Analysis Of Complex Survey Data Research Articles

Factors associated with health-seeking patterns among internally displaced persons in complex humanitarian emergency, Northeast Nigeria: a cross-sectional study

BackgroundCurrently, over two million persons are internally displaced because of the complex humanitarian emergency in Nigeria’s northeast region. Due to crowded and unsanitary living conditions, the risk of communicable disease transmission, morbidity, and mortality among this population is high. This study explored patterns and factors associated with health-seeking among internally displaced persons (IDPs) in northeast Nigeria to inform and strengthen disease surveillance and response activities.MethodsIn a cross-sectional study conducted during June–October 2022, we employed stratified sampling technique to select 2,373 IDPs from 12 IDPs camps. A semi-structured tool was used to collect data on health-seeking patterns, socio-demographics, households, and IDPs camps characteristics. We classified health-seeking patterns into three outcome categories: ‘facility care’ (reference category), ‘non-facility care’ (patent medicine vendors, chemists, traditional healers, religious centers), and ‘home care/no care’. We performed complex survey data analysis and obtained weighted statistical estimates. Univariate analysis was conducted to describe respondents’ characteristics and health-seeking patterns. We fitted weighted multivariable multinomial logistic regression models to identify factors associated with health-seeking patterns.ResultsOf 2,373 respondents, 71.8% were 18 to 39 years old, 78.1% were females, and 81.0% had no formal education. Among the respondents, 75.7% (95% CI: 72.9–78.6) sought ‘facility care’, 11.1% (95% CI: 9.1–13.1) sought ‘non-facility care’, while 13.2% (95% CI: 10.9–15.4) practiced ‘home care/no care’. Respondents who perceived illness was severe (Adjusted Odds Ratio (AOR) = 0.15, [95% CI: 0.08–0.30]) and resided in officially-recognized camps (AOR = 0.26, [95% CI: 0.17–0.39]) were less likely to seek ‘non-facility care’ compared to ‘facility care’. Similarly, respondents who resided in officially-recognized camps (AOR = 0.58, [95% CI: 0.36–0.92]), and received disease surveillance information (AOR = 0.42, [95% CI: 0.26–0.67) were less likely to practice ‘home care/no care’ rather than seek ‘facility care’.ConclusionsThis population exhibited heterogeneous patterns of health-seeking at facility and non-facility centers. Perception of illness severity and camps’ status were major factors associated with health-seeking. To enhance surveillance, non-facility care providers should be systematically integrated into the surveillance network while ramping up risk communication to shape perception of illness severity, prioritizing unofficial camps.

Implementing machine learning methods with complex survey data: Lessons learned on the impacts of accounting sampling weights in gradient boosting.

Despite the prominent use of complex survey data and the growing popularity of machine learning methods in epidemiologic research, few machine learning software implementations offer options for handling complex samples. A major challenge impeding the broader incorporation of machine learning into epidemiologic research is incomplete guidance for analyzing complex survey data, including the importance of sampling weights for valid prediction in target populations. Using data from 15, 820 participants in the 1988-1994 National Health and Nutrition Examination Survey cohort, we determined whether ignoring weights in gradient boosting models of all-cause mortality affected prediction, as measured by the F1 score and corresponding 95% confidence intervals. In simulations, we additionally assessed the impact of sample size, weight variability, predictor strength, and model dimensionality. In the National Health and Nutrition Examination Survey data, unweighted model performance was inflated compared to the weighted model (F1 score 81.9% [95% confidence interval: 81.2%, 82.7%] vs 77.4% [95% confidence interval: 76.1%, 78.6%]). However, the error was mitigated if the F1 score was subsequently recalculated with observed outcomes from the weighted dataset (F1: 77.0%; 95% confidence interval: 75.7%, 78.4%). In simulations, this finding held in the largest sample size (N = 10,000) under all analytic conditions assessed. For sample sizes <5,000, sampling weights had little impact in simulations that more closely resembled a simple random sample (low weight variability) or in models with strong predictors, but findings were inconsistent under other analytic scenarios. Failing to account for sampling weights in gradient boosting models may limit generalizability for data from complex surveys, dependent on sample size and other analytic properties. In the absence of software for configuring weighted algorithms, post-hoc re-calculations of unweighted model performance using weighted observed outcomes may more accurately reflect model prediction in target populations than ignoring weights entirely.

