Assessment of the predictive power of a causal variable: An application to the Head Start impact study.

In a study attempting to estimate a causal effect of a causal variable, an assessment of the predictive power of the causal variable can shed light on the heterogeneity around its average effect. Using data from the Head Start Impact Study, a randomized controlled trial of the Head Start, a nation-wide early childhood education program in the United States, we provide a parallel comparison between measures of average effect and predictive power of the Head Start on five cognitive outcomes. We observed that one year of the Head Start increased scores for all five outcomes, with effect sizes ranging from 0.12 to 0.19 standard deviations. Percent variation explained by the Head Start ranged from 0.56 to 1.62%. For binary versions of the outcomes, the overall pattern remained; the Head Start on average improved the outcomes by meaningful magnitudes. In contrast, in a fully adjusted model, the Head Start only improved area under the curve (AUC) by less than 1% and its influence on the variance of predicted probabilities was negligible. The Head-Start-only model only achieved AUC ranging from 50.22 to 55.24%. Negligible predictive power despite the significant average effect suggests that the heterogeneity in effects may be large. The average effect estimates may not generalize well to different populations or different Head Start program settings. Assessment of the predictive power of a causal variable in randomized data should be a routine practice as it can provide helpful information on the causal effect and especially its heterogeneity.

Distributional effects on children's cognitive and social-emotional outcomes in the Head Start Impact Study: A quantile regression approach.

Heterogeneity in treatment effects of the Head Start, a federally funded early childhood development program in the United States, has previously been found in the Head Start Impact Study (HSIS), a nationally representative randomized controlled trial. While individual characteristics have been extensively examined as sources of effect heterogeneity, treatment effects may vary as a function of outcome distribution (i.e., distributional effect). Using quantile regressions, we investigated distributional effects of the Head Start on eight child developmental outcomes for first year, second year, third year, and the 3rd grade year follow-up in the HSIS data. For PPVT and Applied Problems, the effects varied substantially across quantiles in the first follow-up, but they were positive overall. The effects at the lower quantiles were larger and were sustained beyond the first follow-up (PPVT [95% CI] at 10th and 90th quantiles: 8.74 [6.22, 11.27], 3.32 [0.82, 5.81]) in the first follow-up and 5.72 [2.66, 8.77], -1.66 [-3.69, 0.37] in the second follow-up). For Behavior Problems, the effects were only positive for the lower quantiles in the first follow-up, but they became null in the latter follow-ups. For Letter-Word Identification, Spelling, and Pre-Academic, the effects were positive in the first follow-up with moderate variation across quantiles. In the second follow-up, only the effects at the lower quantiles were statistically significant, although they faded in the latter follow-ups. For Oral Comprehension and Social Skills, effects were null for all follow-ups. The Head Start had meaningful distributional effects for a range of child developmental outcomes, and distributional effects should be routinely assessed for better understanding of child developmental programs.

Assessment of heterogeneous Head Start treatment effects on cognitive and social-emotional outcomes.

Head Start is a federally funded, nation-wide program in the U.S. for enhancing school readiness of children aged 3-5 from low-income families. Understanding heterogeneity in treatment effects (HTE) is an important task when evaluating programs, but most attempts to explore HTE in Head Start have been limited to subgroup analyses that rely on average treatment effects by subgroups. This study applies an extension of multilevel modelling, complex variance modelling, to data from a randomized controlled trial of Head Start, Head Start Impact Study (HSIS). The treatment effects on the variance, in addition to the mean, of nine cognitive and social-emotional outcomes were assessed for 4,442 children aged 3-4 years who were followed until their 3rd grade year. Head Start had positive short-term effects on the means of multiple cognitive outcomes while having no effect on the means of social-emotional outcomes. Head Start reduced the variances of multiple cognitive and one social-emotional outcomes, meaning that substantial HTE exists. In particular, the increased mean and decreased variance reflect the ability of Head Start to improve the outcomes and reduce their variability. Exploratory secondary analyses suggested that larger benefits for children with Spanish as a primary language and low parental educational level partly explained the reduced variability, but the HTE remained and the variability was reduced even within these subgroups. Routinely monitoring the treatment effects on the variance, in addition to the mean, would lead to a more comprehensive program evaluation that describes how a program performs on average and on the entire distribution.

Geographic variation in caesarean delivery in India.

