35.06.02 · health-medicine / public-health

Health disparities and social determinants: income, race, geography, and life expectancy

stub3 tiersLean: none

Anchor (Master): Marmot, M., The Health Gap: The Challenge of an Unequal World (Bloomsbury, 2015)

Intuition Beginner

Your zip code predicts your health better than your genetic code: life expectancy can vary by 20-30 years between neighborhoods in one city. Income, race, education, and geography are "social determinants of health" — the conditions in which people are born, grow, live, work, and age. Marmot's Whitehall studies showed a social gradient: British civil servants at every grade had worse health than those above. In the US, Black women die in childbirth at three times the rate of White women, regardless of income or education. Williams showed racism, as chronic stress, gets "under the skin" via cortisol and inflammation; Krieger's ecosocial theory explains how discrimination becomes biology. The richest 1% of American men outlive the poorest by 15 years.

Visual Beginner

The diagram makes two points. First, no single layer causes health on its own: individual behavior is embedded in networks, neighborhoods, institutions, and macro-level distributions of money and power. Second, the gradient bar reframes inequity as a continuous slope rather than a threshold between "the poor" and "everyone else." Each rung up the ladder — not only escape from poverty — buys better health, which is why interventions aimed only at the very bottom cannot, by themselves, flatten the curve.

Worked example Beginner

Worked example: a fourteen-year life-expectancy gap

Two Chicago rail stations roughly seven miles apart illustrate the stakes. Near the downtown Loop, life expectancy at birth is about 83 years. In West Englewood, a predominantly Black, low-income neighborhood, it is about 69 years — a fourteen-year gap inside a single city, larger than the gap between the United States and many low-income countries.

How should we read this? Not as a genetic difference, and not as the isolated result of individual choices. The gap reflects concentrated, overlapping disadvantage: housing shaped by decades of segregation and disinvestment; fewer supermarkets and more liquor and fast-food outlets; higher exposure to violence and air pollution; hospital closures and longer travel to care; underfunded schools; and chronic, corrosive stress. Each factor multiplies the others, and each maps onto the income, race, and geography axes introduced above. The example shows why "personal responsibility" framings, while not irrelevant, cannot explain why a seven-mile move predicts fourteen years of life — only structural explanations can.

Check your understanding Beginner

Question 1: Marmot's Whitehall studies are most notable for showing that:

A) poverty alone explains poor health
B) health improves at every step up the social hierarchy, not only at the extremes
C) civil servants are healthier than the general population
D) genetics determine the social gradient

Answer: B. The gradient runs from top to bottom of the hierarchy, even among employed, non-poor office workers.

Question 2: True or false: in the United States, Black women with a college education have maternal-mortality outcomes that meet or exceed those of White women who did not finish high school.

Answer: True. The pattern persists after controlling for education and income, pointing to racism and weathering rather than deprivation alone.

Question 3: Holding the relative disparity (rate ratio) constant while the overall rate of a disease falls, the absolute disparity (rate difference) will:

A) grow without bound
B) stay the same
C) shrink
D) first rise, then fall

Answer: C. With a fixed multiplier, lower baseline rates produce a smaller absolute gap.

Question 4: The single best phrase for "the conditions in which people are born, grow, live, work, and age" is:

A) genetic risk
B) the social determinants of health
C) the epidemiologic transition
D) herd immunity

Answer: B.

Formal definition Intermediate+

Social determinants, inequity, and the social gradient

The WHO defines the social determinants of health as the conditions in which people are born, grow, live, work, and age — the housing, education, income, employment, social inclusion, and health-care access shaped by distributions of money, power, and resources. A health disparity is a measured difference in health between groups; a health inequity is a disparity that is unfair, avoidable, and unjust. Distinguishing the two is nontrivial: it requires a normative judgment about which differences are unjust and a counterfactual about what could otherwise be achieved.

