Letter to the Editor
Claiming Positive Results From Negative Trials: A Cause for Concern in Randomized Controlled Trial Research
J Med Internet Res 2013;15(8):e179)
One of the challenging issues facing the randomized controlled trial (RCT) researcher is how to interpret the results of studies where there are improvements in the behaviour under study but where the degree of improvement does not differ between the experimental conditions . This is especially a challenge when the RCT involves the comparison of two or more interventions rather than an intervention compared to some form of no-intervention control group.
One possible cause of the observed improvement in such trials is that both interventions were “active” - that both interventions were effective in facilitating or causing a change among participants. Unfortunately, there is no way to determine if this claim is definitely true from the results of a negative RCT. Other interpretations of the results include: 1) that the change over time is due to regression to the mean [, ]; 2) due to natural history maturation (meaning that participants were in a period in their lives where, on average, a downward trend in quantity of drinking could be expected); or 3) the trial recruited participants who were already motivated to change and who would have done so anyway without exposure to the interventions under study [ ].
Any of these alternate explanations could apply to the recent trial by Hester and colleagues . Further, there is a well-established finding in the alcohol research field that participants in the no intervention control condition of intervention trials show improvements in their drinking from baseline to follow-up [ ]. This may be particularly the case in trials recruiting participants from the community rather than from treatment settings where intractable alcohol problems are more common [ ]. Essentially, the assumption that any changes over time are due to the intervention in a negative trial is predicated on the assumption that the participants would show no improvement without receiving some type of intervention. There may be some behaviours where this is the case. However, alcohol abuse is demonstrably not one of them. Thus, it is unwise to favour an intervention effect explanation over other causes when faced with the results of an RCT where participants show improvement over time but that there are no significant statistical differences between intervention conditions.
- Hester RK, Lenberg KL, Campbell W, Delaney HD. Overcoming Addictions, a Web-Based Application, and SMART Recovery, an Online and In-Person Mutual Help Group for Problem Drinkers, Part 1: Three-Month Outcomes of a Randomized Controlled Trial. J Med Internet Res 2013;15(7):e134 [FREE Full text] [CrossRef] [Medline]
- Barnett AG, van der Pols JC, Dobson AJ. Regression to the mean: what it is and how to deal with it. Int J Epidemiol 2005 Feb;34(1):215-220 [FREE Full text] [CrossRef] [Medline]
- Morton V, Torgerson DJ. Effect of regression to the mean on decision making in health care. BMJ 2003 May 17;326(7398):1083-1084 [FREE Full text] [CrossRef] [Medline]
- Cunningham JA. Regression to the mean: what does it mean? Alcohol Alcohol 2006 Oct;41(5):580; author reply 581 [FREE Full text] [CrossRef] [Medline]
- Moyer A, Finney JW. Outcomes for untreated individuals involved in randomized trials of alcohol treatment. J Subst Abuse Treat 2002 Oct;23(3):247-252. [Medline]
- Cunningham JA, McCambridge J. Is alcohol dependence best viewed as a chronic relapsing disorder? Addiction 2012 Jan;107(1):6-12 [FREE Full text] [CrossRef] [Medline]
Edited by G Eysenbach; submitted 23.07.13; peer-reviewed by A Moyer, R Hester; accepted 15.08.13; published 19.08.13
©John A Cunningham. Originally published in the Journal of Medical Internet Research (http://www.jmir.org), 19.08.2013.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.