Research Library

Sets out allostatic load: the stress hormones that protect the body in the short run damage it when activation is sustained. The physiological frame behind the whole platform.

Indexes combining neuroendocrine, immune, metabolic and cardiovascular markers predict illness and death beyond what standard clinical measures predict.

Summarizes the evidence that chronic stress contributes to depression, cardiovascular disease and the course of HIV and cancer, through both behavior and biology. Short and widely cited; a starting point, not a data paper.

Chronic stress predicts first coronary events; work stress, social isolation and loneliness each raise the risk; acute stress triggers events in people with advanced atherosclerosis.

Severe childhood stress raises the risk of multiple adult chronic conditions. Adult stress effects are, in the authors' words, generally less marked than childhood stress and classic risk factors such as smoking, blood pressure and cholesterol; adult stress matters most as a trigger in people with existing plaque and as a determinant of outcome in those with existing disease.

Men in the top 1 percent of income lived 14.6 years longer than men in the bottom 1 percent; women, 10.1 years. The gap widened over the period. Among low-income people, life expectancy varied by about 4.5 years across local areas, and that variation tracked smoking and other health behaviors, not access to medical care, income inequality or local labor markets.

Twenty years after the first Whitehall study, lower employment grade still meant more angina, more ECG evidence of ischemia, more chronic bronchitis, worse self-rated health, more smoking, less exercise, less job control and less social support. The gradient ran through every grade, not only the bottom.

Socioeconomic status is associated with health at every level, not only below the poverty line, which means the mechanisms have to work across the whole range.

Job strain raised the risk of a first coronary event by 23 percent after adjustment, with a population attributable risk of 3.4 percent. The authors state the effect is much smaller than standard risk factors such as smoking.

Short sleep raised the risk of death by 12 percent; long sleep by 30 percent. No publication bias found.

Short sleep raised mortality 12 percent, diabetes 37 percent, hypertension 17 percent, cardiovascular disease 16 percent, coronary heart disease 26 percent and obesity 38 percent. Mortality rose in a straight line below six hours.

Short sleep raised new type 2 diabetes 28 percent; trouble falling asleep, 57 percent; trouble staying asleep, 84 percent. Quality of sleep predicted diabetes as strongly as quantity.

Defines the adult recommendation of seven or more hours and the short-sleep measure used in PLACES.

The two-item PHQ-2 (depressed mood and loss of interest over two weeks) at a score of 3 or more detected major depression with 83 percent sensitivity and 92 percent specificity. HeadWater's interest and mood items follow this form.

One in five patients had an anxiety disorder; 41 percent of them were receiving no treatment. The two core GAD items (nervous or on edge; unable to stop worrying) screened well for all four anxiety disorders studied. HeadWater's anxiety item follows this form.

The Perceived Stress Scale, the most used measure of how uncontrollable and overloaded people find their lives. HeadWater's stress, overwhelm and control items are in its family.

People with stronger social relationships had 50 percent higher odds of survival over the follow-up period, consistent across age, sex, initial health and cause of death. The effect was strongest for measures of social integration and weakest for simply living with others.

After controlling for confounders, social isolation raised the likelihood of death 29 percent, loneliness 26 percent and living alone 32 percent. Effects were larger in samples under age 65.

A higher sense of purpose in life was associated with 17 percent lower all-cause mortality and 17 percent fewer cardiovascular events, consistent across countries, questionnaires and age. Source for HeadWater's meaning item.

Compared with none, four or more categories of childhood abuse or household dysfunction meant 4 to 12 times the risk of alcoholism, drug abuse, depression and suicide attempt; 2 to 4 times the risk of smoking and poor self-rated health; and a graded rise in heart disease, cancer, lung disease, fractures and liver disease. Source of HeadWater's childhood experiences item.

Four or more ACEs raised every health outcome studied: weak to modest for inactivity, obesity and diabetes; moderate for smoking, heavy drinking, cancer, heart disease and respiratory disease; strong for mental illness and problem drinking; strongest, over sevenfold, for problem drug use and interpersonal and self-directed violence.

CDC's definition of ACEs and the national prevalence figures from BRFSS.

Each year about 4 to 16 percent of children are physically abused and one in ten is neglected or psychologically abused; official substantiated rates capture less than a tenth of this. Maltreatment has lasting effects on mental health, drug and alcohol misuse, obesity and criminal behavior. Neglect is at least as damaging in the long term as physical or sexual abuse and has received the least attention.

Compared with none: emotional abuse tripled the odds of depression (3.06) and suicide attempts (3.37); physical abuse raised depression 54 percent, drug use 92 percent and suicide attempts 3.4-fold; neglect doubled depression (2.11) and raised suicide attempts 95 percent. Evidence for chronic physical disease was suggestive, not confirmed.

A history of sexual abuse was associated with roughly triple the odds of anxiety disorder (3.09), 2.7 times depression and eating disorders, 2.3 times PTSD and 4.1 times suicide attempts, regardless of the victim's sex or age at the abuse. No association with schizophrenia.

Self-reported childhood sexual abuse: 12.7 percent overall, 18 percent of women and 7.6 percent of men. Informant studies find about one thirtieth of that, which shows how much depends on who is asked. Context for HeadWater's anonymous self-report design.

Children who witnessed violence between their parents had significantly worse outcomes than children who did not (d = 0.40) and than children from verbally aggressive homes, and outcomes not significantly different from children who were themselves physically abused. Effects were similar across outcome types, slightly larger for preschoolers.

Exposure to intimate partner violence predicted later behavior and emotional problems, and the association grew stronger over time. Exposure at younger ages predicted more behavior problems. Child sex did not change the result.

