Using measures of race to make clinical predictions: Decision making, patient health, and fairness.

The use of race measures in clinical prediction models is contentious. We seek to inform the discourse by evaluating the inclusion of race in probabilistic predictions of illness that support clinical decision making. Adopting a static utilitarian framework to formalize social welfare, we show that patients of all races benefit when clinical decisions are jointly guided by patient race and other observable covariates. Similar conclusions emerge when the model is extended to a two-period setting where prevention activities target systemic drivers of disease. We also discuss non-utilitarian concepts that have been proposed to guide allocation of health care resources.

Poverty-related bandwidth constraints reduce the value of consumption.

Poverty confers many costs on individuals, primarily through direct material deprivation. We hypothesize that these costs may be understated: poverty may also reduce human welfare by decreasing the experiential value of what little the poor are able to consume via reduced bandwidth (cognitive resources)-exerting a de facto "tax" on the value of consumption. We test this hypothesis using a randomized controlled trial in which we experimentally simulate key aspects of poverty that impair bandwidth via methods commonly used in laboratory studies (e.g., memorizing sequences) and via introducing stressors commonly associated with life in poverty (e.g., thinking about financial security and experiencing thirst). Participants then engaged in consumption activities and were asked to rate their enjoyment of these activities. Consistent with our hypothesis, the randomly assigned treatments designed to reduce bandwidth significantly and meaningfully reduced ratings of the consumption activities, with the strongest effects on the consumption of food. Our results shed additional light on how the consequences of poverty on human welfare may compound and motivate future work on the full scope of returns to poverty alleviation efforts.

Changes in Health Care Workers' Economic Outcomes Following Medicaid Expansion.

Importance: The extent to which changes in health sector finances impact economic outcomes among health care workers, especially lower-income workers, is not well known. Objective: To assess the association between state adoption of the Affordable Care Act's Medicaid expansion-which led to substantial improvements in health care organization finances-and health care workers' annual incomes and benefits, and whether these associations varied across low- and high-wage occupations. Design, Setting, and Participants: Difference-in-differences analysis to assess differential changes in health care workers' economic outcomes before and after Medicaid expansion among workers in 30 states that expanded Medicaid relative to workers in 16 states that did not, by examining US individuals aged 18 through 65 years employed in the health care industry surveyed in the 2010-2019 American Community Surveys. Exposure: Time-varying state-level adoption of Medicaid expansion. Main Outcomes and Measures: Primary outcome was annual earned income; secondary outcomes included receipt of employer-sponsored health insurance, Medicaid, and Supplemental Nutrition Assistance Program benefits. Results: The sample included 1 322 263 health care workers from 2010-2019. Health care workers in expansion states were similar to those in nonexpansion states in age, sex, and educational attainment, but those in expansion states were less likely to identify as non-Hispanic Black. Medicaid expansion was associated with a 2.16% increase in annual incomes (95% CI, 0.66%-3.65%; P = .005). This effect was driven by significant increases in annual incomes among the top 2 highest-earning quintiles (ß coefficient, 2.91%-3.72%), which includes registered nurses, physicians, and executives. Health care workers in lower-earning quintiles did not experience any significant changes. Medicaid expansion was associated with a 3.15 percentage point increase in the likelihood that a health care worker received Medicaid benefits (95% CI, 2.46 to 3.84; P < .001), with the largest increases among the 2 lowest-earning quintiles, which includes health aides, orderlies, and sanitation workers. There were significant decreases in employer-sponsored health insurance and increases in SNAP following Medicaid expansion. Conclusion and Relevance: Medicaid expansion was associated with increases in compensation for health care workers, but only among the highest earners. These findings suggest that improvements in health care sector finances may increase economic inequality among health care workers, with implications for worker health and well-being.

Death by Robots? Automation and Working-Age Mortality in the United States.

