Lower income, higher morbidity and mortality

 A new article confirmed the relationship between low family income and high morbidity/mortality in children and adolescents, using a more robust database and study design than prior studies (see income kids and inc morbidity JAMA2022 in dropbox, or doi:10.1001/jama.2022.22778) 

  

Details: 

-- 795,000 individuals from age 5 to 17 who were enrolled in Medicaid from 2011-2012 with family income <200% of the federal poverty threshold, were followed through December 2021 

-- mean age 10.6 years (44% aged 5 to 9, 56% aged 10 to 17), 50% female, 35% white/34% Hispanic/25% non-Hispanic Black  

-- 5 states with the largest sample: California 14%/Texas 10%/Illinois 6%/New York 6%/Ohio 5% 

-- income level below median income for these Medicaid recipients: 387,000 people; higher than median income in 408,000 

-- income-to-poverty ratio: those below median income 46%, above median income 134% 

    -- this measure is a ratio of the total family income, adjusted for family size, divided by the dollar amount assessed by the federal poverty threshold <200% 

        -- the <200% for the federal poverty threshold for a family of 3 corresponded to $36,212 in 2011 and $43,622 in 2021

-- this study linked national Medicaid claims data in 2011-2012, the American Community Survey (2008-2013, a cross-sectional nationally representative household survey of 1% of the US population, with 3 to 3.5 million addresses selected annually and with a response rates 90-97%), the US Census Bureau version of the Social Security Administration death records (to measure the association between income and prevalence of medical conditions)

-- Main outcomes: prevalence of infection, mental health disorder (aggregation of ADHD, conduct disorders, and hyperkinetic syndromes; anxiety disorders; autism spectrum disorder; bipolar disorders; depressive disorders; personality disorders; PTSD; and schizophrenia/psychoses), injury, asthma, anemia, substance use disorders; including alcohol, tobacco, or drug use disorders), and death within the 10 years of observation, based on ICD-9 codes in children and adolescents enrolled in Medicaid or in the Children’s Health Insurance Program (CHIP) 

-- clinical outcomes overall: infection in 33%, mental health disorder 13%, injury 6%, asthma 5%, anemia 2%, substance use disorder 1% 

    -- overall mortality, 0.6% with mean age at death: 20yo 

-- a few states (Florida and Idaho) did not have full databases 

 

Results: 

-- prevalence for each condition/mortality assessed was lower in children and adolescents in families with above median income (p<0.001), except for substance use disorder 

    -- mortality had the largest difference: 15% lower among children and adolescents in families having income above versus below the median

    -- anemia and asthma were both 13% lower in higher income, mental health disorders 10% lower, injuries 8% lower, and infection 6% lower in families above the median

-- prevalence of conditions and mortality differed between those 5-9yo versus 10-17yo: 

    -- those aged 10 to 17 versus 5 to 9: 

        -- infection: 1.2 percentage points lower 

        -- asthma: 1.2 percentage points lower 

        -- mental health disorders: 6.6 percentage points more likely, 69% times higher 

            -- of note, for both age groups ADHD was the dominant mental health disorder, though anxiety disorders, depressive disorders, bipolar disorders, PTSD, and schizophrenia/psychosis were prevalent in both age groups) 

        -- injury: 3.9 percentage points more likely, 105% times higher 

        -- substance use disorder: 1.5 percentage points more likely, 573% times higher 

        -- 10-year mortality: 0.8 percentage points higher (more than 3-fold) 

 

-- children aged 5-9 living in a family with 100% more income relative to the federal poverty threshold: 

    -- infection: 2.3 percentage points lower, 7% lower 

        -- those 10-17yo: 12% lower 

    -- mental health disorders 1.9 percentage points lower, 19% lower 

        -- those 10-17yo: 15% lower 

    -- injury: 0.7 percentage points lower, 18% lower 

        -- those 10-17yo: 12% lower 

    -- asthma: 0.3 percentage points lower, 5% lower 

        -- those 10-17yo: 12% lower 

    -- anemia: 0.2 percentage points lower, 9% lower 

    -- substance use disorder: 0.6 percentage points lower, 22% lower 

        -- those 10-17yo: 21% lower 

    -- mortality within 10 years: no significant difference 

        -- those 10-17yo: 21% lower 

 

-- the graph below depicts the relationship between outcomes measured and family income relative to the federal poverty threshold: 

 

 

Graph: prevalence of income-to-poverty ratio, adjusted for sex, race and ethnicity, state of residence, year that age was measured, and year that income was observed. Note the much higher slopes of the curves for infection and mental health disorders for both groups, and 10-year mortality for those 10-17yo

 

Commentary: 

    -- prior studies finding a relationship between income and clinical outcomes were much more primitive and contaminated: socioeconomic status was not well measured, health status was typically self-reported, and studies were often small or limited to a single area. This current study was much more rigorous and expansive, leading to more reliable results

-- the many thousands of studies of stress (and stress would generally be higher in those with lower income) have found an association with dramatic increases in clinical disease: for example, see http://gmodestmedblogs.blogspot.com/2022/01/stress-induced-cardiovascular-disease.html for studies on cardiovascular disease 

