Dataset Information

Cocaine Dependence Treatment Data: Methods for Measurement Error Problems With Predictors Derived From Stationary Stochastic Processes.

ABSTRACT: In a cocaine dependence treatment study, we use linear and nonlinear regression models to model posttreatment cocaine craving scores and first cocaine relapse time. A subset of the covariates are summary statistics derived from baseline daily cocaine use trajectories, such as baseline cocaine use frequency and average daily use amount. These summary statistics are subject to estimation error and can therefore cause biased estimators for the regression coefficients. Unlike classical measurement error problems, the error we encounter here is heteroscedastic with an unknown distribution, and there are no replicates for the error-prone variables or instrumental variables. We propose two robust methods to correct for the bias: a computationally efficient method-of-moments-based method for linear regression models and a subsampling extrapolation method that is generally applicable to both linear and nonlinear regression models. Simulations and an application to the cocaine dependence treatment data are used to illustrate the efficacy of the proposed methods. Asymptotic theory and variance estimation for the proposed subsampling extrapolation method and some additional simulation results are described in the online supplementary material.

SUBMITTER: Guan Y

PROVIDER: S-EPMC3188406 | biostudies-literature | 2011 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Cocaine Dependence Treatment Data: Methods for Measurement Error Problems With Predictors Derived From Stationary Stochastic Processes.

Guan Yongtao Y Li Yehua Y Sinha Rajita R

Journal of the American Statistical Association 20110601 493

In a cocaine dependence treatment study, we use linear and nonlinear regression models to model posttreatment cocaine craving scores and first cocaine relapse time. A subset of the covariates are summary statistics derived from baseline daily cocaine use trajectories, such as baseline cocaine use frequency and average daily use amount. These summary statistics are subject to estimation error and can therefore cause biased estimators for the regression coefficients. Unlike classical measurement e ...[more]

PMID: 21984854

Similar Datasets

Project description:Deficits in cognitive control are implicated in cocaine dependence. Previously, combining functional magnetic resonance imaging and a stop signal task, we demonstrated altered cognitive control in cocaine-dependent individuals. However, the clinical implications of these cross-sectional findings and, in particular, whether the changes were associated with relapse to drug use, were not clear. In a prospective study, we recruited 97 treatment-seeking individuals with cocaine dependence to perform the stop signal task during functional magnetic resonance imaging and participate in follow-up assessments for 3 months, during which time cocaine use was evaluated with timeline follow back and ascertained by urine toxicology tests. Functional magnetic resonance imaging data were analysed using general linear models as implemented in Statistical Parametric Mapping 8, with the contrast 'stop error greater than stop success trials' to index error processing. Using voxelwise analysis with logistic and Cox regressions, we identified brain activations of error processing that predict relapse and time to relapse. In females, decreased error-related activations of the thalamus and dorsal anterior cingulate cortex predicted relapse and an earlier time to relapse. In males, decreased error-related activations of the dorsal anterior cingulate cortex and left insula predicted relapse and an earlier time to relapse. These regional activations were validated with data resampling and predicted relapse with an average area under the curve of 0.849 in receiver operating characteristic analyses. These findings provide direct evidence linking deficits in cognitive control to clinical outcome in a moderate-sized cohort of cocaine-dependent individuals. These results may provide a useful basis for future studies to examine how psychosocial factors interact with cognitive control to determine drug use and to evaluate the efficacy of pharmacological or behavioural treatment in remediating deficits of cognitive control in cocaine addicts.

Project description:The relative impact of chronic vs acute cocaine on dependence-related variability in reward processing in cocaine-dependent individuals (CD) is not well understood, despite the relevance of such effects to long-term outcomes. To dissociate these effects, CD (N=15) and healthy controls (HC; N=15) underwent MRI two times while performing a monetary incentive delay task. Both scans were identical across subjects/groups, except that, in a single-blind, counterbalanced design, CD received intravenous cocaine (30 mg/70 kg) before one session (CD+cocaine) and saline in another (CD+saline). Imaging analyses focused on activity related to anticipatory valence (gain/loss), anticipatory magnitude (small/medium/large), and reinforcing outcomes (successful/unsuccessful). Drug condition (cocaine vs saline) and group (HC vs CD+cocaine or CD+saline) did not influence valence-related activity. However, compared with HC, magnitude-related activity for gains was reduced in CD in the left cingulate gyrus post-cocaine and in the left putamen in the abstinence/saline condition. In contrast, magnitude-dependent activity for losses increased in CD vs HC in the right inferior parietal lobe post-cocaine and in the left superior frontal gyrus post-saline. Across outcomes (ie, successful and unsuccessful) activity in the right habenula decreased in CD in the abstinence/saline condition vs acute cocaine and HC. Cocaine-dependent variability in outcome- and loss-magnitude activity correlated negatively with ratings of cocaine craving and positively with how high subjects felt during the scanning session. Collectively, these data suggest dissociable effects of acute cocaine on non-drug reward processes, with cocaine consumption partially ameliorating dependence-related insensitivity to reinforcing outcomes/'liking', but having no discernible effect on dependence-related alterations in incentive salience/'wanting'. The relationship of drug-related affective sequelae to non-drug reward processing suggests that CD experience a general alteration of reward function and may be motivated to continue using cocaine for reasons beyond desired drug-related effects. This may have implications for individual differences in treatment efficacy for approaches that rely on reinforcement strategies (eg, contingency management) and for the long-term success of treatment.

Project description:Measurement errors commonly occur in 24-h hormonal data and may affect the outcomes of such studies. Measurement errors often appear as outliers in such data sets; however, no well-established method is available for their automatic detection. In this study, we aimed to compare performances of different methods for outlier detection in hormonal serial data. Hormones (glucose, insulin, thyroid-stimulating hormone, cortisol, and growth hormone) were measured in blood sampled every 10 min for 24 h in 38 participants of the Leiden Longevity Study. Four methods for detecting outliers were compared: (1) eyeballing, (2) Tukey's fences, (3) stepwise approach, and (4) the expectation-maximization (EM) algorithm. Eyeballing detects outliers based on experts' knowledge, and the stepwise approach incorporates physiological knowledge with a statistical algorithm. Tukey's fences and the EM algorithm are data-driven methods, using interquartile range and a mathematical algorithm to identify the underlying distribution, respectively. The performance of the methods was evaluated based on the number of outliers detected and the change in statistical outcomes after removing detected outliers. Eyeballing resulted in the lowest number of outliers detected (1.0% of all data points), followed by Tukey's fences (2.3%), the stepwise approach (2.7%), and the EM algorithm (11.0%). In all methods, the mean hormone levels did not change materially after removing outliers. However, their minima were affected by outlier removal. Although removing outliers affected the correlation between glucose and insulin on the individual level, when averaged over all participants, none of the 4 methods influenced the correlation. Based on our results, the EM algorithm is not recommended given the high number of outliers detected, even where data points are physiologically plausible. Since Tukey's fences is not suitable for all types of data and eyeballing is time-consuming, we recommend the stepwise approach for outlier detection, which combines physiological knowledge and an automated process.

Dataset Information

Cocaine Dependence Treatment Data: Methods for Measurement Error Problems With Predictors Derived From Stationary Stochastic Processes.

Publications

Cocaine Dependence Treatment Data: Methods for Measurement Error Problems With Predictors Derived From Stationary Stochastic Processes.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets