When the primary outcome is difficult to collect, surrogate endpoint is typically used as a substitute. It is possible that for every individual, treatment has a positive effect on surrogate, and surrogate has a positive effect on primary outcome, but for some individuals, treatment has a negative effect on primary outcome. For example, a treatment may be substantially effective in preventing the stroke for everyone, and lowering the risk of stroke is universally beneficial for a longer survival time, however, the treatment may still cause death for some individuals. We define such paradoxical phenomenon as individual surrogate paradox. The individual surrogate paradox is preposed to capture the treatment effect heterogeneity, which is unable to be described by either the surrogate paradox based on average causal effect (ACE) (Chen et al., 2007) or that based on distributional causal effect (DCE) (Ju and Geng, 2010). We investigate existing surrogate criteria in terms of whether the individual surrogate paradox could manifest. We find that only the strong binary surrogate can avoid such paradox without additional assumptions. Utilizing the sharp bounds, we propose novel criteria to exclude the individual surrogate paradox. Our methods are illustrated in an application to determine the effect of the intensive glycemia on the risk of development or progression of diabetic retinopathy.

Missing data occur frequently in empirical studies in health and social sciences, often compromising our ability to make accurate inferences. An outcome is said to be missing not at random (MNAR) if, conditional on the observed variables, the missing data mechanism still depends on the unobserved outcome. In such settings, identification is generally not possible without imposing additional assumptions. Identification is sometimes possible, however, if an instrumental variable (IV) is observed for all subjects which satisfies the exclusion restriction that the IV affects the missingness process without directly influencing the outcome. In this paper, we provide necessary and sufficient conditions for nonparametric identification of the full data distribution under MNAR with the aid of an IV. In addition, we give sufficient identification conditions that are more straightforward to verify in practice. For inference, we focus on estimation of a population outcome mean, for which we develop a suite of semiparametric estimators that extend methods previously developed for data missing at random. Specifically, we propose inverse probability weighted estimation, outcome regression-based estimation and doubly robust estimation of the mean of an outcome subject to MNAR. For illustration, the methods are used to account for selection bias induced by HIV testing refusal in the evaluation of HIV seroprevalence in Mochudi, Botswana, using interviewer characteristics such as gender, age and years of experience as IVs.

The average treatment effect (ATE) is popularly used to assess the treatment effect. However, the ATE implicitly assumes a homogenous treatment effect even amongst individuals with different characteristics. In this paper, we mainly focus on assessing the treatment effect heterogeneity, which has important implications in designing the optimal individual treatment regimens and in policy making. The treatment benefit rate (TBR) and treatment harm rate (THR) have been defined to characterize the magnitude of heterogeneity for binary outcomes. When the outcomes are continuous, we extend the definitions of the TBR and THR to compare the difference between potential outcomes with a pre-specified level c. Unlike the ATE, these rates involve the joint distribution of the potential outcomes and can not be identified without further assumptions even in randomized clinical trials. In this article, we assume the potential outcomes are independent conditional on the observed covariates and an unmeasured latent variable. Under this assumption, we prove the identification of the TBR and THR in non-separable (generalized) linear models for both continuous and binary outcomes. We then propose estimators and derive their asymptotic distributions. In the simulation studies, we implement our proposed methods to assess the performance of our estimators and carry out a sensitive analysis for different underlying distribution for the latent variable. Finally, we illustrate the proposed methods in two randomized controlled trials.

When the primary outcome is hard to collect, surrogate endpoint is typically used as a substitute. However, even when the treatment has a positive average causal effect (ACE) on the surrogate endpoint, which also has a positive ACE on the primary outcome, it is still possible that the treatment has a negative ACE on the primary outcome. Such a phenomenon is called the surrogate paradox and greatly challenges the use of surrogate. In this paper, we provide novel criteria to exclude the surrogate paradox. Unlike other conditions previously proposed, our conditions are testable since they only involve observed data. Furthermore, our criteria are optimal in the sense that they are sufficient and "almost necessary" to exclude the paradox: if the conditions are satisfied, the surrogate paradox is guaranteed to be absent while if the conditions fail, there exists a data generating process with surrogate paradox that can generate the same observed data. That is, our criteria capture all the information in the observed data to exclude the surrogate paradox rather than relying on unverifiable distributional assumptions.

