Recai Yucel

Financial Burden for Families of Children with Special Health Care Needs

Objectives: We describe family finance-related burden experienced by families with children with special health care needs (CSHCN). The paper further seeks to describe correlates of family financial burden. Methods: We examined correlates of family finance-related burden using multivariate methods and the National Survey of CSHCN, a nationally representative cross-sectional survey of CSHCN. We also examined state-level correlations. Results: Fully 40% of families with CSHCN, or 3,746,000 families nation-wide, experience financial burden related to their child’s condition. Experiencing a finance-related problem is negatively associated with Maternal and Child Health Bureau (MCHB) indicators and positively associated with poor-child health status. States that better meet MCHB indicators generally have lower levels of family finance-related problems. Conclusions: Families with CSHCN have high levels of finance-related family problems. Development of appropriate systems of care appears to offer a mechanism for alleviating the financial burdens of these families.

Plausibility of multivariate normality assumption when multiply imputing non-Gaussian continuous outcomes: a simulation assessment

Multiple imputation under the assumption of multivariate normality has emerged as a frequently used model-based approach in dealing with incomplete continuous data in recent years. Despite its simplicity and popularity, however, its plausibility has not been thoroughly evaluated via simulation. In this work, the performance of multiple imputation under a multivariate Gaussian model with unstructured covariances was examined on a broad range of simulated incomplete data sets that exhibit varying distributional characteristics such as skewness and multimodality that are not accommodated by a Gaussian model. Behavior of efficiency and accuracy measures was explored to determine the extent to which the procedure works properly. The conclusion drawn is that although the real data rarely conform with multivariate normality, imputation under the assumption of normality is a fairly reasonable tool, even when the assumption of normality is clearly violated; the fraction of missing information is high, especially when the sample size is relatively large. Although we discourage its uncritical, automatic and, possibly, inappropriate use, we report that its performance is better than we expected, leading us to believe that it is probably an underrated approach.

Access to Care and Use of Preventive Services by Hispanics: State-based Variations From 1991 to 2004

Background:State-level disparities in access to physicians and preventive services between Hispanics and whites may have changed over time. Objective:To assess state-based changes in Hispanics’ access to physicians and preventive services from 1991 to 2004. Methods:Using data from the Behavioral Risk Factor Surveillance System in the 10 states with the largest Hispanic populations, we examined 4 preventive services for eligible adults (mammography, Papanicolaou testing, colorectal cancer screening, and cholesterol testing) and 2 measures of access to physicians (obtaining routine checkup in prior 2 years and avoiding seeing physician when needed due to cost in prior year). In each state we assessed unadjusted and adjusted Hispanic-white access gaps and changes over time. Results:Hispanic-white access gaps persisted over time and varied widely by state. Disparities narrowed and became nonsignificant in 2 states (Arizona and California) for mammography and 3 states (Nevada, New Mexico, and New York) for Pap testing. Other disparities increased and became significant (mammography in Texas; colorectal cancer screening in California, Colorado, and Texas; cholesterol testing in Florida and Nevada; routine checkups in Arizona and New Mexico). Disparities in lacking doctor visits due to cost remained large and significant over time in all states. Insurance status and education were the main contributors to Hispanic-white disparities and their impact increased over time. Conclusions:Although use of preventive services and access to physicians improved for both whites and Hispanics nationally, access gaps varied widely among states. Therefore, efforts to monitor and eliminate disparities should be conducted at both the national and state levels.

Characteristics of Medical School Faculty Members Serving on Institutional Review Boards: Results of a National Survey.

Purpose To understand the characteristics of medical school faculty members who serve on institutional review boards (IRBs) in U.S. academic health centers. Method Between October 2001 and March 2002, a questionnaire was mailed to a stratified random sample of 4,694 faculty members in 121 four-year medical schools in the United States (excluding Puerto Rico). The sample was drawn from the Association of American Medical Colleges faculty roster database for 1999. The primary independent variable was service on an IRB. Data were analyzed using standard statistical procedures. Results A total of 2,989 faculty members responded (66.5%). Eleven percent of respondents reported they had served on an IRB in the three years before the study. Of these, 73% were male, 81% were white (non-Hispanic). Virtually all faculty IRB members (94%) conducted some research in the three years before the study, and, among these, 71% reported conducting clinical research, and 47% served as industrial consultants to industry. Underrepresented minority faculty members were 3.2 times more likely than white faculty members to serve on the IRB. Clinical researchers were 1.64 times more likely to be on an IRB than were faculty members who conducted nonclinical research. No significant difference was found in the average number of articles published in the three years before the study comparing IRB faculty to non-IRB faculty. Conclusions The faculty members who serve on IRBs tend to have research experience and knowledge that may be used to inform their IRB-related activities. However, the fact that almost half of all faculty IRB members serve as consultants to industry raises potential conflicts of interest.

Multiple imputation inference for multivariate multilevel continuous data with ignorable non-response

Methods specifically targeting missing values in a wide spectrum of statistical analyses are now part of serious statistical thinking due to many advances in computational statistics and increased awareness among sophisticated consumers of statistics. Despite many advances in both theory and applied methods for missing data, missing-data methods in multilevel applications lack equal development. In this paper, I consider a popular inferential tool via multiple imputation in multilevel applications with missing values. I specifically consider missing values occurring arbitrarily at any level of observational units. I use Bayesian arguments for drawing multiple imputations from the underlying (posterior) predictive distribution of missing data. Multivariate extensions of well-known mixed-effects models form the basis for simulating the posterior predictive distribution, hence creating the multiple imputations. The discussion of these topics is demonstrated in an application assessing correlates to unmet need for mental health care among children with special health care needs.

Using Calibration to Improve Rounding in Imputation

Since the 1990s, imputation methods have become increasingly accessible in standard software that typically assume a multivariate normal (MVN) distribution for incompletely observed variables. When these variables are not normally distributed but rather categorical (binary or ordinal), practitioners are often advised to round the MVN imputations to the nearest integer, but this simple procedure can lead to biased estimates. We propose practical rounding rules to be used with the existing imputation methods (e.g., under MVN) to obtain usable imputations with small biases for estimation of means and correlations. The rounding rules are calibrated in the sense that values reimputed for observed data have distributions similar to those of the observed data. Calibration in this sense is a form of posterior predictive check that can be used to evaluate any imputation procedure. It is readily implemented by duplicating the data and comparing the distributions of observed and imputed data. We calculate asymptotic biases of marginal means and slope coefficients under plausible models to assess the performance of our method.

Longitudinal analysis of physical activity, fluid intake, and graft function among kidney transplant recipients

Self‐care is recommended to kidney transplant recipients as a vital component to maintain long‐term graft function. However, little is known about the effects of physical activity, fluid intake, and smoking history on graft function. This longitudinal study examined the relationship between self‐care practices on graft function among 88 new kidney transplant recipients in Chicago, IL and Albany, NY between 2005 and 2008. Participants were interviewed, completed surveys, and medical charts were abstracted. Physical activity, fluid intake, and smoking history at baseline were compared with changes in estimated glomerular filtration rate (eGFR) (every 6 months up to 1 year) using bivariate and multivariate regression analysis, while controlling for sociodemographic and clinical transplant variables. Multivariate analyses revealed that greater physical activity was significantly (P < 0.05) associated with improvement in GFR at 6 months; while greater physical activity, absence of smoking history, and nonwhite ethnicity were significant (P < 0.05) predictors of improvement in GFR at 12 months. These results suggest that increasing physical activity levels in kidney recipients may be an effective behavioral measure to help ensure graft functioning. Our findings suggest the need for a randomized controlled trial of exercise, fluid intake, and smoking history on GFR beyond 12 months.

Medicaid Managed Care and the Unmet Need for Mental Health Care among Children with Special Health Care Needs

OBJECTIVE To determine the association between Medicaid managed care pediatric behavioral health programs and unmet need for mental health care among children with special health care needs (CSHCN). DATA SOURCE The National Survey of CSHCN (2000-2002), using subsets of 4,400 CSHCN with Medicaid and 1,856 CSHCN with Medicaid and emotional problems. Additional state-level sources were used. STUDY DESIGN Multilevel models investigated the association between managed care program type (carve-out, integrated) or fee-for-service (FFS) and reported unmet mental health care need. DATA COLLECTION/EXTRACTION METHODS The National Survey of CSHCN conducted telephone interviews with a sample representative at both the national and state levels. PRINCIPAL FINDINGS In multivariable models, among CSHCN with only Medicaid, living in states with Medicaid managed care (odds ratio [OR]=1.81; 95 percent confidence interval: 1.04-3.15) or carve-out programs (OR=1.93; 1.01-3.69) were associated with greater reported unmet mental health care need compared with FFS programs. Among CSHCN on Medicaid with emotional problems, the association between managed care and unmet need was stronger (OR=2.48; 1.38-4.45). CONCLUSIONS State Medicaid pediatric behavioral health managed care programs were associated with greater reported unmet mental health care need than FFS programs among CSHCN insured by Medicaid, particularly for those with emotional problems.

Impact of non-normal random effects on inference by multiple imputation: A simulation assessment

Multivariate extensions of well-known linear mixed-effects models have been increasingly utilized in inference by multiple imputation in the analysis of multilevel incomplete data. The normality assumption for the underlying error terms and random effects plays a crucial role in simulating the posterior predictive distribution from which the multiple imputations are drawn. The plausibility of this normality assumption on the subject-specific random effects is assessed. Specifically, the performance of multiple imputation created under a multivariate linear mixed-effects model is investigated on a diverse set of incomplete data sets simulated under varying distributional characteristics. Under moderate amounts of missing data, the simulation study confirms that the underlying model leads to a well-calibrated procedure with negligible biases and actual coverage rates close to nominal rates in estimates of the regression coefficients. Estimation quality of the random-effect variance and association measures, however, are negatively affected from both the misspecification of the random-effect distribution and number of incompletely-observed variables. Some of the adverse impacts include lower coverage rates and increased biases.

Random-covariances and mixed-effects models for imputing multivariate multilevel continuous data

Principled techniques for incomplete data problems are increasingly part of mainstream statistical practice. Among many proposed techniques so far, inference by multiple imputation (MI) has emerged as one of the most popular. While many strategies leading to inference by MI are available in cross-sectional settings, the same richness does not exist in multilevel applications. The limited methods available for multilevel applications rely on the multivariate adaptations of mixed-effects models. This approach preserves the mean structure across clusters and incorporates distinct variance components into the imputation process. In this paper, I add to these methods by considering a random covariance structure and develop computational algorithms. The attraction of this new imputation modelling strategy is to correctly reflect the mean and variance structure of the joint distribution of the data and allow the covariances differ across the clusters. Using Markov chain Monte Carlo techniques, a predictive distribution of missing data given observed data is simulated leading to creation of MIs. To circumvent the large sample size requirement to support independent covariance estimates for the level-1 error term, I consider distributional impositions mimicking random-effects distributions assigned a priori. These techniques are illustrated in an example exploring relationships between victimization and individual and contextual level factors that raise the risk of violent crime.