Comparison of Health Outcomes Among High- and Low-Income Adults Aged 55 to 64 Years in the US vs England | Geriatrics | JAMA Internal Medicine | JAMA Network
[Skip to Navigation]
Figure 1.  Adjusted Risk Ratios (ARRs) for US-England Health Differences at Ages 55 to 64 Years for 2008-2016
Adjusted Risk Ratios (ARRs) for US-England Health Differences at Ages 55 to 64 Years for 2008-2016

Adjustment factors included survey year, age, sex, race, foreign-born status (outside the US for HRS; outside the UK for ELSA), household size, marital status, educational level in International Standard Classification of Education categories, and household income in decile. Estimates were calculated holding all adjustment factors at their grand mean values (ie, constant across countries). ADL indicates activities of daily living; CES-D8, 8-item Center for Epidemiologic Studies Depression Scale; CRP, C-reactive protein; HbA1c, hemoglobin A1c; and IADL, instrumental ADL.

Figure 2.  Adjusted Prevalence of Self-assessed Health Outcomes at Ages 55 to 64 Years for 2008-2016 by Country-Specific Income Decile
Adjusted Prevalence of Self-assessed Health Outcomes at Ages 55 to 64 Years for 2008-2016 by Country-Specific Income Decile

Adjustment factors include age, year, sex, foreign-born status (outside the US for HRS; outside the UK for ELSA), race, household size, and marital status. Estimates were calculated holding all adjustment factors at their grand mean values (ie, constant across countries and income deciles). Error bars show 95% CIs. ADL indicates activities of daily living; ELSA, English Longitudinal Study of Ageing; HRS, Health and Retirement Study; and IADL, instrumental ADL.

Figure 3.  Adjusted Prevalence of Measured Health Outcomes at Ages 55 to 64 Years for 2008-2016 by Country-Specific Income Decile
Adjusted Prevalence of Measured Health Outcomes at Ages 55 to 64 Years for 2008-2016 by Country-Specific Income Decile

Adjustment factors included age, year, sex, foreign-born status (outside the US for HRS; outside the UK for ELSA), race, household size, and marital status. Estimates were calculated holding all adjustment factors at their grand mean values (ie, constant across countries and income deciles). Error bars show 95% CIs. CRP indicates C-reactive protein; ELSA, English Longitudinal Study of Ageing; HbA1c, hemoglobin A1c; and HRS, Health and Retirement Study. SI conversion factors: To convert CRP to milligrams per liter, multiply by 10; HbA1c to proportion of total Hb, multiply by 0.01.

Figure 4.  Adjusted Prevalence of Self-reported Physician-Diagnosed Health Conditions at Ages 55 to 64 Years for 2008-2016 by Country-Specific Income Decile
Adjusted Prevalence of Self-reported Physician-Diagnosed Health Conditions at Ages 55 to 64 Years for 2008-2016 by Country-Specific Income Decile

Adjustment factors included age, year, sex, foreign-born status (outside the US for HRS; outside the UK for ELSA), race, household size, and marital status. Estimates were calculated holding all adjustment factors at their grand mean values (ie, constant across countries and income deciles). Error bars show 95% CIs. ELSA indicates English Longitudinal Study of Ageing; HRS, Health and Retirement Study.