The decline of manufacturing employment is frequently invoked as a key cause of worsening U.S. population health trends, including rising mortality due to "deaths of despair." Increasing automation-the use of industrial robots to perform tasks previously done by human workers-is one structural force driving the decline of manufacturing jobs and wages. In this study, we examine the impact of automation on age- and sex-specific mortality. Using exogenous variation in automation to support causal inference, we find that increases in automation over the period 1993-2007 led to substantive increases in all-cause mortality for both men and women aged 45-54. Disaggregating by cause, we find evidence that automation is associated with increases in drug overdose deaths, suicide, homicide, and cardiovascular mortality, although patterns differ by age and sex. We further examine heterogeneity in effects by safety net program generosity, labor market policies, and the supply of prescription opioids.

Mortality Before and After Border Wall Construction Along the US-Mexico Border, 1990-2017.

Objectives. To evaluate changes in mortality in US counties along the US-Mexico border in which there was substantial new border wall construction after the Secure Fence Act of 2006 relative to border counties in which there was no such border wall construction. Methods. Using complete 1990 to 2017 mortality microdata and a quasi-experimental difference-in-differences design, we evaluated changes in overall (all-cause) mortality, mortality from drug overdose, and mortality from homicide in the 10 counties with substantial new border wall construction and 11 counties with no such construction. We fit a linear model, adjusting for population characteristics and county and year fixed effects, with Bonferroni adjustments for multiple comparisons. Sensitivity analyses included the addition of adjacent inland counties and modifications to the statistical model. Results. Relative to counties without substantial new border wall construction, counties in which a substantial amount of new border wall was constructed exhibited a nonsignificant 0.02-percentage-point increase (95% confidence interval [CI] = -0.06, 0.10; P > .99) in overall mortality after construction. Border wall construction was not associated with changes in either deaths from overdose or deaths from homicide. Conclusions. Wall construction along the US-Mexico border after the Secure Fence Act of 2006 was not associated with discernible changes in mortality.

Association Between County-Level Change in Economic Prosperity and Change in Cardiovascular Mortality Among Middle-aged US Adults.

Importance: After a decline in cardiovascular mortality for nonelderly US adults, recent stagnation has occurred alongside rising income inequality. Whether this is associated with underlying economic trends is unclear. Objective: To assess the association between changes in economic prosperity and trends in cardiovascular mortality in middle-aged US adults. Design, Setting, and Participants: Retrospective analysis of the association between change in 7 markers of economic prosperity in 3123 US counties and county-level cardiovascular mortality among 40- to 64-year-old adults (102 660 852 individuals in 2010). Exposures: Mean rank for change in 7 markers of economic prosperity between 2 time periods (baseline: 2007-2011 and follow-up: 2012-2016). A higher mean rank indicates a greater relative increase or lower relative decrease in prosperity (range, 5 to 92; mean [SD], 50 [14]). Main Outcomes and Measures: Mean annual percentage change (APC) in age-adjusted cardiovascular mortality rates. Generalized linear mixed-effects models were used to estimate the additional APC associated with a change in prosperity. Results: Among 102 660 852 residents aged 40 to 64 years living in these counties in 2010 (51% women), 979 228 cardiovascular deaths occurred between 2010 and 2017. Age-adjusted cardiovascular mortality rates did not change significantly between 2010 and 2017 in counties in the lowest tertile for change in economic prosperity (mean [SD], 114.1 [47.9] to 116.1 [52.7] deaths per 100 000 individuals; APC, 0.2% [95% CI, -0.3% to 0.7%]). Mortality decreased significantly in the intermediate tertile (mean [SD], 104.7 [38.8] to 101.9 [41.5] deaths per 100 000 individuals; APC, -0.4% [95% CI, -0.8% to -0.1%]) and highest tertile for change in prosperity (100.0 [37.9] to 95.1 [39.1] deaths per 100 000 individuals; APC, -0.5% [95% CI, -0.9% to -0.1%]). After accounting for baseline prosperity and demographic and health care-related variables, a 10-point higher mean rank for change in economic prosperity was associated with 0.4% (95% CI, 0.2% to 0.6%) additional decrease in mortality per year. Conclusions and Relevance: In this retrospective study of US county-level mortality data from 2010 to 2017, a relative increase in county-level economic prosperity was significantly associated with a small relative decrease in cardiovascular mortality among middle-aged adults. Individual-level inferences are limited by the ecological nature of the study.

