life — the end-of-life disability burden is largely unchanged.
• Three-factor structure of health: The 19 health dimensions in the MCBS reduce to three factors: F1 (activities of daily living / instrumental activities of daily living [ADL/IADL] limitations), F2 (functional limitations such as walking and lifting), and F3 (sensory impairments). Factor analyses confirm stable structure across cohorts.
• MCBS is the right data source: The MCBS is nationally representative of all Medicare beneficiaries including the institutionalized, is linked to death records, and covers 10,000–18,000 individuals annually — enabling the time-until-death framing that standard cross-sectional surveys cannot support.

Concepts Introduced or Extended

• Compression of Morbidity
• Healthy Life Expectancy
• Morbidity-Mortality Distinction

Entities Mentioned

• David Cutler
• Kaushik Ghosh
• Mary Beth Landrum
• Michael Chernew
• Eric Stallard

Quotes

"We find evidence for significant compression of morbidity... disability-free life expectancy rose by more than total life expectancy... years with disability fell."

"The primary mechanism driving the decline in disability... is that existing diseases are becoming less disabling, not that the prevalence of disease is falling."

My Take

This paper provides the clearest direct evidence for compression of morbidity in the U.S. elderly using the gold-standard data source (MCBS with mortality linkage) and the most analytically useful framing (time-until-death rather than calendar time). The Oaxaca decomposition — showing disease-disability decoupling rather than prevalence reduction — is the most policy-relevant finding. However, the evidence is strictly confined to age 65+: Global Burden of Disease (GBD) evidence for working ages (20–64) shows the opposite trend. The time-until-death framing also explains why compression and expansion evidence can coexist: compression is happening among the non-dying; the end-of-life burden is stable. See Morbidity-Mortality Distinction for the full evidential picture including the working-age expansion.