     -- stress is associated with an array of adverse effects on physical and mental health: for a review, see stress and health JHealth  and Social Behavior2010 in dropbox or DOI: 10.1177/0022146510383499

     -- stress is also associated with sleep disorders, which themselves are associated with significant adverse cardiometabolic and cognitive problems: see https://link.springer.com/article/10.1007/s11920-013-0418-8 

-- one real advantage of this study is they assessed people who had insurance (Medicaid) and who were consistently enrolled for a long period of time (eliminating the bias of people coming in and out of Medicaid and at times having poorer insurance coverage over the 10 years) 

-- this study tracked those enrolled in Medicaid for 10 years. A few relevant points: 

    -- we cannot extend the results of this study to those without health insurance: those living in poverty without Medicaid have much less access to health care and likely worse clinical outcomes 

        -- there are some data suggesting that getting on Medicaid (and likely with the current Medicaid expansion) leads to more access to medical care; one study found that there was an association with improved outcomes in pregnant women and infants  (see medicaid expansion health outcomes HealthAffairs2022 in dropbox, or doi: 10.1377/hlthaff.2021.01150 HEALTH AFFAIRS 41, NO. 1 (2022): 60–68)

            -- we need more data on adults, a group pretty consistently found to have a stronger inverse association between income and adverse clinical outcomes. And this database would need to be evaluated over many years, since measurable population differences may take years to be seen (especially in terms of routine preventative care interventions)

        -- that being said, Medicaid expansion and improved health outcomes is still likely a very small benefit as compared to the effects of poverty itself

        -- and, the narrow assessment of Medicaid participants does not allow for comparison to those of higher income.  But, per the graphs above, the curves/slopes of change with income in those with Medicaid but higher incomes was quite profound, especially for infection, mental health disorders and mortality: all of this suggests that the improvements in clinical outcomes would likely continue with increasing income (though would likely plateau at some point)

    -- the finding of more adverse effect in the older children reinforces the likelihood that the effects of stress are cumulative: the more stress over more time is increasingly detrimental. This cumulative effect of stress might also be part of the explanation why adults have a stronger relationship between stress and adverse clinical outcomes than kids

 

Limitations: 

-- there were some notable differences in the demographics of average family income in families having income above the median: they were 8.2 percentage points less likely to be non-Hispanic Black, 7.2 percentage points more likely to be non-Hispanic white (p<0.001), and 0.3 percentage points more likely to be Hispanic (p=0.02) 

-- the study was a retrospective one assessing associations between income and clinical events, and therefore does not have the rigor of a randomized controlled trial in proving causality (ie, there are likely unanticipated confounders that could alter the results, eg diet, exercise, etc) 

-- there is a potential for reverse causation here: were some of those on Medicaid because they had health problems interfering with their ability to work?  To the extent this is true, it would bias the results 

-- this was a non-random sample of Medicaid recipients, with some states over-represented and some not represented (Florida and Idaho were not included because of insufficient databases)

-- there was no information about children <5yo (the authors felt this group was likely uncercounted by the Census Bureau) 

-- diseases were only recorded if the individual “sought formal health care and received a diagnosis and for which Medicaid claims were generated”, possibly missing some clinical outcomes and relying on billing systems to work well 

-- the assessment of income relied on the initial baseline information, and this possibly changed over the 10-year period. And the income was self-reported and potentially inaccurate 

-- and, the other limitations noted above by including only Medicaid patients 

 

So, a few comments 

-- not so surprising that the old and less reliable studies on income and health outcomes were validated in this higher quality study
-- the role of stress in disease, both physical and mental, as is really clear in the myriad of articles published, is profound. And low income is a pretty great stressor for most people 

    -- the prior articles on stress, however, are a bit mixed. It seems that a person's perception of stress and their level of social support tends to moderate their physiologic reaction, both in terms of quantifying levels of the stress hormones (eg cortisol and catecholamines) as well as measured blood pressure.  For example, a study done in 2 states several decades ago found that people living in poorer areas who felt that their areas were bad (as measured by a Social Insecurity Scale asking about perceptions of crime in the neighborhood, crowded housing, poor housing stock, fears of going outside, etc) had higher blood pressure than those feeling that their areas were okay. 

    -- and other older studies have found that people with higher levels of personal supports had lower levels of biomarkers of stress 

-- this article also points to the likely importance of cumulative stress being a driver in adverse outcomes. Animal studies have found that acute stressors lead to transient increases in blood pressure, but chronic continued acute stress led to sustained hypertension. There have been several suggestive human studies: eg a study where workers who were moved to a more stressful job in a workplace reorganization had more cardiac symptoms and increased left ventricular mass vs those moved to a less stressful work situation (and sympathetic stimulation/catecholamines are associated with increased left ventricular mass: https://academic.oup.com/eurheartj/article-abstract/10/11/1036/411112?redirectedFrom=fulltext )

-- though there are real benefits to having health insurance, it is pretty clear (though not well studied) that the main issue here is poverty/income inequality

    -- so, obviously, poverty needs to be addressed as a basic public health issue....  (as well as a basic social one)

geoff