BACKGROUND: Antiretroviral therapy (ART) administered in clinical trial settings virtually eliminates the sexual transmission of human immunodeficiency virus (HIV) in serodiscordant couples, but effectiveness of treatment as prevention in the community is debated. Conflicting results from previous analyses in a Chinese cohort underscore the importance of determining effectiveness of ART delivered in resource limited settings.\n\nMETHODS: All available years of data (2006-2012) from local disease control records of HIV patients and their seronegative spouses in Henan Province, China, were analyzed using marginal structural Cox models to estimate the effect of ART in the initially infected partner his or her partner's HIV seroconversion risk.\n\nRESULTS: We observed 157 seroconversion events in 4916 serosdiscordant couples, for an incidence rate of 0.59 cases per 100 person-years (PY) (95% confidence interval [CI], .51-.70). Of these, 84 occurred after the index partner had initiated ART (0.43/100PY; 95% CI, .35-.53) and 73, whereas index partners were untreated (5.87/100 PY; 95% CI, 4.65-7.42). In a marginal structural Cox model weighted for confounding and censoring, the hazard ratio (HR) for HIV transmission was 0.52 (95% CI, .34-.82). ART efficacy varied significantly by time period; least effective in the early phase from 2006 to 2008 (HR, 0.68; 95% CI, .34-1.36) but far more protective from 2009 onward (HR, 0.33; 95% CI, .20-.55).\n\nCONCLUSIONS: ART can provide HIV-infected persons in resource-limited setting substantial protection against sexual transmission. Effectiveness in the Henan cohort appears to have increased over time, suggesting that quality of care and service infrastructure may be integral to successful use of treatment for prevention.

BACKGROUND In 2010, Nicaragua implemented an adult immunization program with the 23-valent pneumococcal polysaccharide vaccine (PPSV-23) and a pediatric immunization program with the 13-valent pneumococcal conjugate vaccine (PCV-13). We assessed incidence rates of ambulatory visits and hospitalizations for pneumonia and pneumonia-related mortality in adults over the age of 50 years before and after the program's implementation in the Department of León, Nicaragua. METHODS We collected visit diagnoses from all 107 public health facilities between 2008 and 2012 in León. We compared incidence rates of ambulatory visits for pneumonia, pneumonia hospitalizations, and pneumonia-related mortality in the pre-vaccine (2008–2009) and vaccine (2011–2012) periods among older adults using Poisson regression with generalized estimating equations (GEE), controlling for age group, municipality, and proportions of adults who were immunized against influenza. Exposure time was estimated by official municipality population estimates. RESULTS We did not observe lower incidence rates of ambulatory visits or hospitalizations for pneumonia among adults during the vaccine period versus the pre-vaccine period. However, pneumonia-related mortality was lower in the vaccine period versus the pre-vaccine period, with an adjusted incidence rate ratio (IRRa) of 0.73 (0.56, 0.94) among adults aged 50–64 years, and 0.55 (0.43, 0.70) among adults aged ≥65 years. CONCLUSIONS These early results following introduction of a combined pediatric and adult pneumococcal immunization program in Nicaragua show a probable impact of the program on the reduction of pneumonia-related deaths in older adults, but a less clear impact on the reduction of health facility visits for pneumonia.

In observational studies, treatments are typically not randomized and therefore estimated treatment effects may be subject to confounding bias. The instrumental variable (IV) design plays the role of a quasi-experimental handle since the IV is associated with the treatment and only affects the outcome through the treatment. In this paper, we present a novel framework for identification and inference using an IV for the marginal average treatment effect amongst the treated (ETT) in the presence of unmeasured confounding. For inference, we propose three different semiparametric approaches: (i) inverse probability weighting (IPW), (ii) outcome regression (OR), and (iii) doubly robust (DR) estimation, which is consistent if either (i) or (ii) is consistent, but not necessarily both. A closed-form locally semiparametric efficient estimator is obtained in the simple case of binary IV and outcome and the efficiency bound is derived for the more general case.