The social gradient describes the monotonic association between socioeconomic position and health: each step up the hierarchy carries better health, not only at the extremes. Marmot's Whitehall studies established that employment grade predicts all-cause and cardiovascular mortality even within a narrowly employed, non-poor population, ruling out absolute deprivation as the sole driver and locating the mechanism in relative position, control, and chronic stress.

Disparity metrics

Rate ratio: $R R = r_{d} / r_{a}$ , the rate in a disadvantaged group divided by the rate in an advantaged group. Rate difference: $R D = r_{d} - r_{a}$ , the absolute gap. The Slope Index of Inequality ranks subgroups from lowest to highest socioeconomic position and regresses their health on cumulative rank, weighting by group size, to estimate the gap between the most and least advantaged. The Concentration Index analogously orders individuals by socioeconomic rank and weights their health outcomes, yielding a single number that is negative when ill health concentrates among the poor.

The population attributable fraction, $P A F = p_{e} (R R - 1) / (1 + p_{e} (R R - 1))$ , gives the share of cases that would be eliminated if the disadvantaged group had the advantaged group's rate, where $p_{e}$ is the proportion exposed. These metrics answer different questions: $R R$ compares relative risk, $R D$ measures absolute burden, and $P A F$ quantifies population impact. A single disparity may look large or small depending on which is reported, so responsible practice reports all three.

Key theorem with proof Intermediate+

Key result: absolute and relative disparities can move in opposite directions

Claim. As the overall level of a health outcome declines, a constant relative disparity (fixed $R R$ ) forces the absolute disparity ( $R D$ ) to shrink; conversely, holding $R D$ fixed forces $R R$ to grow. The two metrics can therefore trend in opposite directions, so whether a gap is "closing" depends entirely on the chosen measure.

Proof (numerical). Suppose group A (advantaged) has rate $r_{a}$ and group D (disadvantaged) has $r_{d} = 2 r_{a}$ , a twofold disparity.

Era 1: $r_{a} = 100$ , $r_{d} = 200$ . Here $R R = 2$ and $R D = 100$ .
Era 2 (overall improvement): $r_{a} = 10$ , $r_{d} = 20$ . Now $R R$ is still 2, but $R D$ has fallen from 100 to 10.

The relative gap is unchanged while the absolute gap has collapsed by 90 percent. An analyst reporting $R D$ declares the disparity nearly eliminated; one reporting $R R$ declares it unchanged. Both are correct, and the policy conclusions diverge. Reporting both is therefore obligatory. The same arithmetic shows that driving the absolute gap to zero ( $R D \to 0$ ) is compatible with an arbitrarily large relative gap as rates approach zero, and conversely that a fixed relative gap can coexist with vanishing absolute burden. Choosing the metric is a normative act, not a technical detail.

Derivation: population attributable fraction from a two-by-two table

For an exposure present in proportion $p_{e}$ of a population, with rate ratio $R R$ and baseline rate $r_{0}$ in the unexposed, the overall population rate is $r_{p o p} = r_{0} (1 + p_{e} (R R - 1))$ . Removing the excess risk would lower the population rate to $r_{0}$ . The eliminated fraction is therefore $(r_{p o p} - r_{0}) / r_{p o p} = p_{e} (R R - 1) / (1 + p_{e} (R R - 1))$ . This links individual risk to population burden and shows why a common exposure with a modest $R R$ can outweigh a rare exposure with a large $R R$ .

Exercises Intermediate+

Using a rate ratio of 3.0 for Black versus White maternal mortality and a White mortality rate of 13 per 100,000, compute the absolute gap. Then assume an intervention halves both rates; report the new $R R$ and $R D$ . Which changed, and what does this imply about declaring success?
A disparity is reported as $R R = 1.8$ and $R D = 40$ deaths per 100,000. The overall rate then drops by 70 percent while $R R$ stays constant. Report the new $R R$ and $R D$ , and explain which framing a journalist should use.
Define and distinguish health disparity, health inequity, and the social gradient, giving a concrete example of each.
Link and Phelan argue that socioeconomic status is a "fundamental cause." Give two historical episodes in which the specific mechanism linking status to mortality changed while the association persisted.
Sketch an ecosocial pathway from residential segregation to adult cardiovascular disease, naming at least four intermediary steps and the physiological mediators at each.
An algorithm trained on past health-care cost flags patients for extra care. Costs are lower for Black patients at equal need because of access barriers. Explain why the algorithm appears race-neutral yet discriminates, and propose a fix grounded in the Obermeyer audit.
Compare the United States and United Kingdom approaches to health coverage; identify two disparities each system reduces and one it does not, and explain the structural reason.

Advanced results Master

Fundamental cause theory (Link and Phelan)

Socioeconomic status is a "fundamental cause" of disease because it bundles flexible resources — money, knowledge, power, prestige, beneficial social connections — that individuals deploy to avoid risks and minimize disease consequences regardless of which specific risk factors dominate a given era. Four properties define a fundamental cause: it associates with multiple outcomes, operates through multiple pathways, relies on replaceable mechanisms, and its association with health persists across time and place.

The theory explains an awkward empirical pattern: as public health vanquishes one risk factor, status-based disparities re-emerge through new ones. When infectious disease receded, higher-status groups quit smoking first; as smoking declined, disparities reappeared through diet, exercise, and access to new treatments. Each transition preserves the association while swapping the mechanism. The policy implication is sobering: interventions that target downstream risk factors may raise average health while widening disparities, because advantaged groups adopt and benefit from them faster. Reducing disparities therefore requires acting on the resources themselves — the distribution of money, knowledge, and power — not only on the risk factors they currently enable.

Developmental origins and the life course (Barker, DOHaD)

The life-course framework traces adult disease to exposures across gestation, childhood, and working life, organized by three models: critical periods (irreversible effects during developmental windows), risk accumulation (the total burden of exposures), and pathways (early exposures set later trajectories). David Barker showed that low birth weight — a marker of prenatal deprivation — predicts adult cardiovascular disease, diabetes, and metabolic syndrome decades later. The modern developmental-origins-of-health-and-disease (DOHaD) program identifies epigenetic modification as a candidate mechanism: the fetus adapts to nutritional scarcity in ways that harm health when later nutrition is abundant.

Cumulative-disadvantage accounts add that insults compound: childhood poverty raises adult disease risk directly and indirectly by shaping adult status, which in turn shapes the next generation. This reframes disparities as multi-generational, with policy levers in prenatal care, early-childhood investment, and intergenerational mobility rather than adult behavior alone.

Ecosocial theory and embodiment (Krieger)

Nancy Krieger's ecosocial theory asks how social discrimination literally becomes biology — a process she calls embodiment. It specifies pathways of exposure (who encounters hazards), susceptibility (whose bodies are already burdened), and resistance (whose communities buffer against harm), and it insists on multilevel, historically grounded analysis spanning cells, individuals, neighborhoods, and political economies. Embodiment rejects both pure voluntarism (behavior is free choice) and crude biological determinism (race is innate).

Instead, chronic exposure to racism, exclusion, and material deprivation recalibrates stress hormones, inflames vasculature, shortens telomeres, and alters gene expression — measurable changes that mediate the social gradient. The framework makes testable predictions: embodied damage should accumulate over the life course and concentrate in groups subjected to sustained discrimination, which is exactly what cardiovascular, metabolic, and mortality data show. It also dissolves the nature/nurture dichotomy that has historically derailed debates over race and health.

Weathering and accelerated aging (Geronimus)

Arline Geronimus's weathering hypothesis holds that chronic exposure to social, economic, and political exclusion accelerates biological aging in marginalized groups — most starkly among Black women in the United States. The evidence is convergent: Black women show earlier onset of hypertension, cardiovascular damage, and telomere shortening than White women of comparable age, and maternal mortality among college-educated Black women exceeds that of White women who did not finish high school.

Weathering reframes the maternal-mortality crisis: the culprit is not poor behavior or lack of education but the cumulative physiological cost of navigating racism across decades, which makes pregnancy more dangerous for the body well before conception. The pattern persists after adjusting for income and education, which is precisely what fundamental-cause and ecosocial accounts predict and what deprivation-only models cannot explain.

