Seven recent reads on mortality inequality and how to measure it
Recent papers on SSI disability, political ideology, Indigenous maternal mortality, survey measurement, and late-life health show how mortality inequality changes when researchers tighten the population, data source, and mechanism under study.
The most useful new work in this lane is getting less satisfied with broad labels like "race," "region," or "politics." It is trying to pin down where disparities enter the measurement chain, the life course, and the institutions that record death and illness.
Coverage for this read: papers and working papers published or posted in May-June 2026, with a bias toward social demography, mortality measurement, health disparities, and population-health methods.
Start with these seven
| Paper | What it adds | Method to watch |
|---|---|---|
| Madeline S. Helfer, Becky Staiger, and Jessica Van Parys, "Mortality Rates by Race and Ethnicity Among People with Disabilities" | Among low-income adults with disabilities receiving SSI, age-adjusted mid-life mortality is highest for American Indian and Alaska Native recipients at 2.9% and White recipients at 2.6%, followed by Black recipients at 2.3%, Hispanic recipients at 1.9%, and Asian recipients at 1.6%. The striking part is the within-population reversal: in this Medicaid-linked SSI population, Black adults outlive White adults. 1 | Medicaid claims from 2017-2021 are used to compare race and ethnicity groups inside a uniformly low-income, disability-selected population, then test how much diagnosed chronic conditions, despair-related conditions, substance use disorders, and disabling conditions attenuate the mortality gaps. 1 |
| Ezra Elder, "The political polarization of health outcomes in the USA" | In Add Health, conservatives and liberals looked similar on measured biomarkers around 2008-2009, but by 2016-2018 the most conservative respondents had worse biomarker health. By 2020-2022, very conservative respondents had a 1.14 percentage-point higher probability of internal-cause death than very liberal respondents, with a 95% CI of 0.18 to 2.11 and P = 0.021. 2 | The paper links individual-level ideology, biomarker data, and death records in a cohort rather than relying on county-level vote shares. The authors are careful that the findings are descriptive, not causal. 2 |
| Albert Dube and coauthors, "Collecting data on HIV/AIDS mortality during household surveys" | In Malawi, adding HIV questions to standard adult and maternal mortality modules produced nearly complete HIV-status data for deceased siblings. Sensitivity was above 0.75 after accounting for partial verification bias, and specificity was above 0.95. 3 | The randomized validation compares face-to-face interviewing with audio computer-assisted self-interviewing against a demographic surveillance reference system. ACASI did not improve accuracy, but it required more respondent training and collection time. 3 |
| Mengxue Chen, Wen Su, and Vladimir Canudas-Romo, "Life expectancy in China and the contribution of regional dynamics" | National life expectancy gains in China from 2010 to 2020 came from mortality improvements across all regions, but population redistribution did not help everywhere. Only urban-east, urban-central, and rural-west made positive contributions through the composition component; other regions partly offset mortality progress. 4 | The useful move is a decomposition that separates mortality change from changing population exposure to regional mortality regimes, using the Chinese Disease Surveillance Points system. 4 |
| Derlis Duarte-Zoilan and coauthors, "Structural ethnic inequities in maternal mortality between Indigenous and non-Indigenous women in Paraguay" | The preprint identifies 907 maternal deaths from 2014-2023, including 112 among Indigenous women. Indigenous women are overrepresented by a factor of 4.8 relative to their population share, with maternal mortality ratios of 317.7 to 773.6 per 100,000 live births versus 58.7 to 145.1 among non-Indigenous women. 5 | The analysis links ethnic status to territorial and institutional access measures. Indigenous women had more than three times the adjusted odds of dying outside health institutions, aOR = 3.41, 95% CI: 2.20-5.29; obstetric hemorrhage was also strongly associated with Indigenous status, aOR = 3.83, 95% CI: 2.31-6.37. 5 |
| D. Mark Anderson, Kerwin Kofi Charles, and Daniel I. Rees, "Medical Innovation and Racial Health Disparities" | Surfactant replacement therapy reduced respiratory-related neonatal mortality for both Black and White infants, but by 1993 the estimated reduction was larger for White neonates, 46%, than for Black neonates, 30%. The earlier Black-White advantage in respiratory-related neonatal mortality essentially disappeared after FDA approval. 6 | The design uses 1980-2000 vital statistics and treats non-respiratory-related mortality as a counterfactual outcome. The authors argue the uneven effect is not explained by socioeconomic status or distance to the nearest NICU, and point instead to racial differences in fetal pulmonary maturation. 6 |
| Paul Bingley, Nabanita Datta Gupta, Malene Kallestrup-Lamb, and Alexander O.K. Marin, "Health Inequalities Among Danish Retirees 2004-2022" | Higher-income Danish retirees aged 60-79 are healthier across functional, diagnosed, comprehensive, mental, and cognitive domains. But the gradients do not move together: functional and comprehensive health gaps narrow, while mental-health gaps remain large and persistent. 7 | The paper is useful because it refuses to collapse late-life health inequality into one score. It tracks multiple health domains in Danish SHARE data from 2004-2022, which makes convergence in physical health and persistence in mental health visible at the same time. 7 |
The thread across the stack
Three patterns are worth carrying into the next read.
First, disability, retirement, region, and political identity are not just controls. They define the population at risk. The SSI paper changes the usual Black-White mortality comparison by restricting the sample to low-income disabled adults. The Danish retirees paper changes the question again by separating physical, mental, diagnosed, comprehensive, and cognitive health instead of asking whether retirees are simply more or less healthy.
Second, measurement is part of the inequality story. The Malawi study is not only about HIV mortality; it is about whether household surveys can measure HIV-related deaths without relying on a full civil-registration infrastructure. If the data system misses a kind of death, the disparity is already distorted before analysis begins.
Third, mechanisms are getting more specific. The surfactant paper does not stop at "access to care." It tests socioeconomic status and distance to neonatal intensive care, then argues that biology linked to fetal pulmonary maturation better explains the uneven mortality response. The Add Health paper does not claim ideology causes death; it separates changing political composition from within-person health divergence and then treats trust in medical professionals as a candidate mechanism.
What I would read first
If you only have time for two papers, read Helfer, Staiger, and Van Parys on SSI mortality first, then Dube and coauthors on HIV/AIDS mortality measurement. The first shows how a single sample restriction can invert a familiar disparity. The second shows how a small survey-module change can make a hard-to-observe mortality category measurable in low-registration settings.
For methods, keep an eye on decomposition and validation designs rather than headline disparities alone. The strongest papers here are not the ones with the most dramatic gap. They are the ones that make clear which population is being compared, which deaths are observable, and which mechanism the data can actually support.
References
- 1
- 2
- 3Collecting data on HIV/AIDS mortality during household surveys
demographic-research.org
- 4Life expectancy in China and the contribution of regional dynamics
demographic-research.org
- 5
- 6
- 7
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.