Case and Deaton 2015 — Rising Morbidity and Mortality in Midlife Among White Non-Hispanic Americans in the 21st Century

Summary

Case and Deaton document that all-cause mortality among white non-Hispanic Americans aged 45–54 reversed its long secular decline after 1998 and rose at approximately $0.5\%$ per year through 2013, while mortality continued falling at $\approx 2\%$/yr in peer high-income nations and in other U.S. racial/ethnic groups. The reversal is driven by three cause-of-death clusters — drug/alcohol poisoning, chronic liver disease/cirrhosis, and suicide — concentrated overwhelmingly in individuals with a high school education (HS) or less. Alongside the mortality reversal, the paper documents parallel deterioration in self-reported health status, pain, functional limitations, and inability to work, pointing to worsening morbidity as the upstream condition.

Key Claims

• All-cause mortality for U.S. white non-Hispanic 45–54 year-olds rose after 1998 ($+0.5\%$/yr); prior trend was $-2\%$/yr; all peer countries continued improving.
• If the 1998 mortality level had merely held flat: $96{,}000$ additional deaths through 2013. If the prior downward trend had continued: $488{,}500$ additional deaths — comparable in scale to U.S. AIDS epidemic deaths through mid-2015.
• Three causes drove the reversal: drug/alcohol poisoning (overtook lung cancer as a cause in 2011), suicide, chronic liver disease and cirrhosis.
• Drug poisoning mortality specifically rose $9.3$ per $100{,}000$ for white non-Hispanic 45–54, 1999–2013.
• Mortality change 1999–2013 per $100{,}000$: White non-Hispanic $+33.9$; Black non-Hispanic $-214.8$; Hispanic $-63.6$ — reversal is racially specific.
• Education gradient is steep: HS or less $+134$ per $100{,}000$; some college $-3.33$; BA+ $-57$. Low:high ratio widened from $2.6$ (1999) to $4.1$ (2013).
• Morbidity worsened in parallel: share reporting excellent/very good health fell $6.7$ percentage points (pp) to $55.9\%$; fair/poor health rose $4.3$ pp; chronic joint pain at $34.7\%$ in 2011–13; unable to work up $4.5$ pp (near-doubling) to $9.2\%$; activities of daily living (ADL) difficulties up $2$–$3$ pp; at-risk heavy drinking up $1.7$ pp; elevated liver enzymes increased.
• The morbidity reversal affected both obese and non-obese whites, and both smokers and non-smokers in some indicators — ruling out obesity and smoking as complete explanations.
• Corroboration from multiple surveys (Behavioral Risk Factor Surveillance System (BRFSS), National Health Interview Survey (NHIS), National Health and Nutrition Examination Survey (NHANES), Health and Retirement Study (HRS)) across different measures reduces reporting artifact risk.

Concepts Introduced or Extended

• Deaths of Despair
• Morbidity-Mortality Distinction
• Income-Mortality Gradient
• DI Beneficiary Mortality

Entities Mentioned

• Anne Case
• Angus Deaton
• Raj Chetty

Quotes

"The change in the mortality rate for US white non-Hispanics 45–54 is so large that it overwhelmed the population-wide declines that resulted from continued progress against cardiovascular disease mortality."

"Half a million people are dead who would not be dead if the 1978 to 1998 trend had continued."

"Poisonings, which include drug overdoses, overtook lung cancer as a cause of death in this age group in 2011."

"The increased morbidity among midlife white non-Hispanics is multidimensional: they report more pain, worse self-assessed health, more difficulty with activities of daily living, greater inability to work, and more mental health problems than previous cohorts."

My Take

The paper's core mortality finding is unambiguous and replicated across multiple data sources. Its interpretation — a common social driver (declining economic prospects for the less-educated working class) rather than three independent epidemics — is persuasive given the simultaneity and demographic clustering, but remains a hypothesis; Case and Deaton do not establish causality. The morbidity data (Table 2) are particularly important for Disability Insurance (DI) research because "unable to work" nearly doubled, directly corroborating the genuine morbidity thesis against the pure moral hazard framing of DI growth. The methodological tension with Chetty et al. (2016) is real but resolvable: they measure different populations (race-pooled vs. race-specific), different outcomes (life expectancy (LE) at 40 vs. age-specific death rates at 45–54), and different time windows; the two findings tell a consistent story of widening inequality with a specific racial/educational channel.