BACKGROUND: The rate of caesarean delivery has increased markedly both globally and within India. However, there is considerable variation within countries. No previous studies have examined the relative importance of multiple geographic levels in shaping the distribution of caesarean delivery and to what extent they can be explained by individual-level risk factors. OBJECTIVES: To describe geographic variation in caesarean delivery and quantify the contribution of individual-level risk factors to the variation in India. METHODS: We conducted four-level logistic regression analysis to partition total variation in caesarean delivery to three geographic levels (states, districts and communities) and quantify the extent to which variance at each level was explained by a set of 20 sociodemographic, medical and institutional risk factors. Stratified analyses were conducted by the type of delivery facility (public/private). RESULTS: Overall prevalence of caesarean delivery was 19.3% in India in 2016. Most geographic variation was attributable to states (44%), followed by communities (32%), and lastly districts (24%). Adjustment for all risk factors explained 44%, 52% and 46% of variance for states, districts and communities, respectively. The proportion explained by individual risk factors was larger in public facilities than in private facilities at all three levels. A substantial proportion of between-population variation still existed even after clustering of individual risk factors was comprehensively adjusted for. CONCLUSIONS: Diverse contextual factors driving high or low rate of caesarean delivery at each geographic level should be explored in future studies so that tailored intervention can be implemented to reduce the overall variation in caesarean delivery.

The relative importance of households as a source of variation in child malnutrition: a multilevel analysis in India.

BACKGROUND: Child malnutrition remains a major public health issue in India. Along with myriad upstream and social determinants of these adverse outcomes, recent studies have highlighted regional differences in mean child malnutrition rates. This research helps policy makers look between urban and rural communities and states to take a population-level approach to addressing the root causes of child malnutrition. However, one gap in this between-population approach has been the omission of households as a unit of analysis. Households could represent important sources of variation in child malnutrition within communities, districts, and states. METHODS: Using the fourth round of India's National Family Health Survey from 2015 to 2016, we analyzed four and five-level multilevel models to estimate the proportion of variation in child malnutrition attributable to states, districts, communities, households, and children. RESULTS: Overall, we found that of the four levels that children were nested in (households, communities, districts, and states), the greatest proportion of variation in child height-for-age Z score, weight-for-age Z score, weight-for-height Z score, hemoglobin, birthweight, stunting, underweight, wasting, anemia, and low birthweight was attributable to households. Furthermore, we found that when the household level is omitted from models, the variance estimates for communities and children are overestimated. CONCLUSIONS: These findings highlight the importance of households as an important source of clustering and variation in child malnutrition outcomes. As such, policies and interventions should address household-level social determinants, such as asset and social deprivations, in order to prevent poor child growth outcomes among the most vulnerable households in India.

Treatment effect heterogeneity in the head start impact study: A systematic review of study characteristics and findings.

There have been consistent efforts to assess treatment effect heterogeneity (TEH) of Head Start using the data from the Head Start Impact Study (HSIS), a randomized controlled trial of a federally funded child development program for a nationally representative sample of low-income parents and their 3- and 4-year-old children in the United States. Including 28 studies on TEH of Head Start, this review found that multiple high-risk subgroups (e.g., children with lower cognitive abilities, Spanish-speaking dual language learners) experienced larger gains across a range of developmental and parental outcomes, but mixed results for several subgroups. Most studies focused on subgroup analyses, cognitive and social-emotional outcomes, and short-term effects. Further studies on distributional effects, health and parental outcomes, and long-term effects are warranted. Finally, suggestions for future research on TEH of Head Start are discussed, which are applicable to other child development programs and policy evaluations.

Multilevel analysis of geographic variation among correlates of child undernutrition in India.

Prior research has identified a number of risk factors ranging from inadequate household sanitation to maternal characteristics as important determinants of child malnutrition and health in India. What is less known is the extent to which these individual-level risk factors are geographically distributed. Assessing the geographic distribution, especially at multiple levels, matters as it can inform where, and at what level, interventions should be targeted. The three levels of significance in the Indian context are villages, districts, and states. Thus, the purpose of this paper was to (a) examine what proportion of the variation in 21 risk factors is attributable to villages, districts, and states in India and (b) elucidate the specific states where these risk factors are clustered within India. Using the fourth National Family Health Survey dataset, from 2015 to 2016, we found that the proportion of variation attributable to villages ranged from 14% to 63%, 10% to 29% for districts and 17% to 62% for states. Furthermore, we found that Bihar, Jharkhand, Madhya Pradesh, and Uttar Pradesh were in the highest risk quintile for more than 10 of the risk factors included in our study. This is an indication of geographic clustering of risk factors. The risk factors that are clustered in states such as Bihar, Jharkhand, Madhya Pradesh and Uttar Pradesh underscore the need for policies and interventions that address a broader set of child malnutrition determinants beyond those that are nutrition specific.

