Mortality rates for women (A) and men (B) aged 20 to 64 years are shown by population group. The US groups are highlighted with lighter shading.
Graphs show AAPCs in all-cause mortality rates for women (A) and men (B) aged 20 to 64 years by population group. Error bars represent 95% CIs. The US groups are highlighted with lighter shading.
eTable 1. Selection of Latin American Countries and Puerto Rico, and Reason of Exclusion
eTable 2. Age-Standardized Premature Mortality (Age 20-64 Years) Rates per 100 000 (Rates in 2015 Compared With Rates in 2001), and Trends During 2001-2015, by Cause, Population Groups, and Sex
eTable 3. Life Expectancy in Latin American Countries, Puerto Rico, and the United States Including Countries Excluded From This Study
eFigure 1. Trends of All-Cause Age-Standardized Premature Mortality Rates per 100 000 During 2001-2015, by Age Groups (20-34, 35-49, 50-64 Years), Population Groups, and Sex
eFigure 2. Age-Standardized Mortality Rates (ASMR) per 100 000 (Age 20-64 Years) in 2015 Stratified by Cause, Population Group, and Sex (US Data Highlighted in Red)
eFigure 3. Absolute Change in All-Cause and Cause-Specific Age-Standardized Mortality Rates (Rates in 2015 vs rates in 2001), Stratified by Cause, Population Group, and Sex
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Chen Y, Freedman ND, Rodriquez EJ, et al. Trends in Premature Deaths Among Adults in the United States and Latin America. JAMA Netw Open. 2020;3(2):e1921085. doi:10.1001/jamanetworkopen.2019.21085
What are the premature mortality rates for US Latino populations compared with those of other US populations and Latin American countries?
In this descriptive cross-sectional study of 16 populations, premature mortality rates were lower among US Latino populations compared with US African American and white populations. The overall favorable mortality trends in 12 Latin American countries were comparable with those of US Latino populations and, in many cases, better than those of US white populations.
Premature mortality rates are lower among US Latino populations and in several Latin American countries than among US white populations, which may be indicative of a broad “Latino or Hispanic paradox.”
Premature death rates vary in the United States by race/ethnicity. Despite their socioeconomic disadvantages, US Latino populations have lower premature mortality rates than do US white populations, a phenomenon termed the “Latino or Hispanic paradox.”
To investigate whether there is a broader Latin American paradox by comparing premature mortality rates in the United States according to race/ethnicity with rates in Latin America and Puerto Rico from 2001 to 2015.
Design, Setting, and Participants
This descriptive cross-sectional study used mortality data from the World Health Organization Mortality Database. All deaths occurring in individuals aged 20 to 64 years among US Latino, African American, white, and Puerto Rican and 12 other Latin American populations from January 2001 to December 2015 were selected. The data analysis began in October 2018.
Age, sex, race/ethnicity, and country.
Main Outcomes and Measures
All-cause mortality, cause-specific mortality, age-standardized mortality rates (AMSRs), and average annual percentage change in mortality rates during 2001 to 2015.
During 2001 to 2015, 22 million deaths (8 million women and 14 million men) occurred among individuals aged 20 to 64 years in the selected populations. Among women, US Latina individuals had the lowest premature mortality rates (ASMR for 2015, 144 deaths per 100 000 population) and US African American women had the highest premature mortality rate (ASMR for 2015, 340 deaths per 100 000 population) of the 16 populations studied. Rates among US white women shifted from the sixth lowest in 2001 (ASMR, 231 deaths per 100 000 population) to the 12th lowest in 2015 (ASMR, 235 deaths per 100 000 population). Among men, Peru had the lowest premature mortality rates (ASMR for 2015, 219 deaths per 100 000 population), and Belize had the highest premature mortality rates (ASMR for 2015, 702 deaths per 100 000 population). White men in the United States shifted from the fifth lowest rates in 2001 (ASMR, 396 deaths per 100 000 population) to the eighth lowest rates in 2015 (ASMR, 394 deaths per 100 000 population). Rates for both women and men decreased in all the populations studied from 2001 to 2015 (average annual percentage change range, 0.4% to 3.8% per year) except among US white populations, for which the rate plateaued (average annual percentage change, 0.02% per year [95% CI, −0.3% to 0.2% per year] for women; −0.2% per year [95% CI, −0.4% to 0.0% per year] for men) and among Nicaraguan men, for whom the rates increased (0.6% per year [95% CI, 0.2% to 1.0% per year]). The populations with the lowest mortality rates in 2015 had lower rates from all major causes, but rates were particularly lower for heart disease (21 deaths per 100 000 population) and cancer (50 deaths per 100 000 population).