Table.  Health Difference Between the Bottom 20% and Top 20% of Income Distribution in the US and England at Ages 55 to 64 Years for 2008-2016a
Health Difference Between the Bottom 20% and Top 20% of Income Distribution in the US and England at Ages 55 to 64 Years for 2008-2016a
1.
Woolf  SH, Aron  LUS.  Health in International Perspective: Shorter Lives, Poorer Health. National Academies Press; 2013.
2.
Preston  S, Ho  J.  International Differences in Mortality at Older Ages: Dimensions and Sources. National Research Council; 2010.
3.
Cohen  B, Preston  SH, Crimmins  EM.  Explaining Divergent Levels of Longevity in High-Income Countries. National Academies Press; 2011.
4.
Case  A, Deaton  A.  Rising morbidity and mortality in midlife among white non-Hispanic Americans in the 21st century.   Proc Natl Acad Sci U S A. 2015;112(49):15078-15083. doi:10.1073/pnas.1518393112 PubMedGoogle ScholarCrossref
5.
Bound  J, Geronimus  AT, Rodriguez  JM, Waidmann  TA.  Measuring recent apparent declines in longevity: the role of increasing educational attainment.   Health Aff (Millwood). 2015;34(12):2167-2173. doi:10.1377/hlthaff.2015.0481 PubMedGoogle ScholarCrossref
6.
Martinson  ML, Teitler  JO, Reichman  NE.  Health across the life span in the United States and England.   Am J Epidemiol. 2011;173(8):858-865. doi:10.1093/aje/kwq325 PubMedGoogle ScholarCrossref
7.
Banks  J, Marmot  M, Oldfield  Z, Smith  JP.  Disease and disadvantage in the United States and in England.   JAMA. 2006;295(17):2037-2045. doi:10.1001/jama.295.17.2037 PubMedGoogle ScholarCrossref
8.
Nolte  E, McKee  CM.  In amenable mortality—deaths avoidable through health care—progress in the US lags that of three European countries.  [published correction appears in Health Aff (Millwood). 2012 Oct;31(10):2356].  Health Aff (Millwood). 2012;31(9):2114-2122. doi:10.1377/hlthaff.2011.0851 PubMedGoogle ScholarCrossref
9.
Avendano  M, Kawachi  I.  The search for explanations of the American health disadvantage relative to the English.   Am J Epidemiol. 2011;173(8):866-869. doi:10.1093/aje/kwq484 PubMedGoogle ScholarCrossref
10.
van Hedel  K, Avendano  M, Berkman  LF,  et al.  The contribution of national disparities to international differences in mortality between the United States and 7 European countries.   Am J Public Health. 2015;105(4):e112-e119. doi:10.2105/AJPH.2014.302344 PubMedGoogle ScholarCrossref
11.
Banks  J, Muriel  A, Smith  JP.  Disease prevalence, disease incidence, and mortality in the United States and in England.   Demography. 2010;47(1)(suppl):S211-S231. doi:10.1353/dem.2010.0008 PubMedGoogle ScholarCrossref
12.
Kinge  JM, Modalsli  JH, Øverland  S,  et al.  Association of household income with life expectancy and cause-specific mortality in Norway, 2005-2015.   JAMA. 2019;321(19):1916-1925. doi:10.1001/jama.2019.4329 PubMedGoogle ScholarCrossref
13.
Jivraj  S.  Are self-reported health inequalities widening by income? an analysis of British pseudo birth cohorts born, 1920-1970.   J Epidemiol Community Health. 2020;74(3):255-259. doi:10.1136/jech-2019-213186 PubMedGoogle ScholarCrossref
14.
Zaninotto  P, Batty  GD, Stenholm  S,  et al.  Socioeconomic inequalities in disability-free life expectancy in older people from England and the United States: a cross-national population-based study.   J Gerontol A Biol Sci Med Sci. 2020;75(5):906-913. doi:10.1093/gerona/glz266PubMedGoogle ScholarCrossref
15.
Chetty  R, Stepner  M, Abraham  S,  et al.  The association between income and life expectancy in the United States, 2001–2014.   JAMA. 2016;315(16):1750-1766. doi:10.1001/jama.2016.4226 PubMedGoogle ScholarCrossref
16.
OECD Data. Inequality. Poverty rate. Accessed January 16, 2020. https://data.oecd.org/inequality/poverty-rate.htm
17.
World Inequality Database. Accessed February 15, 2020. https://wid.world
18.
Case  A, Deaton  A.  Deaths of Despair and the Future of Capitalism. Princeton University Press; 2020.
19.
National Center for Health Statistics. Health, United States, 2014: with special feature on adults aged 55–64. National Center for Health Statistics; 2015.
20.
Lee  J, Smith  JP.  Health, Economic Status, and Aging in High-Income Countries. National Academies Press; 2018.
21.
Makaroun  LK, Brown  RT, Diaz-Ramirez  LG,  et al.  Wealth-associated disparities in death and disability in the United States and England.   JAMA Intern Med. 2017;177(12):1745-1753. doi:10.1001/jamainternmed.2017.3903 PubMedGoogle ScholarCrossref
22.
Choi  H, Schoeni  R. Health of Americans who must work longer to reach Social Security retirement age.  Health Aff (Millwood). 2017;36(10):1815-1819. doi:10.1377/hlthaff.2017.0217 PubMedCrossref
23.
Evans  W, Wolfe  B, Adler  N. The income-health gradient. Accessed October 27, 2019. https://www.irp.wisc.edu/publications/focus/pdfs/foc301b.pdf
24.
Sonnega  A, Faul  JD, Ofstedal  MB, Langa  KM, Phillips  JW, Weir  DR.  Cohort profile: the Health and Retirement Study (HRS).   Int J Epidemiol. 2014;43(2):576-585. doi:10.1093/ije/dyu067 PubMedGoogle ScholarCrossref
25.
NetCen Social Research. ELSA: user guide to the main interview datasets: waves 1 to 8. Accessed August 16, 2018. http://doc.ukdataservice.ac.uk/doc/5050/mrdoc/pdf/5050_waves_1-8_interviewer_data_user_guide_v01.pdf
26.
Phillips  D, Lin  Y, Wight  J, Chien  S, Lee  J. Harmonized ELSA documentation; version E, April 2017. Accessed October 16, 2018. http://doc.ukdataservice.ac.uk/doc/5050/mrdoc/pdf/5050_harmonized_elsa_e.pdf
27.
Schane  RE, Walter  LC, Dinno  A, Covinsky  KE, Woodruff  PG.  Prevalence and risk factors for depressive symptoms in persons with chronic obstructive pulmonary disease.   J Gen Intern Med. 2008;23(11):1757-1762. doi:10.1007/s11606-008-0749-z PubMedGoogle ScholarCrossref
28.
Whelton  PK, Carey  RM, Aronow  WS,  et al.  2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults: executive summary—a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines.   Circulation. 2018;138(17):e426-e483.PubMedGoogle Scholar
29.
Organisation for Economic Co-operation and Development.  Classifying Educational Programmes: Manual for ISCED-97 Implementation in OECD Countries. 1999 Edition. OECD Publications; 1999.
30.
Norton  EC, Miller  MM, Kleinman  LC.  Computing adjusted risk ratios and risk differences in Stata.   Stata J. 2013;13(3):492-509. doi:10.1177/1536867X1301300304 Google ScholarCrossref
31.
Cox  NJ.  Speaking Stata: smoothing in various directions.   Stata J. 2005;5(4):574-593. doi:10.1177/1536867X0500500408 Google ScholarCrossref
32.
Molina  T.  Adjusting for heterogeneous response thresholds in cross-country comparisons of self-reported health.   J Econ Ageing. 2017;10(C):1-20. doi:10.1016/j.jeoa.2017.08.002 Google ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

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.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    July 13, 2020

    Comparison of Health Outcomes Among High- and Low-Income Adults Aged 55 to 64 Years in the US vs England

    Author Affiliations
    • 1Department of Internal Medicine, School of Medicine, University of Michigan, Ann Arbor
    • 2Institute for Healthcare Policy and Innovation, University of Michigan, Ann Arbor
    • 3Department of Behavioural Science and Health, University College London, London, United Kingdom
    • 4Veterans Affairs Ann Arbor Center for Clinical Management Research, Ann Arbor
    • 5Department of Epidemiology, School of Public Health, University of Michigan, Ann Arbor
    • 6Survey Research Center, Institute for Social Research, University of Michigan, Ann Arbor
    • 7Gerald R. Ford School of Public Policy, University of Michigan, Ann Arbor
    • 8Institute of Epidemiology and Health Care, University College London, London, United Kingdom
    JAMA Intern Med. 2020;180(9):1185-1193. doi:10.1001/jamainternmed.2020.2802
    Key Points

    Question  What is the difference in health status between high- and low-income individuals in the US vs England?

