Modeling longitudinal drinking data in clinical trials: An application to the COMBINE study

Abstract

Background There is a lack of consensus in the literature as to how to define drinking outcomes in clinical trials. Typically, separate statistical models are fit to assess treatment effects on a number of summary drinking measures. These summary measures do not capture the complexity of drinking behavior. We used the COMBINE Study to illustrate a statistical approach for examining treatment effects on high-resolution drinking data, which takes into account abstinence and non zero drinking in the same analysis. Methods This is a secondary data analysis of COMBINE (n=1195) participants randomly assigned to naltrexone, acamprosate, with medical management and/or Combined Behavioral Intervention (CBI). Using a Poisson hurdle model, abstinence and number of drinks were simultaneously modeled with treatment, prior drinking, week, and study center as covariates. Odds ratios (OR) for abstinence and relative risk (RR) for drinking are reported. A special emphasis was placed on the evaluation of “risky drinking” defined as 3 drinks per day for women and 4 for men. Results During treatment, naltrexone increased the odds of abstinence versus placebo naltrexone (OR=1.35[1.06,1.65]) but receiving CBI in addition to naltrexone (versus not) obscured this effect; thus, the naltrexone effect was largest in the group not receiving CBI (OR=1.87[1.29,2.46]). Naltrexone versus placebo naltrexone also reduced the risk of drinking in people who resumed risky drinking, defined as more than 3 and 4 drinks/day for women and men, respectively (RR=0.58[0.24,0.93]) and increased the odds of maintaining low risk drinking (OT=1.99[1.07,2.90]). Both effects were strongest in the absence of CBI when only “medical management” was provided. Conclusions Naltrexone promotes both abstinence and reduction in drinking once risky drinking is resumed. The finding that the rate of risky drinking is reduced once a slip has occurred bolsters support for the use of naltrexone, especially since this was observed in the context of a medical management approach that could be delivered in various health care settings. The utilization of a hurdle model adds to prior reports on summary drinking measures, which found no effect of naltrexone on abstinence, did not evaluate its effect closely on risk drinking, and did not analyze weekly drinking behavior.

Zero-inflated count models for longitudinal measurements with heterogeneous random effects

“…Naltrexone significantly decreases the risk of drinking, which is consistent with the original COMBINE report (where naltrexone reduced the number of drinks per day). The original report did not find an effect of naltrexone on abstinence; however, this was detected in the current paper (and in other reports that consider the longitudinal vector of drinking) because of the way abstinence is operationalized in the zero-inflated model 45 .…”

Internet-delivered cognitive-behavioral therapy for significant others of treatment-refusing problem gamblers: A randomized wait-list controlled trial.

“…However, a zero-inflated model might still be useful if there are reasons to suspect two distinct processes that generate zero-observations. DeSantis et al (2013) found that a hurdle-Poisson model worked well to evaluate treatment effects from high-resolution drinking data. They also found that placing the hurdle at a "low-risk"-cutoff of 4 to 5 for the number of drinks per day, fit the data better than a hurdle at zero.…”

Bayesian Joint Modelling of Longitudinal Data on Abstinence, Frequency and Intensity of Drinking in Alcoholism Trials

“…Wang et al (2002) discussed several basic methods for analysing drinking data (such as reducing the repeated measures of drinking over the whole study period to one summary for each subject, such as the proportion of drinking days or average alcohol consumption per day, and then analysing these univariate summaries, or using survival models to analyse time to the first (any or heavy) drinking day) and argued for the use of more sophisticated statistical methods (specifically, recurrent event survival analyses) that give a more comprehensive description of drinking behaviour over time. Other researchers (DeSantis et al, 2013;Zhu et al, 2017) proposed a longitudinal hurdle or zero-inflated Poisson model to analyse the zero-inflated average number of drinks per day (as opposed to the average number of drinks per drinking day) for each week (rounded to the nearest whole number). This artificially inflates the proportion of zeros by rounding and ignores the number of days in which drinking occurred during the week.…”

“…For subject B, who drinks heavily during the weekend (seven drinks on Saturday and seven on Sunday) and abstains for the rest of the week, we model the number of drinking days (2) and the average of seven drinks per drinking day. Previous literature using average drinks per day (DeSantis et al, 2013;Zhu et al, 2017) would solely model an average of 2 in both cases. Distinguishing between the two drinking patterns of subjects A and B is important because these two types of individual could have different future health and psychosocial outcomes, and treatments may be associated with different drinking patterns.…”