Conclusions and Relevance
Premature mortality rates are lower for US Latino populations and several Latin American countries than for US white populations, suggesting that there may be a broader Latin American paradox. This analysis also highlights the high premature mortality rates among US African American populations, especially women, compared with many Latin American populations.
Premature mortality rates vary substantially in the United States by race/ethnicity.1 Despite their socioeconomic disadvantages,2 US Latino populations have lower premature mortality rates than do US non-Hispanic white populations (hereafter referred to as “US white populations”).3 This phenomenon has been termed the “Latino or Hispanic paradox” and has been suggested to be associated with the “healthy immigrant effect,” where healthier individuals come to the United States to work, or to the “salmon bias,” where immigrants with poor health return home at the end of life. However, approximately two-thirds of current US Latino individuals were born in the United States, and US Latino populations have continued to show lower mortality rates than US white populations.1,4 One theory that has not been studied, to our knowledge, is whether premature mortality rates are lower more generally in Latin America.
The aim of this study was to compare recent trends and rates of premature mortality among US Latino, white, and African American populations with those in Latin America and Puerto Rico for all-cause and major causes of death. Monitoring rates and trends in premature mortality is also important more generally to understand how societies can address preventable causes of early death and to highlight priorities needed to achieve the United Nations Sustainable Development Goals and Targets of a one-third reduction in premature mortality from noncommunicable diseases by the year 2030.5
Because we used publicly accessible data without personal identifying information, institutional review was not required, in accordance with 45 CFR §46.102(f). This study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
For this cross-sectional analysis, we used data from the World Health Organization (WHO) Mortality Database and the Multiple Cause of Death database from the Centers for Disease Control and Prevention. The WHO mortality database documents mortality data by age, sex, and cause of death, as reported annually by member countries from their civil registration systems, but it does not have race/ethnicity data for each member country. Therefore, to obtain US mortality data by racial/ethnic groups, we used the Multiple Cause of Death database from the Centers for Disease Control and Prevention, which documents deaths that occur in the United States compiled from death certificates filed in state vital statistics offices. Both the WHO Mortality Database and the Multiple Cause of Death database used the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) coding system during the study period.
Only populations with high-quality mortality data were eligible for inclusion, defined here as less than 10% of deaths with no specific intent coded (ie, ill-defined codes, ICD-10 codes R00-R99).6 To examine trends, we required annual data from 2001 to 2015. Thirteen countries were eligible for analysis: the United States, Argentina, Belize, Brazil, Chile, Colombia, Cuba, Ecuador, Mexico, Nicaragua, Panama, Peru, and Uruguay (accounting for >80% of the population of Latin America). eTable 1 in the Supplement summarizes the quality assessment and the reasons for exclusion. Rates for the United States were assessed separately for Latino, African American, white, and Puerto Rican populations, for a total of 16 eligible populations.
We extracted mortality data for the 12 Latin American countries and Puerto Rico from the WHO Mortality Database by region, year, 5-year age group, sex, and cause of death between 2001 and 20157 and extracted population data from the United Nations World Population Prospects 2017.8 Uruguay was missing mortality data for 2011, and so this year was estimated using the mean from 2010 and 2012. We extracted US mortality data and population data for Latino, non-Hispanic African American, and white populations from the Multiple Cause of Death database of the Centers for Disease Control and Prevention.9
The main outcome of interest was all-cause premature mortality occurring from ages 20 to 64 years. To compare trends and patterns of cause-specific premature mortality, we assessed deaths attributed to 6 major causes, including neoplasms (ICD-10 codes C00-D48), endocrine disorders (ICD-10 codes E00-E88), circulatory diseases (ICD-10 codes I00-I99), respiratory diseases (ICD-10 codes J00-J98), digestive diseases (ICD-10 codes K00-K92), and external causes (ICD-10 codes V01-Y98). Within these 6 major disease categories, the leading causes are uniform across countries. For example, heart disease and stroke account for most deaths within the circulatory disease category; malignant neoplasms (cancer) account for most deaths within the neoplasm category; influenza and pneumonia and chronic respiratory diseases represent most deaths within the respiratory disease category; diabetes accounts for most deaths within the endocrine disorders category; cirrhosis of the liver accounts for most deaths within the digestive diseases category; and injuries, self-harm, assault, and accidental deaths account for most deaths within the external causes category.10
We estimated age-standardized mortality rates (ASMRs) per 100 000 population annually by sex and population for all-cause and cause-specific premature mortality. We used the standard US population in 2000 as the reference for age standardization. Because ASMRs are based on population numbers, estimates are free from sampling variability, and statistical tests for between-group differences are not needed. Rate estimation was conducted using SEER*Stat statistical software version 8.3.6 (National Cancer Institute).11 To examine changes in mortality rates and to identify causes of death contributing to these changes, we estimated the absolute difference in age-standardized all-cause and cause-specific mortality rates in 2015 compared with rates in 2001. We then estimated the average annual percentage change (AAPC) in mortality rates for the periods 2001 to 2015 using the Joinpoint Regression Program statistical software version 220.127.116.11 (National Cancer Institute). We estimated the AAPCs as a weighted mean of annual percentage changes in rates, with the weights equal to the length of each segment over the fixed interval using Joinpoint.12
We compared trends in rates for 3 broad age groups (20-34, 35-49, and 50-64 years). Because Latino populations of Mexican, Puerto Rican, and Cuban origin or descent are the 3 largest country-of-origin groups in the United States, composing 63%, 10%, and 4%, respectively, of the total US Latino population in 2016,4 we also compared premature mortality among other US Latino populations with these 3 populations.
During 2001 to 2015, 22 million deaths (8 million women and 14 million men) occurred among individuals aged 20 to 64 years in the selected populations. Among women, US Latina individuals had the lowest premature mortality rate (ASMR for 2015, 144 deaths per 100 000 population), and US African American women had the highest rate (ASMR for 2015, 340 deaths per 100 000 population) of the 16 populations studied from 2001 to 2015. Women in several Latin American countries, including Peru (ASMR, 147 deaths per 100 000 population), Chile (ASMR, 166 deaths per 100 000 population), and Panama (ASMR, 186 deaths per 100 000 population) had lower mortality rates than US white women, who had the sixth lowest rate in 2001 (ASMR, 231 deaths per 100 000 population) but by 2015 had shifted to the 12th lowest rate (ASMR, 235 deaths per 100 000 population) (Figure 1). Among men, Peru had the lowest premature mortality rate (ASMR for 2015, 219 deaths per 100 000 population) followed by US Latino men (ASMR for 2015, 278 deaths per 100 000 population), and Belize had the highest premature mortality rate (ASMR, 702 deaths per 100 000 population). Mortality rates for US white men ranked fifth lowest in 2001 (ASMR, 396 deaths per 100 000 population) and moved to eighth lowest in 2015 (ASMR, 394 deaths per 100 000 population) (Figure 1).