    Findings  In this cross-sectional study including 18 572 persons (46 887 person-years of observations), the health gap between the bottom 20% and top 20% of income distribution was significantly greater for US adults aged 55 to 64 years than their English peers on 13 of 16 health measures. In addition, for most measures, the health of US adults appeared to be poorer than that of their peers in England, especially those from the lower end of the income distribution.

    Meaning  The findings of this study suggest that the income-health gradient is greater among US middle-aged adults than among their peers in England, with poor health especially pronounced among those with lower income in the US.

    Abstract

    Importance  Socioeconomic differences in life expectancy, health, and disability have been found in European countries as well as in the US. Identifying the extent and pattern of health disparities, both within and across the US and England, may be important for informing public health and public policy aimed at reducing these disparities.

    Objective  To compare the health of US adults aged 55 to 64 years with the health of their peers in England across the high and low ranges of income in each country.

    Design, Setting, and Participants  Using data from the Health and Retirement Study (HRS) and the English Longitudinal Study of Ageing (ELSA) for 2008-2016, a pooled cross-sectional analysis of comparably measured health outcomes, with adjustment for demographic characteristics and socioeconomic status, was conducted. The analysis sample included community-dwelling adults aged 55 to 64 years from the HRS and ELSA, resulting in 46 887 person-years of observations. Data analysis was conducted from September 17, 2019, to May 12, 2020.

    Exposures  Residence in the US or England and yearly income.

    Main Outcomes and Measures  Sixteen health outcomes were compared, including 5 self-assessed outcomes, 3 directly measured outcomes, and 8 self-reported physician-diagnosed health conditions.

    Results  This cross-sectional study included 12 879 individuals and 31 928 person-years from HRS (mean [SD] age, 59.2 [2.8] years; 51.9% women) and 5693 individuals and 14 959 person-years from ELSA (mean [SD] age, 59.3 [2.9] years; 51.0% women). After adjusting for individual-level demographic characteristics and socioeconomic status, a substantial health gap between lower-income and higher-income adults was found in both countries, but the health gap between the bottom 20% and the top 20% of the income distribution was significantly greater in the US than England on 13 of 16 measures. The adjusted US-England difference in the prevalence gap between the bottom 20% and the top 20% ranged from 3.6 percentage points (95% CI, 2.0-5.2 percentage points) in stroke to 9.7 percentage points (95% CI, 5.4-13.9 percentage points) for functional limitation. Among individuals in the lowest income group in each country, those in the US group vs the England group had significantly worse outcomes on many health measures (10 of 16 outcomes in the bottom income decile); the significant differences in adjusted prevalence of health problems in the US vs England for the bottom income decile ranged from 7.6% (95% CI, 6.0%-9.3%) vs 3.8% (95% CI, 2.6%-4.9%) for stroke to 75.7% (95% CI, 72.7%-78.8%) vs 59.5% (95% CI, 56.3%-62.7%) for functional limitation. Among individuals in the highest income group, those in the US group vs England group had worse outcomes on fewer health measures (4 of 16 outcomes in the top income decile); the significant differences in adjusted prevalence of health problems in the US vs England for the top income decile ranged from 36.9% (95% CI, 33.4%-40.4%) vs 30.0% (95% CI, 27.2%-32.7%) for hypertension to 35.4% (95% CI, 32.0%-38.7%) vs 22.5% (95% CI, 19.9%-25.1%) for arthritis.

    Conclusions and Relevance  For most health outcomes examined in this cross-sectional study, the health gap between adults with low vs high income appeared to be larger in the US than in England, and the health disadvantages in the US compared with England are apparently more pronounced among individuals with low income. Public policy and public health interventions aimed at improving the health of adults with lower income should be a priority in the US.

    Introduction

    The US spends more on health care than any other country, but health and life expectancy of US adults have been decreasing compared with many other high-income countries.1 Life expectancy in the US has remained relatively stable in recent years,1-3 although some evidence that mortality rates among middle-aged US adults have increased has been reported.4,5 The US health disadvantage is found not only in life expectancy but also in self-reported health,6 biomarkers for disease,7 and many specific causes of death,6,8-10 which is persistent and substantial, even after controlling for individual risk factors and health care system factors.1,7,11

    Socioeconomic differences in life expectancy,12 health,13 and disability14 have been found in European countries as well as in the US. However, compared with many European countries, the US has higher income inequality15-17 and has a relatively limited government safety net including public health care coverage.7,18 We believe that the relative health disadvantage of the US compared with many European countries would be especially greater among people with low income than people with higher income.

    This study compared the health of US adults aged 55 to 64 years with the health of their peers in England. The US and England have a relatively similar culture, language, and economic system, yet have substantially different health care and social welfare systems.7,18 The US provides limited public health insurance for those younger than 65 years, but England provides publicly funded health care free at the time of service for all individuals under the National Health Service. About 17% of US adults aged 55 to 64 years had only public health insurance coverage, and 13% had no health insurance coverage in 2012-2013.19 The greater level of income inequality16 with limited public safety net in the US compared with England raises the question of whether those with low income in the US are at greater risk for poor health outcomes than low-income adults in England.

