A well-designed and conducted study raises issues whether cognitive behavioral therapy interventions (CBT), including mindfulness-based abstinence therapy (MBAT) are obsolete and no longer cost effective.
- The study reported that intensive CBT and MBAT were no more effective than usual care consisting of 4 5-10 minute individual counseling sessions.
- The article in the study failed to note an obvious reason why the more intensive therapies were no more effective than the lighter usual care.
- Results were spun to create the impression that although MBAT was not more effective in achieving smoking cessation, it may be more effective than CBT or usual care in promoting recovery from relapse.
- Overall, the study raises important issues about whether cognitive behavioral interventions are now obsolete and represents another example of confirmation bias in the mindfulness literature.
The study is
Vidrine JI, Spears CA, Heppner WL, Reitzel LR, Marcus MT, Cinciripini PM, Waters AJ, Li Y, Nguyen NT, Cao Y, Tindle HA. Efficacy of Mindfulness-Based Addiction Treatment (MBAT) for Smoking Cessation and Lapse Recovery: A Randomized Clinical Trial. Journal of Consulting and Clinical Psychology. 2016 May.
The abstract is
Objective: To compare the efficacy of Mindfulness-Based Addiction Treatment (MBAT) to a Cognitive Behavioral Treatment (CBT) that matched MBAT on treatment contact time, and a Usual Care (UC) condition that comprised brief individual counseling. Method: Participants (N _ 412) were 48.2% African American, 41.5% non-Latino White, 5.4% Latino, and 4.9% other, and 57.6% reported a total annual household income _ $30,000. The majority of participants were female (54.9%). Mean cigarettes per day was 19.9 (SD= 10.1). Following the baseline visit, participants were randomized to UC (n =103), CBT (n = 155), or MBAT (n = 154). All participants were given self-help materials and nicotine patch therapy. CBT and MBAT groups received 8 2-hr in-person group counseling sessions. UC participants received 4 brief individual counseling sessions. Biochemically verified smoking abstinence was assessed 4 and 26 weeks after the quit date. Results: Logistic random effects model analyses over time indicated no overall significant treatment effects (completers only: F(2, 236) = 0.29, p = .749; intent-to-treat: F(2, 401) = 0.9, p = .407). Among participants classified as smoking at the last treatment session, analyses examining the recovery of abstinence revealed a significant overall treatment effect, F(2, 103) = 4.41, p = .015 (MBAT vs. CBT: OR = 4.94, 95% CI: 1.47 to 16.59, p =.010, Effect Size = .88; MBAT vs. UC: OR = 4.18, 95% CI: 1.04 to 16.75, p =.043, Effect Size = .79). Conclusion: Although there were no overall significant effects of treatment on abstinence, MBAT may be more effective than CBT or UC in promoting recovery from lapses.
[Note how minimal and nonsignificant the differences were between the treatments in this large N study. What critical detail in the abstract might explain this lack of differences? Are you impressed by the suggestion that MBAT could nonetheless be more effective than the other treatments?
A boxed statement in the article declares the public health implications of the study.
What is the public health significance of this article?
Although there were no significant differences in overall abstinence between Mindfulness-Based Addiction Treatment (MBAT) and traditional Guideline-based treatments within a diverse and relatively low SES sample of smokers, MBAT may be more efficacious than CBT or UC in facilitating lapse recovery.
Exceptional strengths of the study.
The investigators recruited a large sample of low income participants with a strong minority representation.
Given the characteristics of the population, there were reasonably good retention rates. Analyses were conducted both intent-to-treat and for completers only, as seen in the chart below:
Measurement of smoking cessation was by both self-report and an objective measure of abstinence, biologically confirmed carbon monoxide (CO) levels.
Self-reported use of mindfulness meditation before recruitment was assessed, along with actual practice of mindfulness techniques during treatment.
The CBT and MBAT treatments were carefully described in the frequency and intensity.
What the authors fail to note in their discussion of null results.
Although the usual care involved minimal contact and support, levels were far below the levels for the active CBT and MBAT treatments in frequency and intensity. In that sense, the UC group was not adequate as a control group. But it makes the lack of differences in results all the more impressive. All conditions involved distribution of nicotine patches and repeated presentation of instructions for use.
Lack of differences between groups raises issues of whether more intensive cognitive behavioral treatments add anything to supplying nicotine patches with instructions. There have been no previous well powered and conducted studies providing patches across active treatment and control conditions.
Why I don’t put much weight in the authors’ claims about the superiority of MBAT facilitating relapse recovery.
The study involved randomizing patients to one of three treatments: CBT, MBAT or usual care. The analyses from which the authors claim facilitation of relapse were post-hoc subgroup analyses, which lost the benefits of having come from randomized trial. Such analyses are certainly not equivalent to randomizing patients who are smoking to one of these three conditions, which we know from primary analyses did not produce differences. Focusing on these patients introduces unknown biases.
Note also that there were differences across conditions in rates of patients still smoking at the end of the study.
Among participants classified as smoking on the last treatment session (completers only; n = 145), 14.7% in UC, 7.0% in CBT, and 27.8% in MBAT had recovered abstinence 1 week following the end of treatment.
In the subsample of participants still smoking, there were differences in those who recovered abstinence.
Among participants classified as smoking on the last treatment session (completers only; n = 145), 14.7% in UC,7.0% in CBT, and 27.8% in MBAT had recovered abstinence 1 week following the end of treatment.
Although the trial randomized patients to a particular treatment, the post-hoc subgroup analyses focused on the nonrandomized selection of patients still smoking in each group. Rates of such patients were four times higher in the MBAT versus the CBT group. These still-smoking patients are nonrandom (biased) selections from the original sample. The results claimed to be significant for transitions from still smoking at the end of treatment to abstinence.
Let’s return to how these results were presented in the abstract:
Among participants classified as smoking at the last treatment session, analyses examining the recovery of abstinence revealed a significant overall treatment effect, F(2, 103) = 4.41, p = .015 (MBAT vs. CBT: OR = 4.94, 95% CI: 1.47 to 16.59, p =.010, Effect Size = .88; MBAT vs. UC: OR = 4.18, 95% CI: 1.04 to 16.75, p =.043, Effect Size = .79).
So, we are being asked to place a lot of confidence in post-hoc results to get past the impression that this was an utterly null trial. There were no differences in the primary outcome at the end of treatment, but these differences being emphasized emerged in a nonrandom selection of patients. The largest differences in recovery of abstinence was for participants who were sill smoking after MBAT versus CBT, but MBAT participants had the highest rates of still smoking at the end of treatment. As for MBAT versus usual care, I don’t think we should attach much practical importance to p= .043.
Are the authors persuasive in claiming that they have demonstrated a benefit for MBAT over CBT? Over Usual Care?
I think the most straightforward conclusion is that adding intensive CBT and mindfulness-based treatments to simple support of use of nicotine patches does not accomplish much, at least in this population.