Economic influences on population health in the United States: Toward policymaking driven by data and evidence.

Atheendar S. Venkataramani and colleagues discuss economic factors and population health in the United States.

Social distancing to slow the US COVID-19 epidemic: Longitudinal pretest-posttest comparison group study.

BACKGROUND: Social distancing measures to address the US coronavirus disease 2019 (COVID-19) epidemic may have notable health and social impacts. METHODS AND FINDINGS: We conducted a longitudinal pretest-posttest comparison group study to estimate the change in COVID-19 case growth before versus after implementation of statewide social distancing measures in the US. The primary exposure was time before (14 days prior to, and through 3 days after) versus after (beginning 4 days after, to up to 21 days after) implementation of the first statewide social distancing measures. Statewide restrictions on internal movement were examined as a secondary exposure. The primary outcome was the COVID-19 case growth rate. The secondary outcome was the COVID-19-attributed mortality growth rate. All states initiated social distancing measures between March 10 and March 25, 2020. The mean daily COVID-19 case growth rate decreased beginning 4 days after implementation of the first statewide social distancing measures, by 0.9% per day (95% CI -1.4% to -0.4%; P < 0.001). We did not observe a statistically significant difference in the mean daily case growth rate before versus after implementation of statewide restrictions on internal movement (0.1% per day; 95% CI -0.04% to 0.3%; P = 0.14), but there is substantial difficulty in disentangling the unique associations with statewide restrictions on internal movement from the unique associations with the first social distancing measures. Beginning 7 days after social distancing, the COVID-19-attributed mortality growth rate decreased by 2.0% per day (95% CI -3.0% to -0.9%; P < 0.001). Our analysis is susceptible to potential bias resulting from the aggregate nature of the ecological data, potential confounding by contemporaneous changes (e.g., increases in testing), and potential underestimation of social distancing due to spillover effects from neighboring states. CONCLUSIONS: Statewide social distancing measures were associated with a decrease in the COVID-19 case growth rate that was statistically significant. Statewide social distancing measures were also associated with a decrease in the COVID-19-attributed mortality growth rate beginning 7 days after implementation, although this decrease was no longer statistically significant by 10 days.

Correction: Social distancing to slow the US COVID-19 epidemic: Longitudinal pretest-posttest comparison group study.

[This corrects the article DOI: 10.1371/journal.pmed.1003244.].

College affirmative action bans and smoking and alcohol use among underrepresented minority adolescents in the United States: A difference-in-differences study.