    To our knowledge, few studies have examined health differences between the US and England by socioeconomic groups. Zaninotto et al14assessed socioeconomic differences in disability-free life expectancy in England and the US, examining it by occupation, wealth, and educational level. Lee and Smith20 examined the social gradient in health across high-income countries, including the US and England, focusing on educational level and wealth. Makaroun et al21 compared mortality and disability among adults in different wealth quintiles in the US and England. However, these studies did not examine US-England health differences by income group. While educational level, occupation, and wealth are important, income may have an independent association with health and health care access, especially for people with limited public health care coverage. Banks et al7 examined US-England health differences among adults aged 55 to 64 years by income tertile in 2002 and found worse health in the US regardless of income group. However, to our knowledge, little evidence exists on whether these differences by income between the US and England have persisted in recent decades. Since 2002, the baby boom generation (born between 1946-1964) in the US has entered the age 55- to 64-year cohort with worse health than earlier generations,22 which has important implications for the health and health care needs of US adults. In addition, using more granular income groupings than tertiles may be important because income gradients in health are typically steeper at lower levels of income.23

    To address this gap in evidence, we examined differences between the US and England in health among adults aged 55 to 64 years by using more recent years of data and more fine-grained income groups than in prior studies. We examined 16 health outcomes measured in comparable nationally representative data from the US and England from 2008 to 2016.

    Methods
    Data Sources and Sample

    The Health and Retirement Study (HRS) is a longitudinal biennial survey of approximately 20 000 Americans older than 50 years that started in 1992. The HRS collects extensive health information as well as sociodemographic characteristics.24 The HRS was approved by the University of Michigan Institutional Review Board. The English Longitudinal Study of Ageing (ELSA) is a biennial longitudinal household survey of English adults aged 50 years or older that began data collection in 2002 including detailed health and sociodemographic measures. The ELSA was developed with a goal of using methods and survey questions that are comparable to the HRS to facilitate cross-national comparisons.25 Ethical approval for all the ELSA waves was granted by the National Health Service Research Ethics Committees under the National Research and Ethics Service. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies.

    Our study sample included community-dwelling adults aged 55 to 64 years from both the HRS and ELSA over the period 2008-2016. This age group was chosen because 55 years is the youngest age to be nationally representative for all years and 64 years is the oldest age before eligibility for the Medicare program in the US. There were 12 879 persons and 31 928 person-year observations from the HRS and 5693 persons and 14 959 person-year observations from the ELSA. The sample size and sample period varied depending on the health outcomes examined because some health measures were not available for the entire study period.

    Measures

    We examined 16 harmonized measures for the HRS and ELSA.26 Five measures of self-assessed health outcomes were any functional limitation (climbing 1 flight of stairs; climbing several flights of stairs; getting up from a chair; picking up a dime; reaching or extending arms upward; pushing or pulling large objects; sitting for 2 hours; lifting or carrying 4.5 kg; and stooping, kneeling, or crouching), difficulty with any instrumental activity of daily living (IADL) (using the telephone, managing money, taking medication, shopping, and preparing meals), difficulty with any activity of daily living (ADL) (walking across the room, dressing, bathing, eating, getting in or out of bed, and using the toilet), fair or poor self-rated general health, and depressive symptoms according to the 8-item Center for Epidemiologic Studies Depression Scale (CES-D8).27 Three directly measured health outcomes were blood pressure (elevated level defined as ≥140 mm Hg systolic or ≥90 mm Hg diastolic),28 hemoglobin A1c (HbA1c) level, (elevated level defined as >6.5% [to convert to proportion of total hemoglobin, multiply by 0.01]), and C-reactive protein (CRP) level (elevated level defined as >0.3 mg/dL [to convert to milligrams per liter, multiply by 10]). Eight measures of self-reported physician-diagnosed health conditions were hypertension, arthritis, diabetes, heart problems, cancer, stroke, chronic lung disease, and emotional or psychiatric problem. There were fewer than 4.8% missing values in health outcomes in the HRS and fewer than 6.3% in the ELSA.

    Demographic measures included age, sex, foreign-born status (outside the US for HRS; outside the UK for ELSA), race (white vs other), household size (number of people living in the household), and marital status (partnered, separated/divorced, widowed, and never married). Educational level groups included a 3-tiered harmonized scale.29 To adjust for rank in the income distribution, we used income decile within each country, based on household income. For analyses by income groups, we used the income quintiles, deciles, and percentiles, defined within each country, survey year, and age, with the adjustment for household size. There were fewer than 0.5% missing values for all covariates except educational level (6.4%) and income (2.4%) in the ELSA.

    Statistical Analysis

    Data analysis was performed from September 17, 2019, to May 12, 2020. We conducted pooled cross-sectional analyses to compare health differences and their variation by income group (quintile, decile, and percentile). We used multivariable logistic regressions to adjust for demographic and socioeconomic characteristics and estimated the adjusted risk ratios (ARRs) of health problems for the HRS sample vs ELSA sample by evaluating estimated risks at the grand mean values of covariates.30

    To estimate the adjusted prevalence of health problems by income in each country, we included as explanatory variables the indicator variables for each country-specific income decile, the indicator for the HRS sample, an interaction between the indicator variables for each income decile and an indicator for the HRS sample, and demographic covariates. The adjusted prevalence within each income decile for each survey (HRS and ELSA) was calculated. We also calculated the health difference between the bottom 20% and the top 20% in each country as well as between the countries to assess whether the health gap by income varied between the US and England.