Morbidity compression or expansion? A temporal analysis of the age at onset of non-communicable diseases in India.

While there is evidence of morbidity compression in many countries, temporal patterns of non-communicable diseases (NCDs) in developing countries, such as India, are less clear. Age at onset of disease offers insights to understanding epidemiologic trends and is a key input for public health programs. Changes in age at onset and duration of major NCDs were estimated for 2004 (n = 38,044) and 2018 (n = 43,239) using health surveys from the India National Sample Survey (NSS). Survival regression models were used to compare trends by sociodemographic characteristics. Comparing 2004 to 2018, there were reductions in age at onset and increases in duration for overall and cause-specific NCDs. Median age at onset decreased for NCDs overall (57 to 53 years) and for diabetes, hypertension, heart disease, asthma, mental diseases, eye disease, and bone disease in the range of 2-7 years and increased for cancer, neurological disorders, some genitourinary disorders, and injuries/accidents in the range of 2-14 years. Hazards of NCDs were higher among females for cancers (HR 1.51, 95% CI 1.19-1.90) and neurological disorders (HR 1.18, 95% CI 1.06-1.32) but lower for heart diseases (HR 0.88, 95% CI 0.79-0.97) and injuries/accidents (HR 0.87, 95% CI 0.77-0.99). Hazards were greater among those with lower educational attainment at younger ages and higher educational attainment later in life. Unlike many countries, chronic disease morbidity may be expanding in India for many chronic diseases, indicating excess strain on the health system. Public health programs should focus on early diagnosis and prevention of NCDs.

The relative contributions of behavioral, biological, and psychological risk factors in the association between psychosocial stress and all-cause mortality among middle- and older-aged adults in the USA.

Evidence of an association between psychosocial stress and mortality continues to accumulate. However, despite repeated calls in the literature for further examination into the physiological and behavioral pathways though which stress affects health and mortality, research on this topic remains limited. This study addresses this gap by employing a counterfactual-based mediation analysis of eight behavioral, biological, and psychological pathways often hypothesized to play a role in the association between stress and health. First, we calculated the survival rate of all-cause mortality associated with cumulative psychosocial stress (high vs. low/moderate) using random effects accelerated failure time models among a sample of 7108 adults from the Midlife in the United States panel study. Then, we conducted a multiple mediator mediation analysis utilizing a counterfactual regression framework to determine the relative contributions of each mediator and all mediators combined in the association between stress and mortality. Exposure to high psychosocial stress was associated with a 0.76 times reduced survival rate over the follow-up period 1995-2015, while adjusting for age, sex, race, income, education, baseline health, and study design effects. The mediators accounted for 49% of this association. In particular, smoking, sedentary behavior, obesity/BMI, and cardiovascular disease displayed significant indirect effects and accounted for the largest reductions in the total effect of stress on mortality, with natural indirect effects of 14%, 12%, 11%, and 4%, respectively. In conclusion, traditional behavioral and biological risk factors play a significant role in the association between psychosocial stress and mortality among middle and older adults in the US context. While eliminating stress and the socioeconomic disparities that so often deliver people into high-stress scenarios should be the ultimate goal, public health interventions addressing smoking cessation, physical activity promotion, and cardiovascular disease treatment may pay dividends for preventing premature mortality in the near-term.

Developmental timing of initial racial discrimination exposure is associated with cardiovascular health conditions in adulthood.