Socioeconomic and Racial/Ethnic Disparities in Recovery from Childhood Behavioral or Conduct Problems: Evidence from a Nationally Representative Sample of 3-17 Years Old US Children.

Our study provides nationally-generalizable evidence on the racial/ethnic and socioeconomic disparities in diagnosis and recovery from childhood behavioral or conduct problems. We pooled data from 4years (2016-2019) of the National Survey of Children's Health (NSCH) for 3 to 17years old US children (N = 114,476). We performed several logistic regression models using complex survey data analysis statistical methodologies to estimate nationally representative and generalizable results in the Stata MP 16 program. About 20.1% of 3-17years old US children previously diagnosed with behavioral or conduct problems no longer had the current diagnosis (were recovered). Hispanic children [Odds ratio (OR) 0.77, 95% Confidence Interval (CI) 0.65-0-95], immigrant children or children of immigrant parents (first or second generation immigrant children), and children from high-income families [200-399% Federal Poverty Level (FPL)] were about 23%, 38%, and 21% less likely than non-Hispanic White children, children of US native parents, and children with a family income of below 100% FPL to be currently diagnosed with behavioral or conduct problems, respectively. Conversely, Non-Hispanic Black and Hispanic children were about 50% and 40% more likely than non-Hispanic White children to recover from a past diagnosis. Moreover, children from higher-income families (at or above 300% of FPL) were between 1.59 to 1.79 times more likely than those from low-income families (below 100% FPL) to recover from a past diagnosis. Racial/ethnic and socioeconomic disparities in diagnosing appear to persist in recovering from behavioral or conduct problems.

Farm management practices, biosecurity and influenza a virus detection in swine farms: a comprehensive study in Colombia

Biosecurity protocols (BP) and good management practices are key to reduce the risk of introduction and transmission of infectious diseases into the pig farms. In this observational cross-sectional study, survey data were collected from 176 pig farms with inventories over 100 sows in Colombia. We analyzed a complex survey dataset to explore the structure and identify clustering patterns using Multiple Correspondence Analysis (MCA) of swine farms in Colombia, and estimated its association with Influenza A virus detection. Two principal dimensions contributed to 27.6% of the dataset variation. Farms with highest contribution to dimension 1 were larger farrow-to-finish farms, using self-replacement of gilts and implementing most of the measures evaluated. In contrast, farms with highest contribution to dimension 2 were medium to large farrow-to-finish farms, but implemented biosecurity in a lower degree. Additionally, two farm clusters were identified by Hierarchical Cluster Analysis (HCA), and the odds of influenza A virus detection was statistically different between clusters (OR 7.29, CI: 1.7,66, p = < 0.01). Moreover, after logistic regression analysis, three important variables were associated with higher odds of influenza detection: (1) “location in an area with a high density of pigs”, (2) “farm size”, and (3) “after cleaning and disinfecting, the facilities are allowed to dry before use”. Our results revealed two clustering patterns of swine farms. This systematic analysis of complex survey data identified relationships between biosecurity, husbandry practices and influenza status. This approach helped to identify gaps on biosecurity and key elements for designing successful strategies to prevent and control swine respiratory diseases in the swine industry.

Small Domain Estimation of Census Coverage – A Case Study in Bayesian Analysis of Complex Survey Data

Abstract Many countries conduct a full census survey to report official population statistics. As no census survey ever achieves 100% response rate, a post-enumeration survey (PES) is usually conducted and analysed to assess census coverage and produce official population estimates by geographic area and demographic attributes. Considering the usually small size of PES, direct estimation at the desired level of disaggregation is not feasible. Design-based estimation with sampling weight adjustment is a commonly used method but is difficult to implement when survey nonresponse patterns cannot be fully documented and population benchmarks are not available. We overcome these limitations with a fully model-based Bayesian approach applied to the New Zealand PES. Although theory for the Bayesian treatment of complex surveys has been described, published applications of individual level Bayesian models for complex survey data remain scarce. We provide such an application through a case study of the 2018 census and PES surveys. We implement a multilevel model that accounts for the complex design of PES. We then illustrate how mixed posterior predictive checking and cross-validation can assist with model building and model selection. Finally, we discuss potential methodological improvements to the model and potential solutions to mitigate dependence between the two surveys.

Pseudo empirical likelihood inference for nonprobability survey samples

AbstractIn this article, we first provide an overview of two major developments on complex survey data analysis: the empirical likelihood methods and statistical inference with nonprobability survey samples. We highlight the important research contributions to the field of survey sampling in general and the two topics in particular by Canadian survey statisticians. We then propose new inferential procedures for analyzing nonprobability survey samples through the pseudo empirical likelihood approach. The proposed methods lead to point estimators asymptotically equivalent to those discussed in the recent literature but with more desirable features on confidence intervals such as range‐respecting and data‐driven orientation. Results from a simulation study demonstrate the superiority of the proposed methods in dealing with binary response variables.

Joinpoint Regression Methods of Aggregate Outcomes for Complex Survey Data

Abstract Joinpoint regression can model trends in time-specific estimates from aggregated data. These methods have been developed mainly for nonsurvey data such as cancer registry data assuming that the time-specific estimates are uncorrelated from time point to time point. This independence assumption can be violated for trends in time-specific estimates from complex survey samples due to using the same primary sampling units across time and, therefore, the full variance–covariance matrix of the time-specific estimates should be incorporated into the regression model fitting. This article extends these joinpoint methods for analyzing complex survey data within the National Cancer Institute’s Joinpoint software and empirically compares the extended method to existing methods for analyses of time trends in three surveys.

A Comparison of Propensity Score and Linear Regression Analysis of Complex Survey Data

We extend propensity score methodology to incorporate survey weights from complex survey data and compare the use of multiple linear regression and propensity score analysis to estimate treatment effects in observational data from a complex survey. For illustration, we use these two methods to estimate the effect of gender on information technology (IT) salaries. In our analysis, both methods agree on the size and statistical significance of the overall gender salary gaps in the United States in four different IT occupations after controlling for educational and job-related covariates. Each method, however, has its own advantages which are discussed. We also show that it is important to incorporate the survey design in both linear regression and propensity score analysis. Ignoring the survey weights affects the estimates of population-level effects substantially in our analysis.

Exploring determinants of innovation potential of enterprises in Kazakhstan

The study aims to analyze the innovation potential of enterprises, explain the use of the binary regression methodology, and explore different indicators of the enterprises in the regions of Kazakhstan for building an effective management strategy. Methodological substantiation is based on the complex survey data analysis, provided by the World Bank Enterprises Survey (WBES) for Kazakhstan. WBES database has covered a sociological survey of enterprises, which was conducted using a random survey and representativeness among 1,296 enterprises, mainly in the production sector. Besides, the data were collected among companies regarding their experience in environmental perception (including innovation activity), in which they worked. The results have demonstrated that the age of the company, exporter status, type, sector, or activity – all these have a positive influence on the company’s tendency to innovations. However, as part of the study, it has found out that competitors in the marketplace and regions of activity of enterprises predominantly negatively affect the prospects of introducing innovation. It was also in evidence that the same factors (foreign ownership, advanced training, type, size, and sector of an enterprise) are essential determinants of product, technology, R&amp;amp;amp;D, ICT, and innovation. Obtained results can be used by enterprises for building an effective management strategy of own business innovative development, as well as by local authorities to upgrade the competitive performance. AcknowledgmentsThis study has been funded by the Science Committee of the Ministry of Education and Science of the Republic of Kazakhstan (Grant “Priorities and mechanisms of inclusive regional development of Kazakhstan in the context of overcoming the economic recession” No. AP09259004).

Multivariate Logistic Regression Analysis of Complex Survey Data with Application to BRFSS Data

Multiple binary outcomes that measure the presence or absence of medical conditions occur frequently in public health survey research. The multiple possibly correlated binary outcomes may compose of a syndrome or a group of related diseases. It is often of scientic interest to model the interrelationships not only between outcome and risk factors, but also between dierent outcomes. Applied and practical methods dealing with multiple outcomes from complex designed surveys are lacking. We propose a multivariate approach based on the generalized estimating equation (GEE) methodology to simultaneously conduct survey logistic regressions for each binary outcome in a single analysis. The approach has the following attrac- tive features: 1) It enables modeling the complete information from multiple outcomes in a single analysis; 2) it permits to test the correlations between multiple binary outcomes; 3) it allows of discerning the outcome-specic ef- fect and the overall risk factor eect; and 4) it provides the measurement of dierence of the association between risk factors and multiple outcomes. The proposed method is applied to a study on risk factors for heart attack and stroke in 2009 U.S. nationwide Behavioral Risk Factor Surveillance System (BRFSS) data.

Multivariate Logistic Regression Analysis of Complex Survey Data with Application to BRFSS Data

Multivariate Logistic Regression Analysis of Complex Survey Data with Application to BRFSS Data

Contextual influence on formal and informal teacher leadership

Using the 2013 Teaching and Learning International Survey data, this research employed descriptive analysis and multiple logistic regression with the complex survey data analysis approach (resampling with Balanced Repeated Replicate weights), in order to reveal the variations in overall informal and formal teacher leadership, and for different leadership dimensions, respectively, as well as the contextual influence on teacher leadership. The results suggest that there exist significant variations between general informal and formal teacher leadership, and for each of the leadership dimensions internationally. In sum, Eastern and Nordic Europe have a greater extent of teacher leadership, compared to Asia, Latin America, and Latin Europe. There was less prevalent teacher leadership in hiring, setting teacher salary, and budget allocation in general. Formal teacher leaders often make school decisions for disciplinary policy, teacher evaluation, data analysis, and parental communication, while informal teacher leaders frequently decide on instructional content. Female, experienced, and well-educated teachers often make decisions even without positions; schools with increased minority students tend to have a greater extent of teacher leadership, while schools with disproportional poverty students have decreased teacher leadership. Using large-scale and international data, this study has provided nuanced evidence in teacher leadership with the intersection of leadership functions, positions, and contexts.

Teenage pregnancy and experience of physical violence among women aged 15-19 years in five African countries: Analysis of complex survey data.

BackgroundPregnant teenage women are prime targets of violence against women perpetrated by intimate partners, family members, and miscreants in their neighborhoods. This study estimated the prevalence of Teenage pregnancy (TP) and Physical Violence (PV) and further assessed the relationship between TP and PV in five Low-and-Middle-Income Countries (LMICs).MethodsThe study was conducted among five LIMCs (Burkina Faso, Kenya, Malawi, Nigeria, and Tanzania) using data from the most recent Demographic and Health Surveys conducted in these countries. Modified Poisson with the robust standard error was used to quantify the association between TP and PV. All analyses adjusted for the complex survey design structure (clustering, weighting, and stratification).ResultsThe analysis involved a total of 26055 adolescent women aged 15–19 years across the five countries. The overall prevalence of TP was 25.4% (95%CI = 24.4–26.4) with the highest prevalence occurring among Malawians [29.0% (95%CI = 27.4–30.7)]. Meanwhile, the prevalence of TP among older adolescents (18–19 years) was approximately two-thirds significantly higher compared with young adolescents [aPR(95%CI) = 1.60[1.49–1.71)]. The prevalence of PV among teenagers across the five countries was 24.2% (95%CI = 22.3–26.2). The highest prevalence of PV was recorded among Nigerian adolescent women [31.8% (95%CI = 28.5–35.3)]. The prevalence of PV among adolescent women who were pregnant was approximately 5-folds significant compared to those who were not pregnant (adjusted prevalence ratio; aPR = 4.70; 95% CI: 3.86–5.73; p<0.0001).ConclusionThere was a high prevalence of pregnancy among older teenagers aged 18–19 years. Close to a quarter of teenage women ever experienced physical violence. Pregnant teenage women ever experience of physical violence was very high compared to non-pregnant peers. Intervention should target PV and TP by adopting a gender-sensitive approach to eliminate physical violence, particularly among teenagers to prevent TP.