    Auxiliary and sensitivity analyses were conducted. First, we summarized sample characteristics for the HRS and ELSA. Second, we estimated the unadjusted prevalence of health problems in each country and unadjusted risk ratios of health problems for the US vs England. Third, to examine the income distribution, we collected data on median income by income decile for HRS and ELSA. Fourth, to examine more detailed income patterns, we calculated health outcomes for each income percentile (1 for the lowest and 100 for the highest income level) defined within each country by using a kernel-weighted local polynomial regression.31 Fifth, we reassessed health differences after restricting the sample to non-Hispanic white individuals from the HRS sample. Sixth, we adjusted for smoking and body mass index to examine whether these variables were associated with the results. Seventh, we used a multiple imputation technique to impute missing values in covariates using chained equations with 10 replications. Two-tailed unpaired P values were determined from a t test of whether the estimate of difference in the adjusted prevalence was significantly different from 0. Findings were considered significant if 95% CIs were not overlapped and the P value was less than .05.

    All analyses were conducted using Stata, version 15 (StataCorp). Standard errors and population weight were adjusted using the complex survey design from the HRS and ELSA.

    Results

    This cross-sectional study included 12 879 individuals and 31 928 person-years from HRS (mean age, 59.2 years; 51.9% women) and 5693 individuals and 14 959 person-years from ELSA (mean age, 59.3 years; 51.0% women) (eTable 1 in the Supplement). The proportion of foreign-born adults (11.2% in HRS and 12.2% in ELSA) was comparable between the US and England, but the proportion of white individuals was lower in the HRS sample (79.9% vs 93.8%; P < .001). The HRS sample was composed of more single-person households (19.0% vs 15.6%; P < .001) and large households with 5 or more persons (6.2% vs 3.0%; P < .001), as well as those without a spouse (29.6% vs 23.7%; P < .001). The proportion of adults with a high level of education was greater in the US compared with England (31.9% vs 21.9%; P < .001), and US adults were more likely to have never smoked (45.1% vs 39.1%; P < .001) and to be obese (38.5% vs 34.4%; P < .001).

    As depicted in Figure 1, adults in the US compared with England were significantly more likely to report a functional limitation (ARR, 1.46; 95% CI, 1.40-1.52), IADL limitation (ARR, 1.46; 95% CI, 1.27-1.66), and fair or poor self-rated general health (ARR, 1.22; 95% CI, 1.13-1.32). The adjusted prevalence of an ADL limitation and depression was comparable between the US and England. US adults had a significantly greater adjusted prevalence of a high HbA1c level (ARR, 1.48; 95% CI, 1.23-1.79) and a high CRP level (ARR, 1.17; 95% CI, 1.04-1.30). The adjusted prevalence of directly measured high blood pressure was comparable, while self-reported physician-diagnosed hypertension was significantly higher in the US than England (ARR, 1.39; 95% CI, 1.32-1.47). All physician-diagnosed health conditions were more prevalent among US adults, with the ARR ranging from 1.25 (95% CI, 1.14-1.35) for heart problems to 2.17 (95% CI, 1.81-2.54) for chronic lung disease.

    The unadjusted prevalence and risk ratios of health problems in the US vs England are reported in eTable 2 in the Supplement. The risk ratios of 12 of 16 health problems were higher in the US than England ranging from 1.14 (95% CI, 1.04-1.24) for high inflammation to 2.22 (95% CI, 1.86-2.57) for stroke. The overall findings were robust to additional sample restrictions including only non-Hispanic white individuals, additional adjustments for smoking and body mass index, and use of imputed data (eFigure 1, eFigure 2, and eFigure 3 in the Supplement).

    Variation in Health by Income in the US and England

    The unadjusted prevalence of each of the 16 health outcomes by income percentile suggested a strong income-health gradient for most outcomes in each country (eFigure 4 in the Supplement). For example, the prevalence of functional limitation in the 10th percentile vs 90th percentile was 78.5% vs 39.6% for the US and 57.8% vs 30.9% for England. The prevalence of high inflammation in the 10th percentile vs 90th percentile was 42.6% vs 25.9% for the US and 33.6% vs 24.8% for England. The prevalence of diabetes in the 10th percentile vs 90th percentile was 29% vs 11.7% for the US and 14.5% vs 6.0% for England. Furthermore, these figures suggest that, among health outcomes that differ between countries, the gap is never larger and often smaller at the top than the bottom of the income distribution.

    Median household income of adults aged 55 to 64 years was greater in the US than England for all deciles except for the bottom decile (eTable 3 in the Supplement). The median after-tax household income (US dollars) was $3650 in the US vs $5286 in England at the bottom decile in each country during 2008 to 2014. In contrast, the median after-tax household income in the US was more than twice the value in England at the top decile in 2012 ($143 679 vs $70 712), implying higher income inequality in the US (ratio of median within the top vs bottom decile was 39 in the US and 13 in England).

    As presented in the Table, the gap between the bottom 20% and the top 20% on self-assessed health outcomes was substantial for both countries, ranging from a 22 percentage point difference (IADL limitation) to a 38 percentage point difference (fair or poor health) in the US and from a 16 percentage point difference (IADL limitation) to a 33 percentage point difference (fair or poor health) in England. This health gap by income (bottom 20% minus top 20%) was significantly greater in the US than England on functional limitation (US-England difference in the gap was 9.7 percentage points; 95% CI, 5.4-13.9 percentage points), IADL limitation (US-England difference in the gap was 5.3 percentage points; 95% CI, 2.2-8.4 percentage points), ADL limitation (US-England difference in the gap was 4.8 percentage points; 95% CI, 1.3-8.3 percentage points), and fair or poor health (US-England difference in the gap was 4.9 percentage points; 95% CI, 0.7-9.1 percentage points) (Table).