BACKGROUND: College affirmative action programs seek to expand socioeconomic opportunities for underrepresented minorities. Between 1996 and 2013, 9 US states-including California, Texas, and Michigan-banned race-based affirmative action in college admissions. Because economic opportunity is known to motivate health behavior, banning affirmative action policies may have important adverse spillover effects on health risk behaviors. We used a quasi-experimental research design to evaluate the association between college affirmative action bans and health risk behaviors among underrepresented minority (Black, Hispanic, and Native American) adolescents. METHODS AND FINDINGS: We conducted a difference-in-differences analysis using data from the 1991-2015 US national Youth Risk Behavior Survey (YRBS). We compared changes in self-reported cigarette smoking and alcohol use in the 30 days prior to survey among underrepresented minority 11th and 12th graders in states implementing college affirmative action bans (Arizona, California, Florida, Michigan, Nebraska, New Hampshire, Oklahoma, Texas, and Washington) versus outcomes among those residing in states not implementing bans (n = 35 control states). We also assessed whether underrepresented minority adults surveyed in the 1992-2015 Tobacco Use Supplement to the Current Population Survey (TUS-CPS) who were exposed to affirmative action bans during their late high school years continued to smoke cigarettes between the ages of 19 and 30 years. Models adjusted for individual demographic characteristics, state and year fixed effects, and state-specific secular trends. In the YRBS (n = 34,988 to 36,268, depending on the outcome), cigarette smoking in the past 30 days among underrepresented minority 11th-12th graders increased by 3.8 percentage points after exposure to an affirmative action ban (95% CI: 2.0, 5.7; p < 0.001). In addition, there were also apparent increases in past-30-day alcohol use, by 5.9 percentage points (95% CI: 0.3, 12.2; p = 0.041), and past-30-day binge drinking, by 3.5 percentage points (95% CI: -0.1, 7.2, p = 0.058), among underrepresented minority 11th-12th graders, though in both cases adjustment for multiple comparisons resulted in failure to reject the null hypothesis (adjusted p = 0.083 for both outcomes). Underrepresented minority adults in the TUS-CPS (n = 71,575) exposed to bans during their late high school years were also 1.8 percentage points more likely to report current smoking (95% CI: 0.1, 3.6; p = 0.037). Event study analyses revealed a discrete break for all health behaviors timed with policy discussion and implementation. No substantive or statistically significant effects were found for non-Hispanic White adolescents, and the findings were robust to a number of additional specification checks. The limitations of the study include the continued potential for residual confounding from unmeasured time-varying factors and the potential for recall bias due to the self-reported nature of the health risk behavior outcomes. CONCLUSIONS: In this study, we found evidence that some health risk behaviors increased among underrepresented minority adolescents after exposure to state-level college affirmative action bans. These findings suggest that social policies that shift socioeconomic opportunities could have meaningful population health consequences.

Stigma as a fundamental hindrance to the United States opioid overdose crisis response.

Alexander Tsai and co-authors discuss the role of stigma in responses to the US opioid crisis.

Police killings and their spillover effects on the mental health of black Americans: a population-based, quasi-experimental study.

BACKGROUND: Police kill more than 300 black Americans-at least a quarter of them unarmed-each year in the USA. These events might have spillover effects on the mental health of people not directly affected. METHODS: In this population-based, quasi-experimental study, we combined novel data on police killings with individual-level data from the nationally representative 2013-15 US Behavioral Risk Factor Surveillance System (BRFSS) to estimate the causal impact of police killings of unarmed black Americans on self-reported mental health of other black American adults in the US general population. The primary exposure was the number of police killings of unarmed black Americans occurring in the 3 months prior to the BRFSS interview within the same state. The primary outcome was the number of days in the previous month in which the respondent's mental health was reported as "not good". We estimated difference-in-differences regression models-adjusting for state-month, month-year, and interview-day fixed effects, as well as age, sex, and educational attainment. We additionally assessed the timing of effects, the specificity of the effects to black Americans, and the robustness of our findings. FINDINGS: 38 993 (weighted sample share 49%) of 103 710 black American respondents were exposed to one or more police killings of unarmed black Americans in their state of residence in the 3 months prior to the survey. Each additional police killing of an unarmed black American was associated with 0·14 additional poor mental health days (95% CI 0·07-0·22; p=0·00047) among black American respondents. The largest effects on mental health occurred in the 1-2 months after exposure, with no significant effects estimated for respondents interviewed before police killings (falsification test). Mental health impacts were not observed among white respondents and resulted only from police killings of unarmed black Americans (not unarmed white Americans or armed black Americans). INTERPRETATION: Police killings of unarmed black Americans have adverse effects on mental health among black American adults in the general population. Programmes should be implemented to decrease the frequency of police killings and to mitigate adverse mental health effects within communities when such killings do occur. FUNDING: Robert Wood Johnson Foundation and National Institutes of Health.