OBJECTIVE: To examine the association between developmental timing of initial exposure to racial discrimination and cardiovascular health conditions. DESIGN: Using data from the 1995 Detroit Area Study, logistic and negative binomial regression models were used to assess the association between timing of initial exposure to racial/ethnic discrimination, classified as early childhood (0-7), childhood (8-12), adolescence (13-19), and adulthood (>19), on physician-diagnosed cardiovascular health conditions during adulthood. Each analysis adjusted for age, gender, race/ethnicity, income, education, marital status, health-related behaviors, and pre-existing health conditions. RESULTS: Of the 1,106 participants in the final sample, 520 identified as White and 586 identified as Black. Over half (64%) of the sample experienced at least one major cardiovascular health event at the time of the study, with 39% reporting two or more events. Results from logistic regression models showed that initial exposure to racial discrimination during early childhood was associated with a 2.96 (95%CI:1.15, 7.83) times greater odds of having any cardiovascular-related health condition later in life compared to individuals who reported no discrimination. Results from negative binomial regression models demonstrated that individuals who reported initial exposure to racial discrimination during early childhood and adolescence had a CVD incidence rate that was 1.63 (95%CI:1.11, 2.38) and 1.37 (95%CI:1.10, 1.69) times higher than individuals who reported no discrimination. CONCLUSION: Initial exposure to racial discrimination in early childhood and adolescence may increase the risk of cardiovascular conditions later in life. Clinicians and researchers should consider racial discrimination during childhood as a possible risk factor for illness and disease.

Discrimination and systemic inflammation: A critical review and synthesis.

Exposure to discrimination or unfair treatment has emerged as an important risk factor for illness and disease that disproportionately affects racial and ethnic minorities. Discriminatory experiences may operate like other stressors in that they activate physiological responses that adversely affect the maintenance of homeostasis. Research suggests that inflammation plays a critical role in the pathophysiology of stress-related diseases. Recent findings on discrimination and inflammation are discussed. We highlight limitations in the current evidence and provide recommendations for future studies that seek to examine the association between discrimination and inflammation.

Explaining Within- vs Between-Population Variation in Child Anthropometry and Hemoglobin Measures in India: A Multilevel Analysis of the National Family Health Survey 2015-2016.

BACKGROUND: The complex etiology of child growth failure and anemia-commonly used indicators of child undernutrition-involving proximate and distal risk factors at multiple levels is generally recognized. However, their independent and joint effects are often assessed with no clear conceptualization of inferential targets. METHODS: We utilized hierarchical linear modeling and a nationally representative sample of 139,116 children aged 6-59 months from India (2015-2016) to estimate the extent to which a comprehensive set of 27 covariates explained the within- and between-population variation in height-for-age, weight-for-age, weight-for-height, and hemoglobin level. RESULTS: Most of the variation in child anthropometry and hemoglobin measures was attributable to within-population differences (80-85%), whereas between-population differences (including communities, districts, and states) accounted for only 15-20%. The proximate and distal covariates explained 0.2-7.5% of within-population variation and 2.1-34.0% of between-population variation, depending on the indicator of interest. Substantial heterogeneity was observed in the magnitude of within-population variation, and the fraction explained, in child anthropometry and hemoglobin measures across the 36 states/union territories of India. CONCLUSIONS: Policies and interventions aimed at reducing between-population inequalities in child undernutrition may require a different set of components than those concerned with within-population inequalities. Both are needed to promote the health of the general population, as well as that of high-risk children.

Association between numeracy and self-rated poor health in 33 high- and upper middle- income countries.

The association between numeracy proficiency and health outcomes has been the subject of several studies. However, it is not known if this association is independent of educational attainment and literacy proficiency. In this study, we used logistic regression to model numeracy proficiency as a predictor of self-rated poor health after accounting for educational attainment and literacy proficiency. The prevalence of self-rated poor health among 166,863 adults aged 16-65 years from 33 high- and upper middle-income countries was 24%. Compared to those with the highest numeracy proficiency (level 4), the odds ratio of self-rated poor health for those with the lowest numeracy proficiency (level 1) was 2.2 (95% CI 1.9-2.7) and attenuated to 1.8 (95% CI 1.5-2.1) and 1.5 (95% CI 1.1, 2.0), respectively, after sequential addition of self-education and literacy proficiency. For those who were assessed to have low levels of both numeracy and literacy proficiency, the odds ratio of self-rated poor health was 1.4 (95% CI 1.3 to 1.5), relative to those who had high levels of both numeracy and literacy proficiencies. Numeracy and literacy proficiencies show both independent and interdependent correlations with poor self-rated health. Further, these associations varied by sociodemographic characteristics and across countries. Policies aimed at improving numeracy and literacy may be beneficial in preventing adverse health outcomes.

County-level housing affordability in relation to risk factors for cardiovascular disease among middle-aged adults: The National Longitudinal Survey of Youths 1979.

BACKGROUND: Housing is a fundamental social determinant of health yet housing affordability has diminished over much of the twenty-first century. Research on housing affordability as a determinant of health is limited, but studies to date have shown correlations with mental health. However, few studies have examined the relationship between housing affordability and risk factors for cardiovascular disease, the leading cause of morbidity and mortality among Americans. METHODS: Using a nationally-representative sample of middle-aged adults from the National Longitudinal Survey of Youths 1979 (NLSY79) and exploiting quasi-experimental variation before and after the Great Recession, we estimated the associations between the change in median county-level percentage of household income spent on housing (rent/mortgage) between 2000 and 2008 and individual-level risks of incident hypertension, obesity, diabetes, and depression from 2008 to 2014. We employed conditional fixed effects logistic regression models to reduce bias due to time-invariant confounding. RESULTS: Each percentage point increase in county-level median percentage of household income spent on housing was associated with a 22% increase in the odds of incident hypertension (OR = 1.22, 95% CI = 1.06 to 1.42; p = 0.01), a 37% increased odds of obesity (OR = 1.37, 95% CI = 1.00-1.87; p = 0.049), and a 15% increased odds of depression (OR = 1.15, 95% CI = 1.01-1.31; p = 0.03), controlling for individual- and area-level factors. These associations were stronger among renters than homeowners, and among men compared to women. CONCLUSIONS: Our findings suggest that lower levels of housing affordability contribute to worse risk profiles for cardiovascular disease. Policies that make housing more affordable may help to reduce the population burden of cardiovascular disease.

Social capital and physical health: An updated review of the literature for 2007-2018.

PURPOSE: Social capital is frequently indicated as a determinant of population health. Despite an increase in the frequency of public health studies including such measures, our understanding of social capital's effects on health remains unclear. In 2008, a systematic review of the "first decade" of research on social capital and health was published in the textbook Social Capital and Health. Our study intends to update and expand upon this original review to account for developments in the literature over the second decade of research on social capital and health. METHODS: We employed a systematic review of empirical studies investigating the relationship between measures of social capital and physical health outcomes published between January 1, 2007 and December 31, 2018. To identify potential studies, we conducted searches of PubMed, Embase, and PsychINFO databases in January 2019 using combinations of "social capital" and "physical health" search terms. RESULTS: We identified 1,608 unique articles and reviewed 145 studies meeting our inclusion criteria. The most frequently examined health condition was self-reported health (57%), followed by mortality (12%), cardiovascular diseases (10%), obesity (7%), diabetes (6%), infectious diseases (5%), and cancers (3%). Of these studies, 127 (88%) reported at least partial support for a protective association between social capital and health. However, only 41 (28%) reported exclusively positive findings. The majority (59%) of results were mixed, suggesting a nuanced relationship between social capital and health. This finding could also be indicative of differences in study design, which showed substantial variation. CONCLUSIONS: Despite limitations in the literature, our review chronicles an evolution in the field of social capital and health in terms of size and sophistication. Overall, these studies suggest that social capital may be an important protective factor for some physical health outcomes, but further research is needed to confirm and clarify these findings.

A Cluster Randomized Controlled Trial of a Total Worker Health® Intervention on Commercial Construction Sites.

This study evaluated the efficacy of an integrated Total Worker Health® program, "All the Right Moves", designed to target the conditions of work and workers' health behaviors through an ergonomics program combined with a worksite-based health promotion Health Week intervention. A matched-pair cluster randomized controlled trial was conducted on ten worksites (five intervention (n = 324); five control sites (n = 283)). Worker surveys were collected at all sites pre- and post- exposure at one- and six-months. Linear and logistic regression models evaluated the effect of the intervention on pain and injury, dietary and physical activity behaviors, smoking, ergonomic practices, and work limitations. Worker focus groups and manager interviews supplemented the evaluation. After controlling for matched intervention and control pairs as well as covariates, at one-month following the ergonomics program we observed a significant improvement in ergonomic practices (B = 0.20, p = 0.002), and a reduction in incidences of pain and injury (OR = 0.58, p = 0.012) in the intervention group. At six months, we observed differences in favor of the intervention group for a reduction in physically demanding work (B = -0.25, p = 0.008), increased recreational physical activity (B = 35.2, p = 0.026) and higher consumption of fruits and vegetables (B = 0.87, p = 0.008). Process evaluation revealed barriers to intervention implementation fidelity and uptake, including a fissured multiemployer worksite, the itinerant nature of workers, competing production pressures, management support, and inclement weather. The All the Right Moves program had a positive impact at the individual level on the worksites with the program. For the longer term, the multi-organizational structure in the construction work environment needs to be considered to facilitate more upstream, long-term changes.

Effects of exposure to anti-vaping public service announcements among current smokers and dual users of cigarettes and electronic nicotine delivery systems.

INTRODUCTION: Anti-vaping public service announcements (PSAs) are intended to discourage vaping or use of electronic nicotine delivery systems (ENDS). However, vaping portrayals in PSAs may have unintended effects if they increase smoking or vaping urges. This study examined benefits and unintended effects of anti-vaping PSAs with vapor portrayals on smoking and vaping-related outcomes. METHODS: Young adult smokers (N = 171) and dual users (N = 122) aged 21-30 years were randomly assigned to view: 1) anti-vaping PSAs with vapor; 2) anti-vaping PSAs without vapor; 3) physical activity PSAs; or 4) anti-smoking PSAs with smoking cues. Outcomes were changes in vaping and smoking urges before and after viewing PSAs, post-test vaping and smoking intentions in the next hour, and post-test intention to purchase ENDS and traditional cigarettes. RESULTS: Smokers only: Exposure to anti-vaping PSAs with vapor (vs. physical activity) was associated with lower intention to vape and to purchase ENDS (ps < 0.001) and lower intention to smoke and purchase cigarettes (ps < 0.05). Exposure to anti-vaping PSAs with vapor (vs. PSAs without vapor and vs. anti-smoking PSAs with smoking cues) was associated with lower intention to vape in the next hour (ps < 0.05). Exposure to anti-vaping PSAs without vapor (vs. physical activity) was associated with lower change in vaping urge (p < 0.05) and intention to purchase ENDS (p < 0.001). Dual users: Exposure to anti-vaping PSAs without vapor (vs. anti-smoking PSAs) was associated with lower intention to purchase ENDS (p < 0.05). CONCLUSION: Viewing anti-vaping PSAs with vapor was not associated with unintended effects and may have benefits on reducing smoking and vaping-related outcomes.

Quality Improvement Initiative to Decrease Variability of Emergency Physician Opioid Analgesic Prescribing.

INTRODUCTION: Addressing pain is a crucial aspect of emergency medicine. Prescription opioids are commonly prescribed for moderate to severe pain in the emergency department (ED); unfortunately, prescribing practices are variable. High variability of opioid prescribing decisions suggests a lack of consensus and an opportunity to improve care. This quality improvement (QI) initiative aimed to reduce variability in ED opioid analgesic prescribing. METHODS: We evaluated the impact of a three-part QI initiative on ED opioid prescribing by physicians at seven sites. Stage 1: Retrospective baseline period (nine months). Stage 2: Physicians were informed that opioid prescribing information would be prospectively collected and feedback on their prescribing and that of the group would be shared at the end of the stage (three months). Stage 3: After physicians received their individual opioid prescribing data with blinded comparison to the group means (from Stage 2) they were informed that individual prescribing data would be unblinded and shared with the group after three months. The primary outcome was variability of the standard error of the mean and standard deviation of the opioid prescribing rate (defined as number of patients discharged with an opioid divided by total number of discharges for each provider). Secondary observations included mean quantity of pills per opioid prescription, and overall frequency of opioid prescribing. RESULTS: The study group included 47 physicians with 149,884 ED patient encounters. The variability in prescribing decreased through each stage of the initiative as represented by the distributions for the opioid prescribing rate: Stage 1 mean 20%; Stage 2 mean 13% (46% reduction, p<0.01), and Stage 3 mean 8% (60% reduction, p<0.01). The mean quantity of pills prescribed per prescription was 16 pills in Stage 1, 14 pills in Stage 2 (18% reduction, p<0.01), and 13 pills in Stage 3 (18% reduction, p<0.01). The group mean prescribing rate also decreased through each stage: 20% in Stage 1, 13% in Stage 2 (46% reduction, p<0.01), and 8% in Stage 3 (60% reduction, p<0.01). CONCLUSION: ED physician opioid prescribing variability can be decreased through the systematic application of sharing of peer prescribing rates and prescriber specific normative feedback.