    Figure 2 presents the adjusted prevalence of poor health based on the self-assessed health outcomes for each income decile in each country. Adults in the US were more likely to report a functional limitation compared with those in England in all income groups, but the gap was greater in low-income groups (bottom decile: 75.7%; 95% CI, 72.7%-78.8% in HRS vs 59.5%; 95% CI, 56.3%-62.7% in ELSA; top decile: 35.1%; 95% CI, 31.9%-38.4% in HRS vs 29.1%; 95% CI, 26.2%-32.0% in ELSA). The adjusted prevalence of an IADL limitation was significantly greater for US adults in the lowest 2 income deciles (bottom decile: 25.9%; 95% CI, 23.3%-28.5% in HRS vs 18.9%; 95% CI, 16.2%-21.6% in ELSA; second decile: 24.7%; 95% CI, 21.2%-28.1% in HRS vs 17.5%; 95% CI, 15.0%-19.9% in ELSA), but was comparable for all other higher income deciles.

    The adjusted prevalence of the bottom 20% was significantly greater than that of the top 20% on all 3 directly measured health outcomes in the US (gap was 6.6 percentage points; 95% CI, 0.3-13.0 percentage points for high blood pressure; 9.8 percentage points; 95% CI, 5.2-14.5 percentage points for HbA1c level, and 15.3 percentage points; 95% CI, 8.3-22.3 percentage points for CRP level) and on 2 outcomes in England (the gap was 7.0 percentage points; 95% CI, 3.2-10.8 percentage points for HbA1c and 6.8 percentage points; 95% CI, 1.6-12.0 percentage points for CRP levels) (Table). This health gap by income (bottom 20% minus top 20%) was significantly greater in the US than England on high blood pressure (US-England difference in the gap was 8.2 percentage points; 95% CI, 1.1-15.3 percentage points) and elevated CRP level (US-England difference in the gap was 8.5 percentage points; 95% CI, 0.4-16.6 percentage points) (Table). The adjusted prevalence of poor health outcomes based on direct measures at each income decile is depicted in Figure 3.

    There were substantial differences between the bottom 20% and the top 20% for all physician-diagnosed outcomes except cancer (Table), ranging from a 6.4 percentage point difference (stroke: 95% CI, 5.0-7.8 percentage points ) to a 19.5 percentage point difference (arthritis: 95% CI, 16.4-22.6 percentage points ) in the US and from a 2.8 percentage point difference (stroke: 95% CI, 2.0-3.7 percentage points ) to a 14.5 percentage point difference (arthritis: 95% CI, 11.5-17.5 percentage points) in England. This health gap by income (bottom 20% minus top 20%) was significantly greater in the US than England on all physician-diagnosed conditions except cancer (Table). The US-England difference in the gap was 7.0 percentage points (95% CI, 2.2-11.9 percentage points) for hypertension, 5.0 percentage points (95% CI, 1.1-8.9 percentage points) for arthritis, 4.3 percentage points (95% CI, 1.2-7.4 percentage points) for diabetes, 3.7 percentage points (95% CI, 0.7-6.8 percentage points) for heart problems, 3.6 percentage points (95% CI, 2.0-5.2 percentage points) for stroke, 6.0 percentage points (95% CI, 3.7-8.4 percentage points) for chronic lung disease, and 4.6 percentage points (95% CI, 1.6-7.6 percentage points). Figure 4 shows that, at all income levels, adults in the US were more likely to report a physician diagnosis of hypertension, arthritis, diabetes, and emotional/psychiatric problem, with a greater gap among the lower-income groups compared with the higher-income groups. For example, the adjusted prevalence of hypertension in the US vs England was 59.6% (95% CI, 55.2%-64.1%) vs 43.9% (95% CI, 40.6%-47.0%) (P < .001) for the bottom decile but 36.9% (95% CI, 33.4%-40.4%) vs 30.0% (95% CI, 27.2%-32.7%) for the top decile. The prevalence of arthritis in the US vs England was 53.9% (95% CI, 50.6%-57.2%) vs 37.6% (95% CI, 34.3%-40.8%) for the bottom decile and 35.4% (95% CI, 32.0%-38.7%) vs 22.5% (95% CI, 19.9%-25.1%) for the top decile. The adjusted prevalence of a heart problem, stroke, and chronic lung disease were comparable between the US and England for the higher income deciles, but significantly greater in the US for the lower income deciles (bottom decile: 21.0%; 95% CI, 18.7%-23.3% vs 16.4%; 95% CI, 14.0%-18.7% for heart problem; 7.6%; 95% CI, 6.0%-9.3% vs 3.8%; 95% CI, 2.6%-4.9% for stroke; and 13.8%; 95% CI, 11.3%-16.3% vs 7.3%; 95% CI, 5.8%-8.8% for chronic lung disease). Results from sensitivity analyses were mostly consistent with those from the main analyses (eFigure 5, eFigure 6, and eFigure 7 in the Supplement).

    Discussion

    Using data on adults aged 55 to 64 years from comparable nationally representative surveys of the US and England, this study provides evidence on disparities in health across income groups in the US vs England. Analyses of 16 health outcomes found substantial income differences in each country and for all outcomes except cancer in both countries and measured high blood pressure in England. Moreover, the disparity in health between low-income and high-income adults was significantly greater in US adults for 6 of the 8 self-assessed or directly measured outcomes, and 7 of the 8 self-reported physician diagnosis outcomes.

    Cross-country differences in health were in favor of England for all health outcomes except ADL limitations, depression, and measured blood pressure, which were not significantly different between countries. The US and England disparities were pronounced among lower-income adults.