Economic Vulnerability Among US Female Health Care Workers: Potential Impact of a $15-per-Hour Minimum Wage.

OBJECTIVES: To investigate racial/ethnic and gender inequities in the compensation and benefits of US health care workers and assess the potential impact of a $15-per-hour minimum wage on their economic well-being. METHODS: Using the 2017 Annual Social and Economic Supplement to the Current Population Survey, we compared earnings, insurance coverage, public benefits usage, and occupational distribution of male and female health care workers of different races/ethnicities. We modeled the impact of raising the minimum wage to $15 per hour with different scenarios for labor demand. RESULTS: Of female health care workers, 34.9% of earned less than $15 per hour. Nearly half of Black and Latina female health care workers earned less than $15 per hour, and more than 10% lacked health insurance. A total of 1.7 million female health care workers and their children lived in poverty. Raising the minimum wage to $15 per hour would reduce poverty rates among female health care workers by 27.1% to 50.3%. CONCLUSIONS: Many US female health care workers, particularly women of color, suffer economic privation and lack health insurance. Achieving economic, gender, and racial/ethnic justice will require significant changes to the compensation structure of health care.

Association Between Playing American Football in the National Football League and Long-term Mortality.

Importance: Studies of the longevity of professional American football players have demonstrated lower mortality relative to the general population but they may have been susceptible to selection bias. Objective: To examine the association between career participation in professional American football and mortality risk in retirement. Design, Setting, and Participants: Retrospective cohort study involving 3812 retired US National Football League (NFL) players who debuted in the NFL between 1982 and 1992, including regular NFL players (n = 2933) and NFL "replacement players" (n = 879) who were temporarily hired to play during a 3-game league-wide player strike in 1987. Follow-up ended on December 31, 2016. Exposures: NFL participation as a career player or as a replacement player. Main Outcomes and Measures: The primary outcome was all-cause mortality by December 31, 2016. Cox proportional hazards models were estimated to compare the observed number of years from age 22 years until death (or censoring), adjusted for birth year, body mass index, height, and position played. Information on player death and cause of death was ascertained from a search of the National Death Index and web-based sources. Results: Of the 3812 men included in this study (mean [SD] age at first NFL activity, 23.4 [1.5] years), there were 2933 career NFL players (median NFL tenure, 5 seasons [interquartile range {IQR}, 2-8]; median follow-up, 30 years [IQR, 27-33]) and 879 replacement players (median NFL tenure, 1 season [IQR, 1-1]; median follow-up, 31 years [IQR, 30-33]). At the end of follow-up, 144 NFL players (4.9%) and 37 replacement players (4.2%) were deceased (adjusted absolute risk difference, 1.0% [95% CI, -0.7% to 2.7%]; P = .25). The adjusted mortality hazard ratio for NFL players relative to replacements was 1.38 (95% CI, 0.95 to 1.99; P = .09). Among career NFL players, the most common causes of death were cardiometabolic disease (n = 51; 35.4%), transportation injuries (n = 20; 13.9%), unintentional injuries (n = 15; 10.4%), and neoplasms (n = 15; 10.4%). Among NFL replacement players, the leading causes of death were cardiometabolic diseases (n = 19; 51.4%), self-harm and interpersonal violence (n = 5; 13.5%), and neoplasms (n = 4; 10.8%). Conclusions and Relevance: Among NFL football players who began their careers between 1982 and 1992, career participation in the NFL, compared with limited NFL exposure obtained primarily as an NFL replacement player during a league-wide strike, was not associated with a statistically significant difference in long-term all-cause mortality. Given the small number of events, analysis of longer periods of follow-up may be informative.

Accuracy and determinants of perceived HIV risk among young women in South Africa.

