Randomized Controlled Trials (RCT) often provide rather reliable and reproducible results. Thus they are considered the Gold standard. They are very useful when you want to evaluate a few clearly defined distinct options. Like the efficacy of medical therapies.
are in the design and interpreting the meaning of the results. E.g.:
There was a RCT comparing the efficacy of NRT, ecigs, and “cold turkey” [ https://www.ncbi.nlm.nih.gov/pubmed/24029165 ]. While technically solid, the problem is: This trial wasn’t modelled on reality. The results, good as they are, were still disappointing and don’t seem to match with the abundant anecdotes. The explanation is simple. The scientists appearently started with erroneous assumptions:
“All ecigs are basically the same.”
Well, they aren’t. Picking just one model with a specific liquid cripples the results so that they simply can’t match the reality, where ecigs users usually have to try several devices and many liquids before they find combinations they are comfortable with. ->Pleasure Principle ->Don’t Dismiss Differences
“Ecigs are just like any other NRT.”
Some people might use them this way. If you prescribe a regime for the use of ecigs like for NRTs, you get nice reproduceable data. On the downside this tight control also has a demolishing influence on exactly what you are looking for. ->Vaping vs.NRT
The next time you do a RCT to evaluate the efficacy of ecigs used for cessation, simply add another group. A group without any of these restrictions. On the contrary! Try to give them the best case scenario:
- Get them in contact with experienced vapers. Personal contact would be best. But contacts via internet would help, too.
- Give them a reasonable budget to buy what they decide after consultation. Including a set of liquids.
Ok, the numbers of this group won’t be reliable and reproducible because of all the uncontrolled variables. But it would serve as a control on how much “damage” the study design did to the applicability of the results. And please be explicit about your ->Definition of “cessation” Better yet: Differentiate which levels of cessations where reached.
Please consider, what a real scientist – Carl V. Phillips – has to say: http://antithrlies.com/tag/rcts/
Heisenberg’s uncertainty principle
also is relevant for RCTs. It’s rather easy and efficient when you only have just a few distinct option you want to evaluate. But when you try to evaluate a multifactorial phenomenon like the success of vaping you have a Heisenberg problem. The nicotine content is just one of many factors for the appeal and success of vaping vs smoking. Not even the major factor for most (ex-)smokers. You can get a hint about that when you examine the internet surveys of mostly successful switchers from this point of view. They too often are flippantly dismissed because of their Selection Bias.
For a RCT to realistically model vaping as an alternative to smoking you either have to build an extremely complex (and expensive) RCT that examines the many factors or relax the control and allow the participants lots of choices (like nicotine level, flavors, devices). You’ll have to chose a compromise between these mutually exclusive properties for this RCT: cost efficiency, applicability, reliability.
Effectiveness of the Electronic Cigarette: An Eight-Week Flemish Study with Six-Month Follow-up on Smoking Reduction, Craving and Experienced Benefits and Complaints
This is a nice example of a study that avoided some of the pitfalls of control I mention. But still they encouraged the limitation to a single type of liquid. But anyway the results are very encouraging. Just imagine how much better the results might have been for a group where the smokers were allowed to chose and change the device and the liquids ad lib, too. My “theory”. That is what happened with the “smokers” group when they were properly presented with the vaping option after the last lab session, when the limiting restrictions wwere removed. It might explain why this this group hat the best switch rate in the end. I suspect that some in the ecigs groups were conditioned by the limitations of the lab period and simply didn’t consider exploring other possibilities.
More about this study and models in general: Carl V Phillips: This works in practice, now we just need to see if it works in theory
Bias in public health research: with examples from e-cigarette research