Between 2001 and 2015, premature mortality rates decreased in US Latino and US African American populations and in most Latin American countries and Puerto Rico (AAPC range, −2.8% to −0.4% per year in women; −3.8% to −1.0% per year in men). In contrast, premature mortality rates among US white populations plateaued (AAPC, 0.02% per year [95% CI, −0.3% to 0.2%] in women; −0.2% per year [95% CI, −0.4% to 0.0%] in men) and increased in Nicaraguan men (AAPC, 0.6% per year [95% CI, 0.2% to 1.0%]) (Figure 2 and eTable 2 in the Supplement).
These decreases were observed across all age groups (20-34, 35-49, and 50-64 years) among US Latino populations and most other populations examined in the study. We observed increases in mortality rates among US white individuals aged 20 to 34 years after 2010, and rates for the 35 to 49 years and 50 to 64 years age groups plateaued. There were also increases among men in the 50 to 64 years age group in Belize and Panama (eFigure 1 in the Supplement).
The populations with the lowest premature mortality rates in 2015 (eg, US Latino populations and Peru) generally had lower premature mortality rates from all major causes, particularly heart disease (21 deaths per 100 000 population) and cancer (50 deaths per 100 000 population) (eFigure 2 in the Supplement). Substantial decreases in heart disease and cancer mortality contributed to decreasing mortality rates in US Latino and African American populations and in most Latin American countries. Although there was a marked reduction in cancer deaths among US white individuals (absolute change in rates, −19 deaths per 100 000 population in women; −20 deaths per 100 000 population in men), these were negated by increasing deaths due to external causes (absolute change in rates, 19 deaths per 100 000 population in women; 32 deaths per 100 000 population in men), along with smaller increases in deaths attributed to digestive diseases (absolute change in rates, 4 deaths per 100 000 population in women; 4 deaths per 100 000 population in men), respiratory diseases (absolute change in rates, 2 deaths per 100 000 population in women; 1 death per 100 000 population in men), and endocrine disorders (absolute change in rates, 1 death per 100 000 population in women; 4 deaths per 100 000 population in men) (eFigure 3 in the Supplement). In all 16 populations combined, deaths attributed to external causes accounted for 11% of premature deaths among women and 22% of premature deaths among men in 2015. Although there were major decreases in deaths from external causes in several Latin American populations (eg, Belizean and Colombian men), there were also increases in Uruguayan men (absolute change in rates, 15 deaths per 100 000 population) (eFigure 3 in the Supplement).
The increase in premature mortality among Nicaraguan men (AAPC, 0.6% per year) was associated with increases for several major causes, including digestive diseases, endocrine disorders, cancer, and respiratory diseases. Among Nicaraguan women, improvement in deaths due to heart disease and external causes were negated by increasing deaths due to endocrine disorders, respiratory diseases, and digestive diseases. We also observed increases in deaths due to cancer among Brazilian women (AAPC, 0.3% per year), deaths due to endocrine disorders among Mexican men (AAPC, 0.8% per year), and deaths due to respiratory diseases among Argentinean women (AAPC, 1.2% per year) (eFigure 2 and eTable 2 in the Supplement).
All-cause premature mortality of US Latino populations was lower for both sexes compared with Mexican, Puerto Rican, and Cuban populations during the study period. Across the 6 major causes of death, US Latino populations had lower mortality rates, except for external causes, where US Latina women had higher rates (ASMR for 2015, 20 deaths per 100 000 population) than the other 3 populations (Mexican population, 18 deaths per 100 000 population; Cuban population, 15 deaths per 100 000 population; and Puerto Rican population, 11 deaths per 100 000 population).
In this systematic evaluation of premature mortality in 16 populations in the United States and Latin America with high-quality mortality data, we found that recent premature mortality rates were lower in US Latino populations and in several Latin American countries than in US white populations, suggesting that there may be a broader Latin American paradox. Premature mortality rates also decreased in most of the populations studied from 2001 to 2015, in contrast to the trend among US white populations.