    These findings persisted after adjusting for demographic factors, educational level, smoking, and body mass index. With the current data we were not able to assess the role of other factors in explaining these findings. We believe, however, that one important factor is that financial insecurity among low-income adults may lead to vulnerabilities that are not as well addressed in the US compared with England. We found that both the relative income and the absolute level of income were lower in the lowest income group among US adults aged 55 to 64 years compared with their peers in England. Both across and within countries, the lowest-income adults in the US had the worst outcome for nearly all 16 measures. At the same time, none of the 16 health outcomes were better in the US than England among the top 2 deciles despite having at least 80% higher income. Health care access appears to be more limited in the US due to the more complicated and fragmented health care system, and many US adults are uninsured, which may result in substantial out-of-pocket costs, especially among low-income individuals in the US. Social welfare programs in the US, separate from health care, are also limited compared with the UK system. These factors may be associated with worse overall health for lower-income US adults compared with those in England.

    Limitations

    Our study has limitations. First, questions regarding self-assessed health measures may be interpreted differently between countries owing to different health expectations and reference groups.32 Second, there may be systematic differences in the diagnosis of disease, such as screening and public awareness of symptoms, and other unobserved factors that influenced reporting of the physician-diagnosed health conditions. Third, directly measured health outcomes might not reflect the underlying health conditions across the US and England because of treatment differences between the countries. For example, self-reported diagnosis of hypertension was significantly higher in the US, but the rate of high blood pressure based on direct measurement was similar between the countries. Among those who reported a hypertension diagnosis, the rate of directly measured high blood pressure was significantly lower for US adults, suggesting more aggressive blood pressure treatment and/or diagnosis of less-severe cases in the US. Fourth, comparing health in only 2 countries limits the ability to identify the potential causes for cross-country differences in health. While income inequality is lower in England than the US, it is higher than in many other European countries, and England also has a more limited social welfare system than some other European countries.

    Conclusions

    The findings of this study suggest that the gap in health between lower-income vs higher-income adults is substantial and significantly larger in the US than England for most health outcomes that we examined. In addition, lower-income adults in the US appear to have poorer health compared with lower-income adults in England. Identifying the extent and pattern of health disparities, both within and across the US and England, may be important for informing public health and public policy aimed at reducing these disparities. Future studies that better identify the factors, including individual, environmental, health system, and social welfare, that may be associated with the poor health of low-income US adults, would be valuable.

    Back to top
    Article Information

    Accepted for Publication: May 23, 2020.

    Corresponding Author: HwaJung Choi, PhD, Department of Internal Medicine, School of Medicine, University of Michigan, 2800 Plymouth Rd, NCRC Bldg 14, GR109, Ann Arbor, MI 48104 (hwajungc@med.umich.edu).

    Published Online: July 13, 2020. doi:10.1001/jamainternmed.2020.2802

    Author Contributions: Dr Choi and Ms Cho 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: Choi, Heisler, Clarke, Schoeni, Jivraj.

    Acquisition, analysis, or interpretation of data: Choi, Steptoe, Clarke, Schoeni, Jivraj, Cho, Langa.

    Drafting of the manuscript: Choi, Cho.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Choi, Cho.

    Obtained funding: Choi, Clarke, Jivraj, Langa.

    Administrative, technical, or material support: Steptoe, Clarke, Cho.

    Supervision: Choi, Steptoe, Heisler, Schoeni.

    Conflict of Interest Disclosures: Dr Clarke reported receiving grants from the National Institutes of Health (NIH) during the conduct of the study. Dr Jivraj reported receiving grants from the National Institute of Aging (NIA) during the conduct of the study. Dr Langa reported receiving grants from the NIH/NIA during the conduct of the study. No other disclosures were reported.

    Funding/Support: The Health and Retirement Study is funded by NIA grant U01 AG009740) and performed at the Institute for Social Research, University of Michigan. This research was supported by NIA grant R21 AG054818. Dr Choi was supported by NIA grant K01 AG057820. Dr Jivraj was supported by Leverhulme Trust funding RPG-2015-317. Dr Langa was supported by NIA grant R01 AG053972.

    Role of the Funder/Sponsor: The funding organizations 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.