Algorithmic bias in health care (Obermeyer)

When algorithms guide clinical resources, baked-in disparities scale automatically. Obermeyer and colleagues audited a widely used risk score that referred patients to high-intensity care-management programs. Because the algorithm used past health-care cost as a proxy for need, and Black patients at equal need had historically incurred lower cost (less access, less trust, fewer referrals), it systematically under-referred Black patients: at any given risk score, Black patients were sicker than White patients flagged identically. Retraining on a direct health measure rather than cost closed most of the gap.

The episode generalizes: any model trained on data produced by an unequal system will reproduce that inequality unless fairness is designed in, audited, and enforced. This links health disparities directly to computing (see 33.07.*) and AI ethics (see 20.02.06), and it warns that "objective" automation can launder discrimination into clinical authority.

Mass incarceration, segregation, and place

Structural forces beyond the clinic shape bodies at scale. Residential segregation, engineered through redlining and exclusionary zoning, concentrates poverty, pollution, food scarcity, and violence in the same neighborhoods, producing the zip-code life-expectancy gaps noted above. Mass incarceration (see 30.06.02) removes working-age adults, destabilizes families, and leaves formerly incarcerated people with chronic disease, mental illness, and reduced access to care and employment.

Each of these is a health intervention by other means: housing policy is health policy, sentencing policy is health policy, and transportation policy is health policy. Place-based experiments such as Moving to Opportunity, and Chetty's mobility atlas, show that moving children out of high-poverty neighborhoods yields large, persistent gains in adult income, health, and longevity — evidence that the disparities are environmental and remediable, not intrinsic.

Connections Master

Sociology of stratification (see 30.04.*)

Class structure (30.04.02), race and ethnicity (30.04.03), and intersectionality (Crenshaw) supply the categories that disparity research operationalizes. Oliver and Shapiro's Black Wealth/White Wealth shows that wealth gaps — not income alone — structure health opportunity across generations, which is why income-adjusted analyses still miss the durable advantage that accumulated assets confer.

Medical and sociocultural anthropology (see 31.06.*)

Paul Farmer's concept of structural violence and the preferential option for the poor reframes disparities as violence distributed by social structure rather than bad luck. Applied and community-based participatory research (Wallerstein) and community health workers (promotoras) operationalize equity in field practice, linking the clinic to the communities it serves.

Urbanization and neighborhoods (see 30.08.*)

Neighborhood-effects research explains how concentrated disadvantage operates independently of individual traits, and how urban design — transit, green space, food retail, exposure to violence — becomes a health determinant. Life-expectancy gaps within cities are products of this geography.

Stress, allostasis, and homeostasis (see 29.11.03, 35.01.02)

Allostatic load provides the measurable physiological pathway linking chronic psychosocial stress — including the stress of racism — to cardiovascular, metabolic, and immune disease. It is the principal bridge between the social environment and the biological endpoints disparity research measures.

Ethics, justice, and rights (see 20.02., 20.07.)

Rawlsian fairness and the capabilities approach of Sen and Nussbaum ground health as a capability and a matter of distributive justice; the right-to-health tradition grounds it as a legal entitlement, with direct implications for health policy and universal coverage. The framework chosen determines whether a disparity counts as an injustice or merely a regrettable difference.

Computing, AI bias, and data (see 33.07., 20.02.06, 36.)

Algorithmic risk scores, genomic databases, and wearable surveillance increasingly mediate who is seen, treated, and protected. Equity and privacy questions in these systems are now inseparable from the study of health disparities themselves.

Historical and philosophical context Master

Virchow and the political origins of social medicine

Rudolf Virchow, investigating an 1848 typhus epidemic in Upper Silesia, concluded that the cause was not a microbe but poverty, ignorance, and political exclusion — declaring that "politics is nothing else but medicine on a large scale." This founded a tradition of social medicine that treats disease as embedded in political economy, later carried by Latin American social medicine and by Paul Farmer's work on structural violence. The insight remains radical: it implies that clinicians who ignore politics are practicing incomplete medicine.