One in five grew up with a parent who abused alcohol, and those homes carried far more of the other adverse experiences. Adult alcoholism was higher among children of alcoholics at every ACE count. Adult depression among children of alcoholics was explained by the other adverse experiences in those homes, not by the parental drinking itself.

One third of children of a parent with severe mental illness developed a severe mental illness themselves by adulthood, two and a half times the control rate, and the risk extended beyond the parent's own disorder to other types.

After confounders, children of a parent who died by suicide were about twice as likely to die by suicide (adjusted OR 1.94), and children of a parent who attempted suicide were about twice as likely to attempt it (1.95). Maternal suicidal behavior and younger age at exposure carried larger effects; sons and daughters did not differ.

Parental death in childhood raised later hospital-treated self-harm and violent crime, after adjusting for relatives' mental illness. Losing both parents, unnatural causes of death, and death during the preschool years carried the largest increases. Not a HeadWater item on its own; included because suicide in the home is also a parental death.

Children of divorce scored lower across well-being measures, with a median effect of 0.14 of a standard deviation, a small effect. The most consistent support was for family conflict, not parental absence or lost income, as the reason.

The same gaps in achievement, conduct, adjustment, self-concept and social relations held a decade later. The gap narrowed in the 1980s and widened again in the 1990s. Small effects, consistent across forty years of studies.

In the most rigorous studies, parental incarceration was associated with about 10 percent higher risk of antisocial behavior (OR 1.4) after accounting for parental criminality, but not with mental health problems, drug use or poor school performance. The authors rate much of the underlying research as poor quality. An honest entry: the effect is real and smaller than often claimed.

Famine exposure in the womb was associated with glucose intolerance at every stage of gestation; early-gestation exposure with more coronary heart disease, worse lipids, disturbed clotting, higher stress responsiveness, more obesity and, in women, more breast cancer. Effects depended on timing.

Across famine studies, prenatal famine is consistently tied to adult body size, diabetes and schizophrenia. For other outcomes the findings are less robust, and the epigenetic mechanism remains a hypothesis. This is why the shelf is labeled evidence developing.

Food insecurity is consistently associated with worse health after confounders are controlled: food-insecure children at least twice as likely to be in fair or poor health; food-insecure seniors with limitations comparable to food-secure seniors fourteen years older.

Food insecurity was associated with 20 percent higher hypertension and 30 percent higher hyperlipidemia. The association with diabetes was not statistically significant except under a strict definition of food insecurity.

Eleven million deaths in 2017 were attributable to diet, led by high sodium, low whole grains and low fruit. Dietary data were uneven across countries and the authors flag the resulting uncertainty.

Using conservative assumptions, inactivity accounts for 6 percent of coronary heart disease, 7 percent of type 2 diabetes, 10 percent of breast and colon cancer and 9 percent of premature deaths, about 5.3 million deaths in 2008.

CDC's summary of the health benefits of physical activity and the adult guideline of 150 minutes a week of moderate activity.

A comprehensive evaluation of six lines of research gives persuasive evidence that fine particulate air pollution harms cardiopulmonary health, at different time scales of exposure.

Each 10 micrograms per cubic meter of fine particulate matter raised all-cause mortality 7.3 percent, and 13.6 percent when exposure was already below the national standard. Effects were larger for men, Black beneficiaries and people eligible for Medicaid.

CDC's summary of smoking prevalence and the diseases it causes, drawing on the Surgeon General's reports.

Alcohol was the seventh leading risk factor for death in 2016 and the leading risk factor for ages 15 to 49. In this analysis the level of consumption that minimized health loss was zero drinks a week. Later analyses debate the low-dose estimate; the overall burden finding stands.

The three AUDIT-C consumption questions detected heavy drinking as well as the full ten-item AUDIT. Small and male-only; included as the source of HeadWater's drinking item, not as evidence about drinking.

About three in ten adults had a mental disorder in the past year in both decades. Treatment rose from 20 percent to 33 percent of those with a disorder, which means most people with a disorder still received no treatment.

The Behavioral Model of Health Services Use: whether people get care depends on predisposing factors, enabling resources and need. The standard frame for HeadWater's community needs and confidence items.

Life expectancy and age-adjusted death rates by race, ethnicity and sex for 2023, from the National Vital Statistics System. The current federal reference for differences in length of life.

In 2010 life expectancy for the Black population was 3.8 years lower than for the White population (4.7 years for males, 3.3 for females), accounted for by higher death rates from heart disease, cancer, homicide, diabetes, stroke and perinatal conditions. Kept for the cause-of-death breakdown; the 2023 brief carries the current figures.

The income gap in life expectancy, 14.6 years for men and 10.1 for women between the top and bottom 1 percent, with local variation tied to health behaviors.

Summarizes documented differences in health outcomes for people with disabilities and argues they meet the criteria for a health disparity population. Included for the documented differences; the policy recommendations are the authors' own.

How the CDC PLACES county estimates on the Atlas are made: small area estimation from BRFSS responses, Census population data and ACS characteristics. The estimates are modeled, not direct counts.

The exact wording of each PLACES measure on the Atlas, including the BRFSS question behind it.

The telephone survey behind both PLACES and the County Health Rankings behavior and health measures.

How the County Health Rankings are built: more than 30 county measures from national sources, z-scored, weighted and summed within each state. The authors note two-fold differences between the least and most healthy counties in a state and encourage users to check local data.

The 2025 Nebraska data file (version 3) from which the Atlas takes 34 measures.

A research finding is less likely to be true when studies are small, effects are small, many relationships are tested, designs are flexible and there is financial or other interest. The reason this library requires replication and size.

Thirty-six percent of replications were statistically significant versus 97 percent of the originals; replication effects were half the size. Replication success tracked the strength of the original evidence. The reason single studies do not make this library.