BACKGROUND: HIV risk perceptions are a key determinant of HIV testing. The success of efforts to achieve an AIDS-free generation - including reaching the UNAIDS 90-90-90 target - thus depends critically on the content of these perceptions. We examined the accuracy of HIV-risk perceptions and their correlates among young black women in South Africa, a group with one of the highest HIV incidence rates worldwide. METHODS: We used individual-level longitudinal data from the Cape Area Panel Study (CAPS) from 2005 to 2009 on black African women (20-30 years old in 2009) to assess the association between perceived HIV-risk in 2005 and the probability of testing HIV-positive four years later. We then estimated multivariable logistic regressions using cross-sectional data from the 2009 CAPS wave to assess the relationship between risk perceptions and a wide range of demographic, sexual behaviour and psychosocial covariates of perceived HIV-risk. RESULTS: We found that the proportion testing HIV-positive in 2009 was almost identical across perceived risk categories in 2005 (no, small, moderate, great) (χ 2  = 1.43, p = 0.85). Consistent with epidemiologic risk factors, the likelihood of reporting moderate or great HIV-risk perceptions was associated with condom-use (aOR: 0.57; 95% CI: 0.36, 0.89; p < 0.01); having ≥3 lifetime partners (aOR: 2.38, 95% CI: 1.53, 3.73; p < 0.01); knowledge of one's partner's HIV status (aOR: 0.67; 95% CI: 0.43, 1.07; p = 0.09); and being in an age-disparate partnerships (aOR: 1.73; 95% CI: 1.09, 2.76; p = 0.02). However, the likelihood of reporting moderate or great self-perceived risk did not vary with sexually transmitted disease history and respondent age, both strong predictors of HIV risk in the study setting. Risk perceptions were associated with stigmatising attitudes (aOR: 0.53; 95% CI: 0.26, 1.09; p = 0.09); prior HIV testing (aOR: 0.21; 95% CI: 0.13, 0.35; p < 0.01); and having heard that male circumcision is protective (aOR: 0.38; 95% CI: 0.22, 0.64; p < 0.01). CONCLUSIONS: Results indicate that HIV-risk perceptions are inaccurate. Our findings suggest that this inaccuracy stems from HIV-risk perceptions being driven by an incomplete understanding of epidemiological risk and being influenced by a range of psycho-social factors not directly related to sexual behaviour. Consequently, new interventions are needed to align perceived and actual HIV risk.

Economic Opportunity, Health Behaviors, and Mortality in the United States.

OBJECTIVES: We assessed whether economic opportunity was independently associated with health behaviors and outcomes in the United States. METHODS: Using newly available, cross-sectional, county-level data from the Equality of Opportunity Project Database and vital statistics, we estimated associations between all-cause mortality rates (averaged over 2000-2012) and economic opportunity, adjusting for socioeconomic, demographic, and health system covariates. Our measure of economic opportunity was the county-average rank in the national income distribution attained by individuals born to families in the bottom income quartile. Secondary outcomes included rates of age- and race-specific mortality, smoking, obesity, hypertension, and diabetes. RESULTS: An increase in economic opportunity from the lowest to the highest quintile was associated with a 16.7% decrease in mortality. The magnitudes of association were largest for working-age adults and African Americans. Greater economic opportunity was also associated with health behaviors and risk factors. CONCLUSIONS: Economic opportunity is a robust, independent predictor of health. Future work should investigate underlying causal links and mechanisms.

Syndemics and Health Disparities: A Methodological Note.

In the theory of syndemics, diseases are hypothesized to co-occur in particular temporal or geographical contexts due to harmful social conditions (disease concentration) and to interact at the level of populations and individuals, with mutually enhancing deleterious consequences for HIV risk (disease interaction). Since its original elaboration more than 20 years ago, the epidemiological literature on syndemic problems has followed a questionable trajectory, stemming from the use of a specific type of regression model specification that conveys very little information about the theory of syndemics. In this essay we critically review the dominant approaches to modeling in the literature on syndemics; highlight the stringent assumptions implicit in these models; and describe some meaningful public health implications of the resulting analytical ambiguities. We conclude with specific recommendations for empirical work in this area moving forward.