    References
    1.
    Woolf  SH, Aron  LUS.  Health in International Perspective: Shorter Lives, Poorer Health. National Academies Press; 2013.
    2.
    Preston  S, Ho  J.  International Differences in Mortality at Older Ages: Dimensions and Sources. National Research Council; 2010.
    3.
    Cohen  B, Preston  SH, Crimmins  EM.  Explaining Divergent Levels of Longevity in High-Income Countries. National Academies Press; 2011.
    4.
    Case  A, Deaton  A.  Rising morbidity and mortality in midlife among white non-Hispanic Americans in the 21st century.   Proc Natl Acad Sci U S A. 2015;112(49):15078-15083. doi:10.1073/pnas.1518393112 PubMedGoogle ScholarCrossref
    5.
    Bound  J, Geronimus  AT, Rodriguez  JM, Waidmann  TA.  Measuring recent apparent declines in longevity: the role of increasing educational attainment.   Health Aff (Millwood). 2015;34(12):2167-2173. doi:10.1377/hlthaff.2015.0481 PubMedGoogle ScholarCrossref
    6.
    Martinson  ML, Teitler  JO, Reichman  NE.  Health across the life span in the United States and England.   Am J Epidemiol. 2011;173(8):858-865. doi:10.1093/aje/kwq325 PubMedGoogle ScholarCrossref
    7.
    Banks  J, Marmot  M, Oldfield  Z, Smith  JP.  Disease and disadvantage in the United States and in England.   JAMA. 2006;295(17):2037-2045. doi:10.1001/jama.295.17.2037 PubMedGoogle ScholarCrossref
    8.
    Nolte  E, McKee  CM.  In amenable mortality—deaths avoidable through health care—progress in the US lags that of three European countries.  [published correction appears in Health Aff (Millwood). 2012 Oct;31(10):2356].  Health Aff (Millwood). 2012;31(9):2114-2122. doi:10.1377/hlthaff.2011.0851 PubMedGoogle ScholarCrossref
    9.
    Avendano  M, Kawachi  I.  The search for explanations of the American health disadvantage relative to the English.   Am J Epidemiol. 2011;173(8):866-869. doi:10.1093/aje/kwq484 PubMedGoogle ScholarCrossref
    10.
    van Hedel  K, Avendano  M, Berkman  LF,  et al.  The contribution of national disparities to international differences in mortality between the United States and 7 European countries.   Am J Public Health. 2015;105(4):e112-e119. doi:10.2105/AJPH.2014.302344 PubMedGoogle ScholarCrossref
    11.
    Banks  J, Muriel  A, Smith  JP.  Disease prevalence, disease incidence, and mortality in the United States and in England.   Demography. 2010;47(1)(suppl):S211-S231. doi:10.1353/dem.2010.0008 PubMedGoogle ScholarCrossref
    12.
    Kinge  JM, Modalsli  JH, Øverland  S,  et al.  Association of household income with life expectancy and cause-specific mortality in Norway, 2005-2015.   JAMA. 2019;321(19):1916-1925. doi:10.1001/jama.2019.4329 PubMedGoogle ScholarCrossref
    13.
    Jivraj  S.  Are self-reported health inequalities widening by income? an analysis of British pseudo birth cohorts born, 1920-1970.   J Epidemiol Community Health. 2020;74(3):255-259. doi:10.1136/jech-2019-213186 PubMedGoogle ScholarCrossref
    14.
    Zaninotto  P, Batty  GD, Stenholm  S,  et al.  Socioeconomic inequalities in disability-free life expectancy in older people from England and the United States: a cross-national population-based study.   J Gerontol A Biol Sci Med Sci. 2020;75(5):906-913. doi:10.1093/gerona/glz266PubMedGoogle ScholarCrossref
    15.
    Chetty  R, Stepner  M, Abraham  S,  et al.  The association between income and life expectancy in the United States, 2001–2014.   JAMA. 2016;315(16):1750-1766. doi:10.1001/jama.2016.4226 PubMedGoogle ScholarCrossref
    16.
    OECD Data. Inequality. Poverty rate. Accessed January 16, 2020. https://data.oecd.org/inequality/poverty-rate.htm
    17.
    World Inequality Database. Accessed February 15, 2020. https://wid.world
    18.
    Case  A, Deaton  A.  Deaths of Despair and the Future of Capitalism. Princeton University Press; 2020.
    19.
    National Center for Health Statistics. Health, United States, 2014: with special feature on adults aged 55–64. National Center for Health Statistics; 2015.
    20.
    Lee  J, Smith  JP.  Health, Economic Status, and Aging in High-Income Countries. National Academies Press; 2018.
    21.
    Makaroun  LK, Brown  RT, Diaz-Ramirez  LG,  et al.  Wealth-associated disparities in death and disability in the United States and England.   JAMA Intern Med. 2017;177(12):1745-1753. doi:10.1001/jamainternmed.2017.3903 PubMedGoogle ScholarCrossref
    22.
    Choi  H, Schoeni  R. Health of Americans who must work longer to reach Social Security retirement age.  Health Aff (Millwood). 2017;36(10):1815-1819. doi:10.1377/hlthaff.2017.0217 PubMedCrossref
    23.
    Evans  W, Wolfe  B, Adler  N. The income-health gradient. Accessed October 27, 2019. https://www.irp.wisc.edu/publications/focus/pdfs/foc301b.pdf
    24.
    Sonnega  A, Faul  JD, Ofstedal  MB, Langa  KM, Phillips  JW, Weir  DR.  Cohort profile: the Health and Retirement Study (HRS).   Int J Epidemiol. 2014;43(2):576-585. doi:10.1093/ije/dyu067 PubMedGoogle ScholarCrossref
    25.
    NetCen Social Research. ELSA: user guide to the main interview datasets: waves 1 to 8. Accessed August 16, 2018. http://doc.ukdataservice.ac.uk/doc/5050/mrdoc/pdf/5050_waves_1-8_interviewer_data_user_guide_v01.pdf
    26.
    Phillips  D, Lin  Y, Wight  J, Chien  S, Lee  J. Harmonized ELSA documentation; version E, April 2017. Accessed October 16, 2018. http://doc.ukdataservice.ac.uk/doc/5050/mrdoc/pdf/5050_harmonized_elsa_e.pdf
    27.
    Schane  RE, Walter  LC, Dinno  A, Covinsky  KE, Woodruff  PG.  Prevalence and risk factors for depressive symptoms in persons with chronic obstructive pulmonary disease.   J Gen Intern Med. 2008;23(11):1757-1762. doi:10.1007/s11606-008-0749-z PubMedGoogle ScholarCrossref
    28.
    Whelton  PK, Carey  RM, Aronow  WS,  et al.  2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults: executive summary—a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines.   Circulation. 2018;138(17):e426-e483.PubMedGoogle Scholar
    29.
    Organisation for Economic Co-operation and Development.  Classifying Educational Programmes: Manual for ISCED-97 Implementation in OECD Countries. 1999 Edition. OECD Publications; 1999.
    30.
    Norton  EC, Miller  MM, Kleinman  LC.  Computing adjusted risk ratios and risk differences in Stata.   Stata J. 2013;13(3):492-509. doi:10.1177/1536867X1301300304 Google ScholarCrossref
    31.
    Cox  NJ.  Speaking Stata: smoothing in various directions.   Stata J. 2005;5(4):574-593. doi:10.1177/1536867X0500500408 Google ScholarCrossref
    32.
    Molina  T.  Adjusting for heterogeneous response thresholds in cross-country comparisons of self-reported health.   J Econ Ageing. 2017;10(C):1-20. doi:10.1016/j.jeoa.2017.08.002 Google ScholarCrossref
    ×