The healthy immigrant effect has long been proposed to partly explain the better health outcomes among the US Latino population than the US white population, because healthier Latino populations might be more likely to immigrate to the United States.13 A recent meta-analysis of international studies confirmed this pattern of lower death rates among migrant populations compared with their counterparts who reside in their country of origin.14 Our findings of lower mortality rates among US Latino populations than most of the other Latin American populations studied, especially compared with Mexican, Puerto Rican, and Cuban populations, provide some support for the healthy immigrant effect. However, the proportion of foreign-born US Latino individuals has been decreasing since 2000, and by 2015, two-thirds of US Latino individuals were born in the United States.15 The more general finding that premature death rates were lower in several Latin American countries than among US white individuals raises the question of whether there is a broader Latin American paradox.
The decrease in premature mortality rates in most of the Latin American countries we studied is consistent with reports16 of increasing life expectancy in these countries over the same period. The low premature mortality rates in Mexico were also previously reported by Norheim et al,5 who showed that Mexico had the seventh lowest premature mortality rates of the 25 most populous countries worldwide in 2010. Our findings for the reduction of heart disease–associated mortality were comparable with those of recent studies examining premature mortality associated with cardiovascular disease in Latin America,17 and this trend is projected to continue in the region.18 Although ischemic heart disease and stroke are the leading causes of premature deaths worldwide,10 Latin America has largely diverged from this trend, where cancer was the leading cause of premature deaths among women and external causes (eg, violence and accidents) were the leading cause of premature death among men.
In addition to the long-proposed healthy immigrant effect, our findings of lower premature mortality rates among US Latino populations may also reflect cultural, lifestyle, or undefined biological differences among Latino populations in general. In the United States, factors including lower smoking rates among Latina women and social and behavioral factors (eg, cohesive social networks) are likely associated with better health outcomes among Latino populations2,19 and clearly warrant further investigation. Since 1990, especially during 2000 to 2010, health levels in some Latin American countries improved faster than income levels, in contrast to countries with high development indices, such as the United States.20 The lower mortality rates in several Latin American countries are likely associated with modest economic development, notably improvement in cardiovascular health, lower rates of cancer associated with reductions in tobacco use, and better treatment and management of chronic diseases after the control of infectious diseases improved.16,21 However, these improvements are not uniform across each country, and health problems associated with emerging challenges, such as high obesity rates, increased rates of diabetes, and increasing alcohol consumption and related liver disease, are worrisome.22-24 Obesity rates have been increasing in Latin America.25 For example, in Mexican men we observed an increase in mortality associated with endocrine disorders, in contrast to the overall decrease in mortality in the country. Obesity prevalence among US Latino populations is also increasing and is higher than that of US white populations.26 Without effective public health strategies, high obesity prevalence may result in a reversal of mortality trends in Latin America during the coming decades.
The primary strength of this study is the use of a consistent and established coding of mortality data from a single source and restriction of the analysis to regions with high-quality mortality data across the study period. This enables direct comparisons in trends and absolute rates across countries and groups. Although this approach minimizes artifactual differences between countries, it cannot completely eliminate the potential differences associated with variations in coding practice and data quality over time across different countries.
This study also has several limitations. First, to provide a comprehensive comparison of the premature mortality rates among US populations with those among groups in Latin America, we aimed to include as many Latin American countries as possible. Therefore, we used a broad category of ICD-10 codes R00 to R99 as an indicator for ill-defined cause of death for data-quality assessment. This may result in underestimated mortality rates in those countries. However, it is unlikely to explain the observed mortality advantage of several countries compared with US white populations. In addition, the data quality has been improving,27 and current data still indicate large mortality differences.
