Identifiability of PBPK models with applications to dimethylarsinic acid exposure.

Any statistical model should be identifiable in order for estimates and tests using it to be meaningful. We consider statistical analysis of physiologically-based pharmacokinetic (PBPK) models in which parameters cannot be estimated precisely from available data, and discuss different types of identifiability that occur in PBPK models and give reasons why they occur. We particularly focus on how the mathematical structure of a PBPK model and lack of appropriate data can lead to statistical models in which it is impossible to estimate at least some parameters precisely. Methods are reviewed which can determine whether a purely linear PBPK model is globally identifiable. We propose a theorem which determines when identifiability at a set of finite and specific values of the mathematical PBPK model (global discete identifiability) implies identifiability of the statistical model. However, we are unable to establish conditions that imply global discrete identifiability, and conclude that the only safe approach to analysis of PBPK models involves Bayesian analysis with truncated priors. Finally, computational issues regarding posterior simulations of PBPK models are discussed. The methodology is very general and can be applied to numerous PBPK models which can be expressed as linear time-invariant systems. A real data set of a PBPK model for exposure to dimethyl arsinic acid (DMA(V)) is presented to illustrate the proposed methodology.

Fixed and random effects selection in mixed effects models.

We consider selecting both fixed and random effects in a general class of mixed effects models using maximum penalized likelihood (MPL) estimation along with the smoothly clipped absolute deviation (SCAD) and adaptive least absolute shrinkage and selection operator (ALASSO) penalty functions. The MPL estimates are shown to possess consistency and sparsity properties and asymptotic normality. A model selection criterion, called the IC(Q) statistic, is proposed for selecting the penalty parameters (Ibrahim, Zhu, and Tang, 2008, Journal of the American Statistical Association 103, 1648-1658). The variable selection procedure based on IC(Q) is shown to consistently select important fixed and random effects. The methodology is very general and can be applied to numerous situations involving random effects, including generalized linear mixed models. Simulation studies and a real data set from a Yale infant growth study are used to illustrate the proposed methodology.

VARIABLE SELECTION FOR REGRESSION MODELS WITH MISSING DATA.

We consider the variable selection problem for a class of statistical models with missing data, including missing covariate and/or response data. We investigate the smoothly clipped absolute deviation penalty (SCAD) and adaptive LASSO and propose a unified model selection and estimation procedure for use in the presence of missing data. We develop a computationally attractive algorithm for simultaneously optimizing the penalized likelihood function and estimating the penalty parameters. Particularly, we propose to use a model selection criterion, called the IC(Q) statistic, for selecting the penalty parameters. We show that the variable selection procedure based on IC(Q) automatically and consistently selects the important covariates and leads to efficient estimates with oracle properties. The methodology is very general and can be applied to numerous situations involving missing data, from covariates missing at random in arbitrary regression models to nonignorably missing longitudinal responses and/or covariates. Simulations are given to demonstrate the methodology and examine the finite sample performance of the variable selection procedures. Melanoma data from a cancer clinical trial is presented to illustrate the proposed methodology.

Variable selection in the cox regression model with covariates missing at random.

We consider variable selection in the Cox regression model (Cox, 1975, Biometrika 362, 269-276) with covariates missing at random. We investigate the smoothly clipped absolute deviation penalty and adaptive least absolute shrinkage and selection operator (LASSO) penalty, and propose a unified model selection and estimation procedure. A computationally attractive algorithm is developed, which simultaneously optimizes the penalized likelihood function and penalty parameters. We also optimize a model selection criterion, called the IC(Q) statistic (Ibrahim, Zhu, and Tang, 2008, Journal of the American Statistical Association 103, 1648-1658), to estimate the penalty parameters and show that it consistently selects all important covariates. Simulations are performed to evaluate the finite sample performance of the penalty estimates. Also, two lung cancer data sets are analyzed to demonstrate the proposed methodology.

Multicultural medication adherence: a comparative study.

The purpose of this study was to evaluate the effect of several interventions on improving medication adherence among White, Black, and Hispanic older women. A total of 109 women older than age 65 who were participating in a clinical osteoporosis trial were recruited for this 12-month study examining medication adherence. After baseline medication adherence was assessed, participants underwent standardized teaching. Participants were contacted monthly by telephone and were seen in a clinic setting every 3 months. All participants used a pillbox for 6 months, and the minority women used an electronic monitoring bottle for 6 months. Adherence was highest in White women. Black women showed significant improvement in adherence at 9 and 12 months, and Hispanic women demonstrated a significant increase in adherence at 12 months. The use of electronic monitors had a positive effect on adherence for the minority women.

Measuring disability in Medicare home care patients: application of Rasch modeling to the outcome and assessment information set.

BACKGROUND: The Outcome and Assessment Information Set (OASIS) is the universal clinical assessment tool for adult nonmaternity patients receiving skilled care at home from Medicare-certified home health agencies in the United States. Anticipating increased use of OASIS data for research purposes, this article explored the usefulness of Rasch modeling to address disability measurement challenges presented by the unique response category structure of the seven activities of daily living (ADL) and eight instrumental ADL (IADL) items in the OASIS. OBJECTIVES: To illustrate how Rasch model statistics can be used to evaluate OASIS ADL and IADL item unidimensionality and model fit; to illustrate how Rasch modeling simultaneously estimates ADL and IADL item difficulty, thresholds between item response categories, and person disability; and to compare Rasch estimates of item difficulty and person disability scores to estimates based on more conventional Likert scoring techniques. SUBJECTS: Medicare-eligible home health care patients (n = 583) served by one of 12 home care agencies in Ohio between November 1999 and September 2000. MEASURES: ADL and IADL items were measured three ways: according to the original OASIS scoring (raw Likert); transformed raw Likert scores accounting for the nonuniform item structure (corrected Likert); and Rasch Partial Credit model scores. RESULTS: The items bathing and telephone use showed evidence of unexpected response patterns; recoding of these items was necessary for good Rasch model fit. Partial Credit model results revealed that interval distances between response categories varied widely across the 15 ADL and IADL items. When ADL and IADL items were ranked by level of difficulty, results were similar between Rasch and corrected Likert measurement approaches; however, corrected Likert person scores were found to be nonlinear at highest and lowest disability levels when plotted against Rasch person scores. CONCLUSIONS: Rasch modeling can help improve the precision of disability measurement in Medicare home care patients when using ADL and IADL items from the OASIS instrument.

Rehabilitation therapy self-efficacy and functional recovery after hip fracture.

Little is known about the role of psychological factors in the functional recovery process of hip fracture patients. This study employed a prospective cohort design to test the hypothesis that hospitalized hip fracture patients with greater reported self-efficacy for conducting rehabilitation therapy would have a greater likelihood of recovering to a pre-fracture level of locomotion function six months after the fracture. This hypothesis was tested controlling for pre-fracture level of function and depressive symptoms reported during hospitalization for surgical repair. An original measure of rehabilitation therapy self-efficacy was evaluated prior to hypothesis testing. Study patients were recruited from two hospitals, interviewed during hospitalization and followed up six months later. Patients included in hypothesis test analyses (n = 24) were mostly women (82%) with a mean age of 79 years. Results showed that patients with higher self-efficacy scores had a greater likelihood of locomotion recovery, controlling for pre-fracture locomotion function level (adjusted odds ratio (AOR) = 1.21; 95% confidence interval (CI) = 1.00-1.45; P= 0.05). This positive association between rehabilitation therapy self-efficacy and likelihood of locomotion recovery persisted after adding depressive symptoms (the Center for Epidemiological Studies-depression (CES-D) score) to this logistic regression model (AOR for self-efficacy = 1.18; 95% CI = 0.99-1.42; P= 0.07). It is concluded that rehabilitation therapy self-efficacy is a potentially important psychological factor in helping hip fracture patients recover locomotion functioning.