Copyright © 2015 Wolters Kluwer Health, Inc. The Public Health Workforce Interests and Needs Survey (PH WINS) has yielded the first-ever nationally representative sample of state health agency central office employees. The survey represents a step forward in rigorous, systematic data collection to inform the public health workforce development agenda in the United States. PH WINS is a Web-based survey and was developed with guidance from a panel of public health workforce experts including practitioners and researchers. It draws heavily from existing and validated items and focuses on 4 main areas: workforce perceptions about training needs, workplace environment and job satisfaction, perceptions about national trends, and demographics. This article outlines the conceptualization, development, and implementation of PH WINS, as well as considerations and limitations. It also describes the creation of 2 new data sets that will be available in public use for public health officials and researchers\-A nationally representative data set for permanently employed state health agency central office employees comprising over 10 000 responses, and a pilot data set with approximately 12 000 local and regional health department staff responses.

BACKGROUND: Several individual-level stressors have been linked to incident coronary heart disease (CHD), but less attention has focused on the influence of neighbourhood-level sources of stress. In this study we examined prospective associations of individual-level and neighbourhood-level stressors with incident CHD. METHODS: Multi-Ethnic Study of Atherosclerosis participants aged 45-84 years at baseline (2000-2002) with complete data were included in the analyses (n=6678 for individual-level and n=6105 for neighbourhood-level stressors). CHD was defined as non-fatal myocardial infarction, resuscitated cardiac arrest or CHD death. Median follow-up was 10.2 years. Multivariable Cox proportional hazards models were fitted to estimate associations of individual-level and neighbourhood-level stressors (categorised into approximate tertiles) with incident CHD. RESULTS: Higher reported individual-level stressors were associated with higher incident CHD. Participants in the high individual-level stressor category had 65% higher risk of incident CHD (95% CI 1.23 to 2.22) than those in the low category after adjusting for sociodemographics (P for trend=0.002). This association weakened but remained significant with further adjustment for behavioural and biological risk factors. There was a non-linear relationship between neighbourhood-level stressors and incident CHD (P for quadratic term=0.01). Participants in the medium category had 49% higher CHD risk (95% CI 1.06 to 2.10) compared with those in the low category; those in the high category had only 27% higher CHD risk (95% CI 0.83 to 1.95). These associations persisted with adjustment for risk factors and individual-level stressors. CONCLUSIONS: Individual-level and neighbourhood-level stressors were independently associated with incident CHD, though the nature of the relationships differed.

BACKGROUND: Few longitudinal studies compare changes in instrumental activities of daily living (IADLs) among stroke-free adults to prospectively document IADL changes among adults who experience stroke. We contrast annual declines in IADL independence for older individuals who remain stroke free to those for individuals who experienced stroke. We also assess whether these patterns differ by sex, race, or Southern birthplace. METHODS: Health and Retirement Study participants who were stroke free in 1998 (n = 17,741) were followed through 2010 (average follow-up = 8.9 years) for self- or proxy-reported stroke. We used logistic regressions to compare annual changes in odds of self-reported independence in six IADLs among those who remained stroke free throughout follow-up (n = 15,888), those who survived a stroke (n = 1,412), and those who had a stroke and did not survive to participate in another interview (n = 442). We present models adjusted for demographic and socioeconomic covariates and also stratified on sex, race, and Southern birthplace. RESULTS: Compared with similar cohort members who remained stroke free, participants who developed stroke had faster declines in IADL independence and lower probability of IADL independence prior to stroke. After stroke, independence declined at an annual rate similar to those who did not have stroke. The black-white disparity in IADL independence narrowed poststroke. CONCLUSION: Racial differences in IADL independence are apparent long before stroke onset. Poststroke differences in IADL independence largely reflect prestroke disparities.