Because of the required completeness of mortality data, there are gaps in the coverage of Latin American countries in our study. Importantly, most of the excluded countries have lower life expectancies than the included populations; therefore, we cannot generalize our findings to the whole region (eTable 3 in the Supplement). Nevertheless, the included countries account for more than 80% of the population of Latin America. We were not able to assess variation within Latin America by racial/ethnic groups. A previous study28 of indigenous and tribal peoples’ health found poorer health outcomes for indigenous populations in several Latin American countries compared with nonindigenous populations. In addition, we do not have data on socioeconomic status in each of the countries or groups studied and, therefore, were not able to provide socioeconomic status–adjusted rates. Previous US data suggested that there is an association between mortality and socioeconomic status by race/ethnicity,29 although comparisons of US Latino populations with Latin American populations are limited. This could indicate an area of future research.
In this study, we examined rates of premature mortality among individuals aged 20 to 64 years. Differences between country rankings in the current analysis and country rankings by life expectancy13 are likely explained by higher childhood mortality rates in some Latin American countries, including Peru, Panama, and Ecuador.
Among the 6 major causes of death, cancer accounted for most premature deaths among Latin American women. Cervical cancer is still the leading cause of cancer-related mortality among women in Latin America, despite decreases in recent years.30 As an infection-related cancer, cervical cancer is highly preventable. By 2014, the human papillomavirus vaccine had been offered through public immunization programs in 18 countries in Central and South America31; however, it will take several decades for vaccination to be associated with cancer death rates, and effective screening (eg, Papanicolaou test) and treatment are still urgently needed. We also observed an increase in cancer-associated mortality among Brazilian women. An analysis32 of recent Brazilian data found increasing mortality associated with breast cancer, particularly among women younger than 50 years. High mortality rates associated with external causes among Latin American men are another major preventable cause of premature death.33
Despite important decreases in premature mortality, our analysis highlights the high premature mortality rates among US African American populations compared with many Latin American populations, especially deaths due to heart disease and cancer. Similarly, a recent study29 analyzing US death certificate data reported increases in drug poisoning mortality among African American populations. Continued efforts are needed to address the disparities. Another important contrast is the plateau in premature mortality rates in US white populations and the high rates of deaths from external causes, driven by drug overdoses.34
In 2015, several Latino populations had lower premature mortality rates than the US white population, and nearly all the Latin American populations we studied had more rapid decreases in mortality rates during 2001 to 2015. Much of the progress has been driven by decreases in heart disease–associated mortality. Deaths due to cancer in women and due to external causes in men are major causes of premature death in Latin America. Concerning increases in the rates of premature mortality among certain populations in the Americas suggest a need to continue efforts against the major causes of death as well as to address emerging causes, such as obesity, with aggressive efforts. Our data also provide a template for potential future studies of migrant populations within well-characterized Latino populations in the United States, such as the Hispanic Community Health Study/Study of Latinos.35
Accepted for Publication: December 14, 2019.
Published: February 12, 2020. doi:10.1001/jamanetworkopen.2019.21085
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Chen Y et al. JAMA Network Open.
Corresponding Author: Yingxi Chen, MD, PhD, National Institutes of Health, National Cancer Institute, 9609 Medical Center Dr, Room 6E-328, Rockville, MD 20850 (firstname.lastname@example.org).
Author Contributions: Dr Chen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Chen, Freedman, Shiels, Perez-Stable, Berrington de González.
Acquisition, analysis, or interpretation of data: Chen, Freedman, Rodriquez, Shiels, Napoles, Withrow, Spillane, Sigel, Berrington de González.
Drafting of the manuscript: Chen, Napoles, Berrington de González.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Chen, Shiels, Spillane, Berrington de González.
Administrative, technical, or material support: Freedman, Rodriquez.
Conflict of Interest Disclosures: Dr Spillane reported receiving grants from Health Research Board Ireland during the conduct of the study. No other disclosures were reported.
Funding/Support: This work was supported by the Intramural Research Programs of the National Cancer Institute, National Heart, Lung, and Blood Institute, and National Institute on Minority Health and Health Disparities within the National Institutes of Health.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Create a personal account or sign in to: