Jonathan Blitstein

Design and Analysis of Group-Randomized Trials: A Review of Recent Methodological Developments

We review recent developments in the design and analysis of group-randomized trials (GRTs). Regarding design, we summarize developments in estimates of intraclass correlation, power analysis, matched designs, designs involving one group per condition, and designs in which individuals are randomized to receive treatments in groups. Regarding analysis, we summarize developments in marginal and conditional models, the sandwich estimator, model-based estimators, binary data, survival analysis, randomization tests, survey methods, latent variable methods and nonlinear mixed models, time series methods, global tests for multiple endpoints, mediation effects, missing data, trial reporting, and software. We encourage investigators who conduct GRTs to become familiar with these developments and to collaborate with methodologists who can strengthen the design and analysis of their trials.

Methods To Reduce The Impact Of Intraclass Correlation In Group-Randomized Trials:

This study reports intraclass correlation (ICC) for dependent variables used in group-randomized trials (GRTs). The authors also document the effect of two methods suggested to reduce the impact of ICC in GRTs; these two methods are modeling time and regression adjustment for covariates. They coded and analyzed 1,188 ICC estimates from 17 published, in press, and unpublished articles representing 21 studies. Findings confirm that both methods can improve the efficiency of analyses shown to be valid across conditions common in GRTs. Investigators planning GRTs should obtain ICC estimates matched to their planned analysis so that they can size their studies properly.

Design and Analysis of Group-Randomized Trials: A Review of Recent Practices

We reviewed group-randomized trials (GRTs) published in the American Journal of Public Health and Preventive Medicine from 1998 through 2002 and estimated the proportion of GRTs that employ appropriate methods for design and analysis. Of 60 articles, 9 (15.0%) reported evidence of using appropriate methods for sample size estimation. Of 59 articles in the analytic review, 27 (45.8%) reported at least 1 inappropriate analysis and 12 (20.3%) reported only inappropriate analyses. Nineteen (32.2%) reported analyses at an individual or subgroup level, ignoring group, or included group as a fixed effect. Hence increased vigilance is needed to ensure that appropriate methods for GRTs are employed and that results based on inappropriate methods are not published.

Design and Analysis of Group-Randomized Trials in Cancer: A Review of Current Practices

BACKGROUND Previous reviews have identified problems in the design and analysis of group-randomized trials in a number of areas. Similar problems may exist in cancer research, but there have been no comprehensive reviews. METHODS We searched Medline and PubMed for group-randomized trials focused on cancer prevention and control that were published between 2002 and 2006. We located and reviewed 75 articles to determine whether articles included evidence of taking group randomization into account in establishing the size of the trial, such as reporting the expected intraclass correlation, the group component of variance, or the variance inflation factor. We also examined the analytical approaches to determine their appropriateness. RESULTS Only 18 (24%) of the 75 articles documented appropriate methods for sample size calculations. Only 34 (45%) limited their reports to analyses judged to be appropriate. Fully 26 (34%) failed to report any analyses that were judged to be appropriate. The most commonly used inappropriate analysis was an analysis at the individual level that ignored the groups altogether. Nine articles (12%) did not provide sufficient information. CONCLUSIONS Many investigators who use group-randomized trials in cancer research do not adequately attend to the special design and analytic challenges associated with these trials. Failure to do so can lead to reporting type I errors as real effects, mislead investigators and policy-makers, and slow progress toward control and prevention of cancer. A collaborative effort by investigators, statisticians, and others will be required to ensure that group-randomized trials are planned and analyzed using appropriate methods so that the scientific community can have confidence in the published results.

Ethnic differences in predictors of adolescent smoking onset and escalation : A longitudinal study from 7th to 12th grade

The present study sought to determine whether risk factors for cigarette smoking onset have a stronger influence on Black versus White children, and whether those patterns differ for nonsmokers versus experimental smokers. As part of a larger study, 5,213 seventh graders were surveyed annually for 5 years. The survey instrument assessed cigarette smoking, along with a variety of background variables thought to influence smoking. Results indicated that rebelliousness predicted smoking onset for Black nonsmokers; peer modeling was more influential for Whites. For both ethnicities, socially isolated children were at increased risk. The variables that predicted the shift from experimental to regular smoking were for the most part different from those that predicted smoking onset. Specifically, youths who lived with smokers, viewed smoking as a means to a goal, and experienced few adverse physiological effects of smoking were more likely to increase their cigarette consumption. Overall, we found that, contrary to previous research, risk factors for smoking differed across ethnicity. Further, smoking onset appears to be shaped by forces different from those that accelerate consumption.

