Proxy Pattern-Mixture Analysis for a Binary Variable Subject to Nonresponse

AbstractGiven increasing survey nonresponse, good measures of the potential impact of nonresponse on survey estimates are particularly important. Existing measures, such as the R-indicator, make the strong assumption that missingness is missing at random, meaning that it depends only on variables that are observed for respondents and nonrespondents. We consider assessment of the impact of nonresponse for a binary survey variable Y subject to nonresponse when missingness may be not at random, meaning that missingness may depend on Y itself. Our work is motivated by missing categorical income data in the 2015 Ohio Medicaid Assessment Survey (OMAS), where whether or not income is missing may be related to the income value itself, with low-income earners more reluctant to respond. We assume there is a set of covariates observed for nonrespondents and respondents, which for the item nonresponse (as in OMAS) is often a rich set of variables, but which may be potentially limited in cases of unit nonresponse. To reduce dimensionality and for simplicity we reduce these available covariates to a continuous proxy variable X, available for both respondents and nonrespondents, that has the highest correlation with Y, estimated from a probit regression analysis of respondent data. We extend the previously proposed proxy-pattern mixture (PPM) analysis for continuous outcomes to the binary outcome using a latent variable approach for modeling the joint distribution of Y and X. Our method does not assume data are missing at random but includes it as a special case, thus creating a convenient framework for sensitivity analyses. Maximum likelihood, Bayesian, and multiple imputation versions of PPM analysis are described, and robustness of these methods to model assumptions is discussed. Properties are demonstrated through simulation and with the 2015 OMAS data.

Download Full-text

The Impact of the Revised WIC Food Package on Maternal Nutrition During Pregnancy and Postpartum

American Journal of Epidemiology ◽

10.1093/aje/kwz098 ◽

2019 ◽

Vol 188 (8) ◽

pp. 1493-1502 ◽

Cited By ~ 7

Author(s):

Rita Hamad ◽

Akansha Batra ◽

Deborah Karasek ◽

Kaja Z LeWinn ◽

Nicole R Bush ◽

...

Keyword(s):

Low Income ◽

Maternal Nutrition ◽

Temporal Trends ◽

Dietary Quality ◽

Sensitivity Analyses ◽

Control Group ◽

Food Package ◽

Pregnancy And Postpartum ◽

Before And After ◽

The Impact

Abstract The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) provides nutritional support for pregnant and postpartum women and young children. The typical food package provided to recipient families was revised in October 2009 to include more whole grains, fruits, vegetables, and low-fat milk. Little is known about whether these revisions improved nutrition among women during this critical period of the life course. We conducted a quasiexperimental difference-in-differences analysis, comparing WIC recipients (“treatment” group) before and after the WIC policy change, while accounting for temporal trends among nonrecipients (“control” group). We examined nutritional outcomes among a cohort of 1,454 women recruited during pregnancy in 2006–2011 in Memphis and surrounding Shelby County, Tennessee. We found improvements in several measures of dietary quality and nutrient intake during pregnancy, although these did not persist into the postpartum period. Results were robust to numerous sensitivity analyses. At a time when federal WIC funding is threatened, this study provides some of the first evidence of the benefits of recent WIC revisions among low-income women.

Download Full-text

Pay for performance in primary care: the contribution of the Programme for Improving Access and Quality of Primary Care (PMAQ) on avoidable hospitalisations in Brazil, 2009–2018

BMJ Global Health ◽

10.1136/bmjgh-2021-005429 ◽

2021 ◽

Vol 6 (7) ◽

pp. e005429

Author(s):

Letícia Xander Russo ◽

Timothy Powell-Jackson ◽

Jorge Otavio Maia Barreto ◽

Josephine Borghi ◽

Roxanne Kovacs ◽

...

Keyword(s):

Primary Care ◽

Low Income ◽

Pay For Performance ◽

Panel Data Analysis ◽

National Level ◽

Age Groups ◽

Hospitalisation Rate ◽

Sensitivity Analyses ◽

Access And Quality ◽

The Impact

BackgroundEvidence on the effect of pay-for-performance (P4P) schemes on provider performance is mixed in low-income and middle-income countries. Brazil introduced its first national-level P4P scheme in 2011 (PMAQ-Brazilian National Programme for Improving Primary Care Access and Quality). PMAQ is likely one of the largest P4P schemes in the world. We estimate the association between PMAQ and hospitalisations for ambulatory care sensitive conditions (ACSCs) based on a panel of 5564 municipalities.MethodsWe conducted a fixed effect panel data analysis over the period of 2009–2018, controlling for coverage of primary healthcare, hospital beds per 10 000 population, education, real gross domestic product per capita and population density. The outcome is the hospitalisation rate for ACSCs among people aged 64 years and under per 10 000 population. Our exposure variable is defined as the percentage of family health teams participating in PMAQ, which captures the roll-out of PMAQ over time. We also provided several sensitivity analyses, by using alternative measures of the exposure and outcome variables, and a placebo test using transport accident hospitalisations instead of ACSCs.ResultsThe results show a negative and statistically significant association between the rollout of PMAQ and ACSC rates for all age groups. An increase in PMAQ participating of one percentage point decreased the hospitalisation rate for ACSC by 0.0356 (SE 0.0123, p=0.004) per 10 000 population (aged 0–64 years). This corresponds to a reduction of approximately 60 829 hospitalisations in 2018. The impact is stronger for children under 5 years (−0.0940, SE 0.0375, p=0.012), representing a reduction of around 11 936 hospitalisations. Our placebo test shows that the association of PMAQ on the hospitalisation rate for transport accidents is not statistically significant, as expected.ConclusionWe find that PMAQ was associated with a modest reduction in hospitalisation for ACSCs.

Download Full-text

A review of the use of controlled multiple imputation in randomised controlled trials with missing outcome data

BMC Medical Research Methodology ◽

10.1186/s12874-021-01261-6 ◽

2021 ◽

Vol 21 (1) ◽

Author(s):

Ping-Tee Tan ◽

Suzie Cro ◽

Eleanor Van Vogt ◽

Matyas Szigeti ◽

Victoria R. Cornelius

Keyword(s):

Sensitivity Analysis ◽

Missing Data ◽

Multiple Imputation ◽

Randomised Controlled Trials ◽

Missing At Random ◽

Sensitivity Analyses ◽

Controlled Trials ◽

Primary Analysis ◽

Randomised Controlled ◽

The Impact

Abstract Background Missing data are common in randomised controlled trials (RCTs) and can bias results if not handled appropriately. A statistically valid analysis under the primary missing-data assumptions should be conducted, followed by sensitivity analysis under alternative justified assumptions to assess the robustness of results. Controlled Multiple Imputation (MI) procedures, including delta-based and reference-based approaches, have been developed for analysis under missing-not-at-random assumptions. However, it is unclear how often these methods are used, how they are reported, and what their impact is on trial results. This review evaluates the current use and reporting of MI and controlled MI in RCTs. Methods A targeted review of phase II-IV RCTs (non-cluster randomised) published in two leading general medical journals (The Lancet and New England Journal of Medicine) between January 2014 and December 2019 using MI. Data was extracted on imputation methods, analysis status, and reporting of results. Results of primary and sensitivity analyses for trials using controlled MI analyses were compared. Results A total of 118 RCTs (9% of published RCTs) used some form of MI. MI under missing-at-random was used in 110 trials; this was for primary analysis in 43/118 (36%), and in sensitivity analysis for 70/118 (59%) (3 used in both). Sixteen studies performed controlled MI (1.3% of published RCTs), either with a delta-based (n = 9) or reference-based approach (n = 7). Controlled MI was mostly used in sensitivity analysis (n = 14/16). Two trials used controlled MI for primary analysis, including one reporting no sensitivity analysis whilst the other reported similar results without imputation. Of the 14 trials using controlled MI in sensitivity analysis, 12 yielded comparable results to the primary analysis whereas 2 demonstrated contradicting results. Only 5/110 (5%) trials using missing-at-random MI and 5/16 (31%) trials using controlled MI reported complete details on MI methods. Conclusions Controlled MI enabled the impact of accessible contextually relevant missing data assumptions to be examined on trial results. The use of controlled MI is increasing but is still infrequent and poorly reported where used. There is a need for improved reporting on the implementation of MI analyses and choice of controlled MI parameters.

