UCLA Fielding School-led study of amenable mortality in Brazil provides new insight into preventable deaths

New research shows that lower income Brazilians, even though they have good access to primary healthcare, are more vulnerable to mortality from amenable causes.

The results are among the findings of first-ever individual-level study of amenable mortality, published Aug. 10 in the peer-reviewed Journal of Aging and Health as “Amenable Mortality Among Older Brazilians: Evidence from the Brazilian Longitudinal Study of Aging (ELSI-Brazil).” The research also found that income, sex, partnership, smoking status, and chronic disease burden, rather than geography or reported access to care, separate which older adults die from preventable causes from those who do not.

“Brazil has built one of the world’s largest primary healthcare systems as part of the Unified Health System, or SUS in Portuguese, and most people in our nationally representative study said they can get care when they need it,” said UCLA Felding’s Dr. James Macinko, professor in the departments of Health Policy and Management and Community Health Sciences, and lead author of the study. “This means the next priority for the SUS must be to enhance the quality, continuity, and coordination of the care that people are already receiving.”

Macinko’s co-authors include Dr. Hiram Beltran-Sanchez, professor in UCLA Fielding’s Department of Community Health Sciences, as well as researchers from the Federal University of Minas Gerais (Brazil) and the University of Illinois. They used data from the Brazilian Longitudinal Study of Aging (ELSI-Brazil), a nationally representative cohort of adults aged 50 and older, and linked participants’ records to Brazil’s national mortality registry (SIM). The study followed 9,412 participants for approximately six years, from 2015 through 2021, and analyzed the 872 deaths with an assigned cause of death.

“Brazil is one of only a handful of countries with a longitudinal aging cohort such as ELSI-Brazil,” Macinko said. “This unique national resource continues to inform public policies that can lead to healthier aging and wellbeing in Brazil and beyond.”

Overall, the researchers found that between 30% to 39% of deaths among Brazilians aged 50 and older are from causes that should not occur when people receive timely, effective and accessible healthcare. The analysis focuses on amenable mortality, sometimes called avoidable mortality, meaning deaths from conditions that should generally not be fatal if people receive appropriate medical care throughout their lifetimes. It is often used as an indicator of overall health system performance.

“This study is the first to examine amenable mortality in Brazil at the individual level,” Beltran-Sanchez said. “Previous research has compared rates across Brazilian municipalities or states, but those aggregate studies could not break out the characteristics of the individuals who died from amenable causes.”

Overall, the study found that amenable deaths occurred at a rate of between 5.6 and 7.3 per 1,000 people per year, depending on the definition used as compared to all-cause mortality of 18.6 per 1,000 per year.

The researchers compared two ways of classifying these deaths: Brazil’s list of avoidable causes, which reflects preventive and curative interventions available through the Unified Health System (SUS), and the definition used by the Pan American Health Organization (PAHO) for comparisons across the Americas.

The PAHO definition classified 39% of deaths as amenable, while the Brazilian list classified 30%. Brazil’s list is broader and includes injuries, violence, alcohol-attributable conditions and endemic infectious diseases such as dengue and Chagas disease but is limited to deaths before age 75. The PAHO list applies no upper age limit but excludes most external causes and several acute conditions. The difference between the two frameworks was largest at the oldest ages-- among adults 80 and over, the PAHO-based rate was 52% higher than the Brazilian one. Despite this, the two definitions resulted in similar conclusions about who is at risk.

Steep inequalities by age, sex, partnership status, income, and smoking

Age was by far the strongest predictor of mortality from all causes. Adults aged 80 and over faced roughly seven to nine times the risk of amenable death of adults aged 50 to 59, as expected.

Women had roughly 40% to 50% lower risk than men, and people living with a spouse or partner had about 40% lower risk than those without one. The authors suggest that partners may help people recognize symptoms, seek care, manage medications and navigate the health system, and that older women access healthcare more often than do men of similar ages.

Income was another important factor. Adults in the highest third of household income had 28% to 36% lower risk of amenable death than those in the lowest third, after adjusting for age, sex and partnership status. Adults in the middle third were not significantly different from the poorest, suggesting the heaviest burden of preventable death is concentrated among the economically vulnerable.

Having two or more chronic conditions was associated with approximately 48% to 50% higher risk under both definitions. The authors argue this reflects less a failure to diagnose or treat any single condition than a broader weakness in delivering coordinated, person-centered care to older adults with several illnesses at once.

People who had ever smoked had a 58% higher risk of amenable death under the PAHO definition. Under the Brazilian definition, the same association was weaker and not statistically significant. When the analysis was restricted to deaths before age 75 (the age ranges the Brazilian list was designed for) smoking again emerged as a strong predictor. The authors identify this as the most important difference between the two classification schemes and caution that studies applying Brazil’s national list without the <75-year age restriction may understate tobacco’s contribution to preventable death.

What the study did not find

Several widely assumed patterns about mortality, specific to Brazil, did not appear in this study and are noteworthy. There was no statistically significant difference in amenable mortality risk between residents of the wealthier South and Southeast and residents of other regions, and no significant difference between urban and rural residents. Early childhood deprivation, such as going hungry, lacking water, or long illness in childhood was not associated with amenable death in adulthood.

Notably, self-reported access to primary healthcare was not associated with lower amenable mortality under either definition. The authors note that access was already high across the cohort: 84% had seen a doctor in the previous 12 months and 61% said they could always easily reach their usual source of care. Notably, all of the individuals in the cohort were born and grew up before Brazil’s health system, the SUS, even existed, meaning that opportunities for early prevention for these cohorts may have been lost.

Implications

For the researchers, the pattern points to a health system whose remaining challenge is less about access than coordination of care as well as a need to continue monitoring the situation as the Brazilian population continues to age at one of the fastest rates in the world. The authors conclude that reducing amenable mortality will require both improvements in geriatric competencies, longitudinal care relationships and active case-finding within the Family Health Strategy as well as action on the social conditions that place some older Brazilians at greater risk in the first place.

Limitations

All predictors were measured once, at the 2015–2016 baseline, and treated as fixed, so changes in income, smoking status or chronic disease burden during follow-up were not captured. About 10% of decedents had no specific cause of death assigned and were excluded. A further 251 deaths had non-informative “garbage codes” as cause of death and were reallocated using standard imputation methods. Mortality rates in the study are nationally representative. Because of the limited number of deaths, risk factors were tested one at a time rather than simultaneously, so the study did not estimate the combined risk facing a person with several of these characteristics at once. 

About the study

Macinko, J., Beltrán-Sánchez, H., Andrade, F. C. D., Passarelli-Araujo, H., & Lima-Costa, M. F. (2026). Amenable mortality among older Brazilians: Evidence from the Brazilian Longitudinal Study of Aging (ELSI-Brazil). Journal of Aging and Health. https://doi.org/10.1177/08982643261476542

The authors are affiliated with the UCLA Fielding School of Public Health (Departments of Health Policy and Management, and Community Health Sciences); the School of Social Work, University of Illinois at Urbana-Champaign; the Center for Development and Regional Planning (Cedeplar), Federal University of Minas Gerais; and the Public Health and Aging Study Group, Oswaldo Cruz Foundation and Federal University of Minas Gerais, Belo Horizonte.

The first two waves of ELSI-Brazil were supported by the Brazilian Ministry of Health. The funders had no role in study design, data collection, analysis, interpretation, writing, or the decision to submit for publication. The authors declare no conflicts of interest. The study was approved by the Research Ethics Committee of the Fundação Oswaldo Cruz, Minas Gerais (CAAE 34649814.3.0000.509). ELSI-Brazil baseline data are publicly available at https://elsi.cpqrr.fiocruz.br/en.