Wednesday, December 9, 2015

Cochrane Review: ADHD medications have lower-quality evidence than most people believe

A new Cochrane review, published on November 25, 2015 by authors Storebø and Zwi, looked at the use of stimulants to treat childhood ADHD (specifically, methylphenidate).  Their conclusions included:

1.  "the low quality of the underpinning evidence means that we cannot be certain of the magnitude of the effects."
2. "the general perception of methylphenidate as an effective drug for all children with ADHD seems out of step with the new evidence."  

The authors found a great deal of industry sponsorship in existing studies, and found that "all 185 trials" had a high risk of bias.   I would add that more recent ADHD studies, involving newer, more expensive medications, are most likely at an even higher risk of similar biases.  

In general, the weaknesses of existing data are similar to the weaknesses in much other psychiatric research:  studies are usually brief, rather than long-term, despite  treatments often being given for many years or permanently.  

Other authors, such as Hinshaw (2015)[http://www.ncbi.nlm.nih.gov/pubmed/26262927] have reached similar conclusions, including
"the diminution of medication's initial superiority [was apparent] once the randomly assigned treatment phase turned into naturalistic follow-up. The key paradox is that while ADHD clearly responds to medication and behavioral treatment in the short term, evidence for long-term effectiveness remains elusive"

How should this information inform our understanding or management of ADHD?

First, I do not think it is necessary to stop using stimulants as a treatment.  However, I do think it is necessary to step away from the assumption that long-term stimulant use is appropriate for every person with ADHD symptoms.   Other ways of using stimulant medication could often be more appropriate for many, such as using stimulants sporadically, to manage attentional symptoms for brief periods of time.  

The evidence also does not strongly support the long-term effectiveness of behavioural therapies.  This, too, is not really surprising to me.  

I think that the answer lies in moving away from a highly medicalized, reductionistic approach entirely.  Phenomena such as ADHD have broad biopsychosocial underpinnings:  some factors exist within the individual, while many others exist in family, social, and educational structures.   In some ways this is similar to other public health issues, such as obesity or addictions:  a single medication or behavioural treatment is very unlikely to be a remarkably effective strategy to help with these problems.    Yet, each of these strategies has a role, provided that the role is not overvalued by those offering it.  Other larger social factors are extremely important as well, including factors relating to poverty, economic equality, community supports, provision of justice & public safety, etc. 

So in conclusion, I see -- not surprisingly -- that we must not have exaggerated expectations of medication for treating ADHD or any other psychiatric phenomena.  I do think stimulants have an important role, however, for many people, provided that the expectations are modest, and provided that side effect risks are not discounted by an over-enthusiastic prescriber with biased beliefs about long-term effectiveness vs. risk. 

It is also important not to be biased against any particular treatment.  In some cases, for example, balanced medication treatment of ADHD could reduce various types of risks, including substance use problems and traffic accidents, etc.   It is just that the magnitude of such protective effects are likely to be exaggerated in most practioners' minds, due to the biases described above. 

 As with other life issues, I believe it is necessary to have a very broad view about helping strategies, which includes other types of therapeutic support if desired, as well as attention given to community, educational, cultural, and family resources--not in isolation, but in a comprehensive and holistic way.


No comments: