Application of random-effects pattern-mixture models for missing data in longitudinal studies.

Hedeker, Donald; Gibbons, Robert D.

doi:10.1037/1082-989x.2.1.64

Cited by 803 publications

(792 citation statements)

References 64 publications

Supporting

Mentioning

785

Contrasting

Order By: Relevance

“…The inclusion of a term for dropout was based on the pattern-mixture approach to controlling for the nonrandom nature of missing data-i.e. for the possibility that dropouts differed in some systematic way from study completers (Hedeker and Gibbons, 1997).…”

Section: Discussionmentioning

confidence: 99%

Self-report of illicit benzodiazepine use on the Addiction Severity Index predicts treatment outcome

Ghitza

Epstein

Preston

2008

Drug and Alcohol Dependence

View full text Add to dashboard Cite

The relationship between pretreatment illicit benzodiazepine use (days of use in the last 30) assessed on the Addiction Severity Index (ASI) and treatment outcome was investigated by retrospective analysis of data from two controlled clinical trials in 361 methadone maintained cocaine/opiate users randomly assigned to 12-week voucher-or prize-based contingency management (CM) or control interventions. Based on screening ASI, participants were identified as nonusers (BZD-N; 0 days of use) or users (BZD-U; >0 days of use). Outcome measures were: urine drug screens (thrice weekly); quality of life and self-reported HIV-risk behaviors (every 2 weeks); and current DSM-IV diagnosis of cocaine and heroin dependence (study exit). In the CM group, BZD-U had significantly worse outcomes on in-treatment cocaine use, quality-of-life scores, needle-sharing behaviors, and current heroin dependence diagnoses at study exit compared to BZD-N. In the control group, BZD-U had significantly higher in-treatment cocaine use but did not differ from BZD-N on psychosocial measures. Thus, in a sample of non-dependent BZD users, self-reported illicit BZD use on the ASI, even at low levels, predicted worse outcome on cocaine use and blunted response to CM.

show abstract

Section: Discussionmentioning

confidence: 99%

Self-report of illicit benzodiazepine use on the Addiction Severity Index predicts treatment outcome

Ghitza

Epstein

Preston

2008

Drug and Alcohol Dependence

View full text Add to dashboard Cite

show abstract

“…The gLCPMM is both a special case of the general mixture SEM model which handles both continuous latent (e.g., growth parameters) and categorical latent (e.g., latent classes) variables and an extension of classical pattern mixture models for non-ignorably missing data (Hedeker & Gibbons, 1997;Schafer, 2003).…”

Section: Methods Primer On Statistical Simulationmentioning

confidence: 99%

“…If there is more than one class, the growth parameters (i.e., conditional growth parameter means, treatment effects) from each class are averaged and weighted by the class proportions and standard errors are calculated using the delta method (see Hedeker & Gibbons, 1997, p.74-76, Morgan-Lopez & Fals-Stewart, 2007 in order to estimate the overall treatment effect.…”

Section: Methods Primer On Statistical Simulationmentioning

confidence: 99%

“…Each dataset was analyzed under three different scenarios: (a) correctly analyzed as a 3-class gLCPMM; (b) incorrectly analyzed as a 2-class gLCPMM (which still accounts for turnover in membership and non-ignorable missingness but estimates too few classes); and (c) incorrectly analyzed as a single-class gLCPMM (analogous to a conventional gLGM model and assumes that membership turnover is irrelevant and data are missing-at-random). The parameter estimates and parameter covariances from each set of analyses were saved; for 2-and 3-class LCPMMs, the weighted averaged treatment effect estimates and standard errors were calculated using the delta method (see Hedeker & Gibbons, 1997, p.74-76, Morgan-Lopez & Fals-Stewart, 2007, supplemental materials available at http://dx.doi.org/10.1037/0022-006X.75.4.580.supp).…”

Section: Population Parametersmentioning

confidence: 99%

See 1 more Smart Citation

Consequences of misspecifying the number of latent treatment attendance classes in modeling group membership turnover within ecologically valid behavioral treatment trials

Morgan‐López

Fals‐Stewart

2008

Journal of Substance Abuse Treatment

View full text Add to dashboard Cite

Historically, difficulties in analyzing treatment outcome data from open enrollment groups have led to their avoidance in use in federally-funded treatment trials, despite the fact that 79% of treatment programs use open enrollment groups. Recently, latent class pattern mixture models (LCPMM) have shown promise as a defensible approach for making overall (and attendance classspecific) inferences from open enrollment groups with membership turnover. We present a statistical simulation study comparing LCPMMs to longitudinal growth models (LGM) to understand when both frameworks are likely to produce conflicting inferences concerning overall treatment efficacy. LCPMMs performed well under all conditions examined; meanwhile LGMs produced problematic levels of bias and Type I errors under two joint conditions: moderate-tohigh dropout (30-50%) and treatment by attendance class interactions exceeding Cohen's d ≈.2. This study highlights key concerns about using LGM for open enrollment data: treatment effect overestimation and advocacy for treatments that may be ineffective in reality.

show abstract

“…In what follows, since the focus is on describing application of the various multilevel regression models, we will make the MAR assumption. A further approach, however, that does not rely on the MAR assumption (e.g., a multilevel pattern-mixture model as described in Hedeker and Gibbons [50]) could be used. Missing data issues are described more fully in chapter 10.…”

Section: Health Services Research Examplementioning

confidence: 99%

Multilevel Models for Ordinal and Nominal Variables

Hedeker

Handbook of Multilevel Analysis

Self Cite

View full text Add to dashboard Cite

IntroductionReflecting the usefulness of multilevel analysis and the importance of categorical outcomes in many areas of research, generalization of multilevel models for categorical outcomes has been an active area of statistical research. For dichotomous response data, several approaches adopting either a logistic or probit regression model and various methods for incorporating and estimating the influence of the random effects have been developed [9,21,34,37,103,115]. Several review articles [31,39,76,90] have discussed and compared some of these models and their estimation procedures. Also, Snijders and Bosker [99, chap. 14] provide a practical summary of the multilevel logistic regression model and the various procedures for estimating its parameters. As these sources indicate, the multilevel logistic regression model is a very popular choice for analysis of dichotomous data.Extending the methods for dichotomous responses to ordinal response data has also been actively pursued [4,29,30,44,48,58,106,113]. Again, developments have been mainly in terms of logistic and probit regression models, and many of these are reviewed in Agresti and Natarajan [5]. Because the proportional odds model described by McCullagh [71], which is based on the logistic regression formulation, is a common choice for analysis of ordinal data, many of the multilevel models for ordinal data are generalizations of this model. The proportional odds model characterizes the ordinal responses in C categories in terms of C−1 cumulative category comparisons, specifically, C−1 cumulative logits (i.e., log odds) of the ordinal responses. In the proportional odds model, the covariate effects are assumed to be the same across these cumulative logits, or proportional across the cumulative odds. As noted by Peterson and Harrell [77], however, examples of non-proportional odds are

show abstract

Application of random-effects pattern-mixture models for missing data in longitudinal studies.

Cited by 803 publications

References 64 publications

Self-report of illicit benzodiazepine use on the Addiction Severity Index predicts treatment outcome

Self-report of illicit benzodiazepine use on the Addiction Severity Index predicts treatment outcome

Consequences of misspecifying the number of latent treatment attendance classes in modeling group membership turnover within ecologically valid behavioral treatment trials

Multilevel Models for Ordinal and Nominal Variables

Contact Info

Product

Resources

About