The Black Report and the rediscovery of inequality

Britain's 1980 Black Report (Douglas Black, Margaret Whitehead) documented persistent class-based health inequalities despite the National Health Service and was famously buried by the Thatcher government. It became foundational in Europe and helped launch modern social epidemiology, establishing a point that still unsettles policy: equal access to care does not by itself produce equal health, because the determinants of health lie largely outside the clinic.

Whitehall and the social gradient

Marmot's Whitehall I (1978) and Whitehall II studies tracked British civil servants for decades and found a stepwise gradient of mortality by employment grade, present even among non-manual, non-poor workers. The finding overturned the assumption that only absolute poverty harms health and motivated the gradient framing central to modern disparities research, shifting attention toward relative position, control over work, and chronic stress as biological agents.

The Commission on Social Determinants of Health

Chaired by Marmot, the WHO Commission's 2008 report Closing the Gap in a Generation argued that avoidable health inequalities are unjust and remediable through action on daily living conditions and the unequal distribution of power and resources. It made equity a central WHO objective, though implementation has consistently lagged the rhetoric, exposing the gap between recognizing social causation and mustering the political will to act on it.

Behavioral versus structural explanations

A persistent debate sets individual-behavior accounts (diet, smoking, exercise) against structural accounts (segregation, exploitation, policy). The strongest current position, consistent with fundamental-cause and ecosocial theory, is that behavior is the proximate mechanism through which structural forces act: behavior matters, but its distribution is itself produced by structure, so effective policy must act on both levels simultaneously rather than choosing between them.

Justice, capabilities, and the right to health

Philosophically, disparities raise the question of what we owe one another. Rawlsian fairness, Nozickian libertarianism, and Sen's capabilities approach give different answers, and the choice frames whether health is a commodity, a right, or a prerequisite for a flourishing life. The human-right-to-health tradition operationalizes the latter in international law; Farmer's clinical ethics operationalizes it at the bedside through a preferential option for the poor.

Bibliography Master

Gordis, L. (2019). Epidemiology (6th ed.). Elsevier/Saunders. ^{[source pending]}
Marmot, M. (2015). The Health Gap: The Challenge of an Unequal World. Bloomsbury. ^{[source pending]}
Krieger, N. (2001). "Theories for social epidemiology in the 21st century: an ecosocial perspective." International Journal of Epidemiology, 30(4), 668-677. ^{[source pending]}
Williams, D. R. (1997). "Race and Health: Basic Questions, Emerging Directions." Annals of Epidemiology, 7(5), 322-333. ^{[source pending]}
Marmot, M. G., Rose, G., Shipley, M., & Hamilton, P. J. (1978). "Employment grade and coronary heart disease in British civil servants." Journal of Epidemiology & Community Health, 32(4), 244-249.
Link, B. G., & Phelan, J. (1995). "Social Conditions as Fundamental Causes of Disease." Journal of Health and Social Behavior, 35, 80-94.
Geronimus, A. T. (1992). "The weathering hypothesis and the health of African-American women and infants." Ethnicity & Disease, 2(3), 207-221.
Commission on Social Determinants of Health (2008). Closing the Gap in a Generation: Health Equity through Action on the Social Determinants of Health. World Health Organization.
Chetty, R., Stepner, M., Abraham, S., et al. (2016). "The Association Between Income and Life Expectancy in the United States." JAMA, 315(16), 1750-1766.
Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2019). "Dissecting racial bias in an algorithm used to manage the health of populations." Science, 366(6464), 447-453.
Farmer, P. (2003). Pathologies of Power: Health, Human Rights, and the New War on the Poor. University of California Press.
Black, D., Morris, J. N., Smith, C., & Townsend, P. (1980). Inequalities in Health: The Black Report. Penguin.
Oliver, M. L., & Shapiro, T. M. (1995). Black Wealth/White Wealth: A New Perspective on Racial Inequality. Routledge.