Download Full-text

Social Integration of Second Generation Students in the Italian School System

Social Indicators Research ◽

10.1007/s11205-021-02801-9 ◽

2021 ◽

Author(s):

Francesco Giovinazzi ◽

Daniela Cocchi

Keyword(s):

Social Integration ◽

School System ◽

Latent Variable ◽

Missing Values ◽

Second Generation ◽

Variable Model ◽

Qualitative Variable ◽

Italian School ◽

Variable Approach ◽

The Impact

AbstractCultural divides and prejudice complicate the processes of integration and acculturation of migrant families living in a foreign country. Evaluating the impact of such phenomenon can be crucial for social stability and policy making. In this context, the education system has a leading role in fostering and attaining social integration, in particular when it comes to younger sections of the migrant population. In this work, we propose a method for the construction of a quantitative indicator capturing social integration of second generation students in the Italian school system according to areas defined by nationality of the students and administrative region in which they attend school. The indicator, based on survey data, is estimated by means of a 2-step methodology. In the first step, we choose an individual qualitative variable capturing social integration at the unit level, and we compute a first direct estimate of the indicator as the proportion of highly integrated students in each area. Such variable is isolated following alternatively a proxy variable approach or a latent variable model approach. In the second step, we make use of two alternative small area models to improve the estimates, dealing with missing values, low sample size and high variability in smaller domains. At the end, the 2-step methodology results in 4 alternative versions of a synthetic indicator of social integration, that can be used to rank nationalities and administrative regions.

Download Full-text

Corporate Default Risk and Environmental Deterioration (International Evidence)

10.21203/rs.3.rs-541507/v1 ◽

2021 ◽

Author(s):

Obaid Ur Rehman ◽

Xiaoxing Liu

Keyword(s):

Low Income ◽

Default Risk ◽

Negative Impact ◽

Market Competition ◽

Preliminary Evidence ◽

Firm Level ◽

Corporate Default ◽

Environmental Deterioration ◽

Variable Approach ◽

The Impact

Abstract This study explores the impact of corporate default risk on environmental deterioration in the international context. We find that corporate bankruptcy is positively associated with CO2 emissions and its decomposed components. These findings are reliable in low-income and highly uncertain countries but weak in countries having more market competition. We also find that the negative impact of corporate default risk on environment is more robust in countries with more population density and less forest area thresholds. Using instrumental variable approach, we provide preliminary evidence that firm-level political risk (for U.S. and Canadian firms only) tend to increase corporate default risk leading to degrading environment. Our research will help environmental authorities to consider corporate-default risk as a determinant when formulating environmental-related strategies.

Download Full-text

Impact of expanding smoke-free policies beyond enclosed public places and workplaces on children’s tobacco smoke exposure and respiratory health: protocol for a systematic review and meta-analysis

BMJ Open ◽

10.1136/bmjopen-2020-038234 ◽

2020 ◽

Vol 10 (10) ◽

pp. e038234

Author(s):

Márta K Radó ◽

Famke JM Mölenberg ◽

Aziz Sheikh ◽

Christopher Millett ◽

Wichor M Bramer ◽

...

Keyword(s):

Tobacco Smoke ◽

Low Income ◽

Respiratory Health ◽

Smoke Exposure ◽

Risk Of Bias ◽

Tobacco Smoke Exposure ◽

Sensitivity Analyses ◽

Publication Date ◽

Public Places ◽

The Impact

IntroductionTobacco smoke exposure (TSE) has considerable adverse respiratory health impact among children. Smoke-free policies covering enclosed public places are known to reduce child TSE and benefit child health. An increasing number of jurisdictions are now expanding smoke-free policies to also cover outdoor areas and/or (semi)private spaces (indoor and/or outdoor). We aim to systematically review the evidence on the impact of these ‘novel smoke-free policies’ on children’s TSE and respiratory health.Methods and analysis13 electronic databases will be searched by two independent reviewers for eligible studies. We will consult experts from the field and hand-search references and citations to identify additional published and unpublished studies. Study designs recommended by the Cochrane Effective Practice and Organisation of Care (EPOC) group are eligible, without restrictions on the observational period, publication date or language. Our primary outcomes are: self-reported or parental-reported TSE in places covered by the policy; unplanned hospital attendance for wheezing/asthma and unplanned hospital attendance for respiratory infections. We will assess risk of bias of individual studies following the EPOC or Risk Of Bias In Non-randomised Studies of Interventions tool, as appropriate. We will conduct separate random effects meta-analyses for smoke-free policies covering (1) indoor private places, (2) indoor semiprivate places, (3) outdoor (semi)private places and (4) outdoor public places. We will assess whether the policies were associated with changes in TSE in other locations (eg, displacement). Subgroup analyses will be conducted based on country income classification (ie, high, middle or low income) and by socioeconomic status. Sensitivity analyses will be undertaken via broadening our study design eligibility criteria (ie, including non-EPOC designs) or via excluding studies with a high risk of bias. This review will inform policymakers regarding the implementation of extended smoke-free policies to safeguard children’s health.Ethics and disseminationEthical approval is not required. Findings will be disseminated to academics and the general public.PROSPERO registration numberCRD42020190563.

Download Full-text

Antisocial Behavior is Associated with Reduced Frontoparietal Activity to Loss in a Population-Based Sample of Adolescents

10.31234/osf.io/ctpds ◽

2021 ◽

Author(s):

Laura Murray ◽

Nestor Lopez-Duran ◽

Colter Mitchell ◽

Christopher S. Monk ◽

Luke Williamson Hyde

Keyword(s):

Antisocial Behavior ◽

Low Income ◽

Latent Variable ◽

Community Sample ◽

Population Based ◽

Neural Response ◽

Health Concern ◽

Negative Consequences ◽

Cu Traits ◽

Variable Approach

Background: Adolescent antisocial behavior (AB) is a public health concern due to the high financial and social costs of AB on victims and perpetrators. Neural systems involved in reward and loss processing are thought to contribute to AB. However, investigations into these processes are limited: few have considered anticipatory and consummatory components of reward, response to loss, nor whether associations with AB may vary by level of callous-unemotional (CU) traits.Methods: A population-based, community sample of 128 predominantly low-income youth (mean age =15.9 years; 42% male) completed a monetary incentive delay task during fMRI. A multi-informant, multi-method latent variable approach was used to test associations between AB and neural response to reward and loss anticipation and outcome, and whether these associations were moderated by CU traits.Results: AB was not associated with neural response to reward but was associated with reduced frontoparietal activity during loss outcomes. This association was moderated by CU traits such that individuals with higher levels of AB and CU traits had the largest reductions in frontoparietal activity. Co-occurring AB and CU traits were also associated with increased precuneus response during loss anticipation.Conclusions: Findings indicate that AB is associated with reduced activity in brain regions involved in cognitive control, attention, and behavior modification during negative outcomes. Moreover, these reductions are most pronounced in youth with co-occurring CU traits. These findings have implications for understanding why adolescents involved in AB continue these behaviors despite severe negative consequences (e.g., incarceration).

Download Full-text

Assessing the impact of extreme values in clinical studies−a latent variable approach

Biometrics & Biostatistics International Journal ◽

10.15406/bbij.2019.08.00276 ◽

2019 ◽

Vol 8 (3) ◽

pp. 85-91

Author(s):

Zhengning Lin

Keyword(s):

Clinical Studies ◽

Latent Variable ◽

Extreme Values ◽

Variable Approach ◽

The Impact ◽

Latent Variable Approach

Download Full-text

Demonstrating Nutrient Cost Gradients: A Brooklyn Case Study

International Journal for Vitamin and Nutrition Research ◽

10.1024/0300-9831/a000210 ◽

2014 ◽

Vol 84 (5-6) ◽

pp. 244-251 ◽

Cited By ~ 2

Author(s):

Robert J. Karp ◽

Gary Wong ◽

Marguerite Orsi

Keyword(s):

Low Income ◽

Energy Content ◽

Food Selection ◽

The United States ◽

United States Department ◽

Micronutrient Deficiencies ◽

Caloric Density ◽

Low Income Families ◽

The Cost ◽

The Impact

Abstract. Introduction: Foods dense in micronutrients are generally more expensive than those with higher energy content. These cost-differentials may put low-income families at risk of diminished micronutrient intake. Objectives: We sought to determine differences in the cost for iron, folate, and choline in foods available for purchase in a low-income community when assessed for energy content and serving size. Methods: Sixty-nine foods listed in the menu plans provided by the United States Department of Agriculture (USDA) for low-income families were considered, in 10 domains. The cost and micronutrient content for-energy and per-serving of these foods were determined for the three micronutrients. Exact Kruskal-Wallis tests were used for comparisons of energy costs; Spearman rho tests for comparisons of micronutrient content. Ninety families were interviewed in a pediatric clinic to assess the impact of food cost on food selection. Results: Significant differences between domains were shown for energy density with both cost-for-energy (p < 0.001) and cost-per-serving (p < 0.05) comparisons. All three micronutrient contents were significantly correlated with cost-for-energy (p < 0.01). Both iron and choline contents were significantly correlated with cost-per-serving (p < 0.05). Of the 90 families, 38 (42 %) worried about food costs; 40 (44 %) had chosen foods of high caloric density in response to that fear, and 29 of 40 families experiencing both worry and making such food selection. Conclusion: Adjustments to USDA meal plans using cost-for-energy analysis showed differentials for both energy and micronutrients. These differentials were reduced using cost-per-serving analysis, but were not eliminated. A substantial proportion of low-income families are vulnerable to micronutrient deficiencies.

Download Full-text