Rapid progression to regular cigarette smoking among nonsmoking adolescents: interactions with gender and ethnicity

BACKGROUND While the correlates and predictors associated with transition through the stages of smoking acquisition have received substantial attention, the identification of factors associated with the time course of progression remain generally unexplored. METHODS We identified adolescents escalating from nonsmoking to regular smoking in 1 year as rapid progressors (n = 98) and those escalating from nonsmoking to experimental smoking in 1 year as slow progressors (n = 555). We entered demographic and psychosocial variables, initial reactions to smoking variables, and their interactions with gender and ethnicity into a logistic regression analysis comparing rapid progressors to slow progressors. RESULTS At baseline, rapid progressors reported more friends who smoked regularly. Girls who reported more concern with dieting and less social success were more likely to progress to regular smoking within 1 year; these variables did not influence progression in boys. Those who associated positive feelings with their initial smoking experience were more likely to progress to regular smoking. Differences between white and African American youth emphasize the importance of physiological interpretations of initial smoking experiences. CONCLUSIONS Our findings suggest that even as nonsmokers, adolescents likely to progress quickly to regular smoking can be identified. Researchers designing prevention programs should take these factors into account.

Assessing Intervention Effects in a School-Based Nutrition Intervention Trial: Which Analytic Model Is Most Powerful?

This article compares four mixed-model analyses valid for group-randomized trials (GRTs) involving a nested cohort design with a single pretest and posttest. This study makes estimates of intraclass correlations (ICCs) available to investigators planning GRTs addressing dietary outcomes. It also provides formulae demonstrating the potential benefits to the standard error of the intervention effect (σΔ) from adjustments for both fixed and time-varying covariates and correlations over time. These estimates will allow other researchers using these variables to plan their studies by estimating a priori detectable differences and sample size requirements for any of the four analytic options. These methods are demonstrated using data from the Teens Eating for Energy and Nutrition at School study. Mixed-model analyses of covariance proved to be the most powerful analysis in that data set. The formulae may be applied to any dependent variable in any GRT given corresponding information for those variables on the parameters that define the formulae.

The reliability and predictive value of adolescents' reports of initial reactions to smoking

A sample of 372 adolescents completed annual surveys regarding their initial reactions to smoking (IRTS) and their current smoking status. Each annual survey asked participants if they had the following five reactions the first time they smoked a cigarette: coughing, or feeling dizzy, sick, high, or relaxed. Time 1 IRTS data were collected within 1 year of the reported initial smoking experience, and Time 2 IRTS data were collected 1 year later. Kappa values for two IRTS items, coughing (.45) and dizziness (.40), indicated modest but acceptable reliability across a 1-year period, but the remaining items showed poor reliability (all < or =.31). Logistic regression analyses indicated that no IRTS item significantly predicted regular (at least weekly) smoking 1 year later. However, when IRTS and smoking status were assessed concurrently (at Time 2), reports of feeling relaxed and not coughing during the initial episode were significantly correlated with regular smoking. Participants who progressed from experimental to regular smoking from Time 1 to Time 2 were more likely to report relaxation as an initial reaction to smoking at Time 2 after denying relaxation at Time 1.

Evaluating the Impact of Six Supplemental Nutrition Assistance Program Education Interventions on Children’s At-Home Diets

Background. Nutrition education in the Supplemental Nutrition Assistance Program Education (SNAP-Ed) is designed to promote healthy eating behaviors in a low-income target population. Purpose. To evaluate the effectiveness of six SNAP-Ed interventions delivered in child care centers or elementary school settings in increasing participating children’s at-home fruit and vegetable (F/V) consumption by 0.3 cups per day and use of fat-free or low-fat milk instead of whole or reduced-fat milk during the prior week. Method. Clustered randomized or quasi-experimental clustered trials took place in child care centers or elementary schools between 2010 and 2012. Parents of children at intervention and control sites completed baseline and follow-up surveys about their child’s at home F/V consumption and other dietary behaviors. Results. One of the six interventions was successful in meeting the objective of increasing children’s F/V consumption by 0.3 cups per day. For three of the six interventions, there was a small but statistically significant increase in F/V consumption and/or use of low-fat or fat-free milk. Conclusion. Although not all interventions were effective, these findings suggest that it is possible for some SNAP-Ed interventions to improve dietary habits among low-income children among some families. The effective interventions appear to have benefited from implementation experience and sustained efforts at intervention refinement and improvement.

Increasing the Degrees of Freedom in Future Group Randomized Trials

This article builds on the previous article by Blitstein et al. (2005), which showed how external estimates of intraclass correlation can be used to improve the precision for the analysis of an existing group randomized trial. The authors extend that work to sample size estimation and power analysis for future group-randomized trials. Often this approach will allow a smaller study than would otherwise be possible without sacrificing statistical power. Such studies are needed, for example, as pilot studies to help plan for a full-scale efficacy trial, as replication studies, or in situations in which resource constraints prohibit a larger trial. The authors discuss the circumstances under which this strategy will be most helpful and the risks associated with conducting smaller studies.