BACKGROUND Nicaragua was the first developing nation to implement routine immunization with the pentavalent rotavirus vaccine (RV5). In this RV5-immunized population, understanding infectious etiologies of childhood diarrhea is necessary to direct diarrhea treatment and prevention efforts. METHODS We followed a population-based sample of children <5 years in León, Nicaragua for diarrhea episodes through household visits. Information was obtained on RV5 history and sociodemographics. Stool samples collected during diarrhea episodes and among healthy children underwent laboratory analysis for viral, bacterial and parasitic enteropathogens. Detection frequency and incidence of each enteropathogen was calculated. RESULTS The 826 children in the cohort experienced 677 diarrhea episodes during 607.5 child-years of exposure time (1.1 episodes per child-year). At least 1 enteropathogen was detected among 61.1% of the 337 diarrheal stools collected. The most common enteropathogens among diarrheal stools were: norovirus (20.4%), sapovirus (16.6%), enteropathogenic Escherichia coli (11.3%), Entamoeba histolytica/dispar (8.3%), Giardia lamblia (8.0%) and enterotoxigenic E. coli (7.7%), with rotavirus detected among 5.3% of diarrheal stools. Enteropathogenic Escherichia coli and enterotoxigenic E. coli were frequently detected among stools from healthy children. Among children with diarrhea, norovirus was more commonly detected among younger children (< 2 years) and G. lamblia was more commonly detected among older children (2-4 years). The mean age of rotavirus detection was 34.6 months. CONCLUSIONS In this Central American community after RV5 introduction, rotavirus was not commonly detected among children with diarrhea. Prevention and appropriate management of norovirus and sapovirus should be considered to further reduce the burden of diarrheal disease.

Recently, there has been increasing interest in making causal inference when interference is possible. In the presence of interference, treatment may have several types of effects. In this article, we consider inference about such effects when the population consists of groups of individuals where interference is possible within groups but not between groups. A two-stage randomization design is assumed where in the first stage groups are randomized to different treatment allocation strategies and in the second stage individuals are randomized to treatment or control conditional on the strategy assigned to their group in the first stage. For this design, the asymptotic distributions of estimators of the causal effects are derived when either the number of individuals per group or the number of groups grows large. Under certain homogeneity assumptions, the asymptotic distributions provide justification for Wald-type confidence intervals (CIs) and tests. Empirical results demonstrate that the Wald CIs have g...

Blood pressure (BP) is a complex trait, with a heritability of 30 to 40%. Several genome wide associated BP loci explain only a small fraction of the phenotypic variation. Family studies can provide an important tool for gene discovery by utilizing trait and genetic transmission information among relative-pairs. We have previously described a quantitative trait locus at chromosome 17q25.3 influencing systolic BP in American Indians of the Strong Heart Family Study (SHFS). This locus has been reported to associate with variation in BP traits in family studies of Europeans, African Americans and Hispanics. To follow-up persuasive linkage findings at this locus, we performed comprehensive genotyping in the 1-LOD unit support interval region surrounding this QTL using a multi-step strategy. We first genotyped 1,334 single nucleotide polymorphisms (SNPs) in 928 individuals from families that showed evidence of linkage for BP. We then genotyped a second panel of 306 SNPs in all SHFS participants (N = 3,807) for genes that displayed the strongest evidence of association in the region, and, in a third step, included additional genotyping to better cover the genes of interest and to interrogate plausible candidate genes in the region. Three genes had multiple SNPs marginally associated with systolic BP (TBC1D16, HRNBP3 and AZI1). In BQTN analysis, used to estimate the posterior probability that any variant in each gene had an effect on the phenotype, AZI1 showed the most prominent findings (posterior probability of 0.66). Importantly, upon correction for multiple testing, none of our study findings could be distinguished from chance. Our findings demonstrate the difficulty of follow-up studies of linkage studies for complex traits, particularly in the context of low powered studies and rare variants underlying linkage peaks.

Abstract. We estimated the incidence of watery diarrhea in the community before and after introduction of the pentavalent rotavirus vaccine in León, Nicaragua. A random sample of households was selected before and after rotavirus vaccine introduction. All children &lt; 5 years of age in selected households were eligible for inclusion. Children were followed every 2 weeks for watery diarrhea episodes. The incidence rate was estimated as numbers of episodes per 100 child-years of exposure time. A mixed effects Poisson regression model was fit to compare incidence rates in the pre-vaccine and vaccine periods. The pre-vaccine cohort (N = 726) experienced 36 episodes per 100 child-years, and the vaccine cohort (N = 826) experienced 25 episodes per 100 child-years. The adjusted incidence rate ratio was 0.60 (95% confidence interval [CI] 0.40, 0.91) during the vaccine period versus the pre-vaccine period, indicating a lower incidence of watery diarrhea in the community during the vaccine period.