Age standardization using reweighting in complex survey data and its application in regression

Age standardization is often used in analyzing complex survey data to compare subgroups with different age distributions. Conventionally, age-standardized estimates are derived as weighted averages of age-specific estimates applying weights based on a standard population of choice. Age adjustment, to be distinguished from age standardization, is compatible with widely-used, regression-based techniques for assessing time trends. Those techniques are much harder to use in conjunction with the conventional age standardization method. A reweighting-based alternative method, which can readily accommodate regression-based time trend assessment, is described in this presentation. The alternative involves modification of sampling weights that force age composition for each subgroup to match that of the standard population. To illustrate an application of the alternative, time trend assessments of erythrocyte metal measurements by race/Hispanic origin subgroup in the National Health and Examination Nutrition Survey (NHANES) 1999-2016 are used. The estimand of interest is age-standardized geometric mean derived from Tobit regression that treats measurements below the lower limit of detection as left-censored. When ordinary least square-type regression was used in place of Tobit regression, the reweighting-based method produced age-standardized estimates that exactly matched those obtained using the conventional method, as expected mathematically. When Tobit regression was used, the conventional and alternative methods produced slightly different age-standardized geometric mean estimates for erythrocyte concentration of lead, cadmium, and mercury by sex, race/Hispanic origin, and survey cycle with the difference within ± 5% in 193 out of 198 metal-subgroup combinations. The reason for the difference was investigated to inform potential method improvement for reducing the difference. Annual percentage changes and p-values from orthogonal contrast-based tests for linear, quadratic, and other higher order trends can be produced easily in conjunction with the reweighting-based age standardization.

Estimating a Multilevel Model with Complex Survey Data: Demonstration using TIMSS

Analysis of complex survey data is demonstrated for the multilevel model. Description of specific aspects of analysis, including plausible values, sampling weights, and replicate weights is provided. Following this, example TIMSS data and models are described and results are presented.

Secondhand Smoke Exposure among Nonsmokers in China.

China signed the World Health Organization Framework Convention on Tobacco Control and conducted a series of activities to protect people from secondhand smoke exposure. This paper explores the changes in prevalence of secondhand smoke exposure among nonsmokers between 2010 and 2015. Data from the 2010 Global Adult Tobacco Survey and 2015 National Adult Tobacco Survey were used in this study. Due to the complex sample design for these surveys, data were weighted and analyzed using the SAS 9.3 complex survey data analysis procedure. The Chi-square test was used for comparison among different groups. From 2010 to 2015, secondhand smoke exposure among nonsmokers decreased in restaurants, government buildings, health-care facilities, schools, and public transportation in China (p<0.05). The relative change was most significant for schools (52.1%), followed by public transportation (49.4%) and government buildings (42.2%). The percentage of secondhand smoke exposure reported in workplaces declined from 55.2% to 45.3%. Secondhand smoke exposure at home reduced from 58.3% to 46.7%. People's awareness that secondhand smoke could cause heart disease in adults, lung illness in children, and lung cancer in adults increased from 24.6% to 36.0%. Additionally, support for smoke-free policies is high among the Chinese population. Even for restaurants, where support for smoke-free policies was lowest, 75.1% of nonsmokers and 55.3% of smokers supported smoke-free policies. Secondhand smoke exposure declined from 2010 to 2015 in China but remains a serious problem. Public awareness about the hazards of secondhand smoke is increasing and Chinese people support smoke-free laws.

Using SPSS to Analyze Complex Survey Data: A Primer

An introduction to using SPSS to analyze complex survey data is given. Key features of complex survey design are described briefly, including stratification, clustering, multiple stages, and weights. Then, annotated SPSS syntax for complex survey data analysis is presented to demonstrate the step-by-step process using real complex samples data.

Food insecurity and social determinants of health among immigrants and natives in Portugal

Literature is scarce on food insecurity in the context of different illnesses and intersecting social hierarchies of gender and ethnicity. This study aims to describe and compare the prevalence of food insecurity between immigrants and natives in Portugal and explore social determinants of poor health outcomes associated with food insecurity. Data were derived from the National Food, Nutrition and Physical Activity Survey 2015–2016, which is a national and regionally representative survey of the Portuguese general population which collected nationwide data on dietary habits (including food insecurity measured by the Radimer/Cornell food security questionnaire) and physical activity, among other dimensions. Demographic, socioeconomic and health factors that could help explain ethnic disparities in food insecurity were explored through complex survey data analysis. The prevalence of food insecurity was 10.7% (95% CI: 6.5–15.0) among immigrants and 10.1% (95% CI: 8.3–11.9) among natives; no significant differences were found. Low family income and low level of education were the main factors associated with food insecurity in Portugal, in both native and immigrant groups. Self-reported diseases and poor self-rated health were also associated with food insecurity. Only among natives, women, older and unmarried subjects had higher food insecurity. Therefore, inter-sectoral policies addressing the social determinants of food insecurity are needed to reduce social inequalities and particular attention should be given for Portuguese women, elderly and unmarried people which are the most vulnerable groups. Promoting equality in household food and nutrition security in Portugal including among immigrant’s populations is a public health priority.

The Pseudo Maximum Likelihood Estimator for Quantiles of Survey Variables

Abstract The pseudo maximum likelihood approach has been employed in multilevel models for analysis of complex survey data. This approach may be inappropriate for many survey variables that are nonsymmetrically distributed with skewness and multimodality characteristics. This article intends to fill this gap in the literature by developing a pseudolikelihood estimator for quantiles of survey variables in the quantile regression framework. This approach is illustrated using a Monte Carlo simulation study and the body mass index data from the 1998–1999 Early Childhood Longitudinal Study. Results show that the proposed estimator is consistent and approximately unbiased for both informative and noninformative sampling designs.

Methods for analysis of complex survey data: an application using the Tanzanian 2015 Demographic and Health Survey and Service Provision Assessment.

BackgroundLow-income and middle-income countries (LMICs) seek to better utilize household and health facility survey data for monitoring and evaluation, as well as for health program planning. However, analysis of this complex survey data are complicated. In Tanzania, the National Evaluation Platform project sought to analyze Demographic and Health Survey (DHS) data and Service Provision Assessment (SPA) data as part of an evaluation of the national One Plan for Maternal and Child Health. To support this evaluation, we used this survey data to answer two key methodological questions: 1) what are the benefits and costs of using sampling weights in rate estimation; and 2) what is the best method for calculating standard errors in these two surveys?MethodsWe conducted a simulation study for each methodologic question. The first simulation study assessed the benefits and costs of using sampling weights in rate estimation. This simulation used weighted and unweighted estimates and examined bias, variance, and the mean squared error (MSE). The second simulation study assessed the best method for calculating standard errors comparing cluster bootstrapped variance estimation, design based asymptotic variance with one level (svy1), and design based asymptotic variance with three levels (svy3). We compared coverage probability and confidence interval length.ResultsOur results showed that although weighted estimates were less biased, unweighted estimates were less variable. The weighted estimates had a lower MSE, indicating that the effect of the bias trade-off was greater than the effect of the variance trade-off for most indicators assessed. The best performer for variance estimation was the cluster bootstrap method, followed by the svy3 method. The svy1 method was the worst performer for most indicators assessed.ConclusionsAs complex survey data become more widely used for policymaking in LMICs, there is a need for guidance on the best methods for analyzing this data. The standard of practice has been a design-based analysis using survey weights and the single-level svy method for calculating standard errors. This study puts forth an alternative approach to analysis. In addition, this study offers practical guidance on determining the best method for analysis of complex survey data.