Does medication improve reading ability?

September 05th, 2016

The Journal of Child and Adolescent Psychopharmocology has published an article by Shaywitz et al that investigated the affects of the drug atomoxetine (sold as Strattera) on reading and attention in children with dyslexia, ADHD and comorbid dyslexia + ADHD.

My understanding is that the experiment was part of a broader study on the effectiveness of atomoxetine as a treatment for ADHD. The study appears to have been funded by Eli Lilly and the second author is reportedly an employee and minor shareholder in Eli Lilly.

Children were defined as having dyslexia if (a) there was at least a 22-point discrepancy (approx. 43 percentile ranks) between IQ and word-reading ability (defined by the WJIII Word Attack test OR the Word Identification OR the WJIII Basic Reading Cluster – an average of Word Attack and Word Id) or (b) the child had a standard score of 89 or lower on at least one of the WJIII tests.

The first thing that springs to mind is that these criteria are not particularly stringent. A standard score of 89 represents the 23rd percentile. It is still within a standard deviation of the mean. Further, the discrepancy criterion presumably means that a child with an IQ score of 122 and a reading score of 100 (exactly average) met criteria for the dyslexia group.

Using these criteria, 209 children who were allocated at random to an atomoxetine or placebo group. Numbers in each group were: dyslexia only (n = 29 atomoxetine, n = 29 placebo), ADHD + dyslexia (n = 64 atomoxetine, n = 60 placebo) and ADHD only (all of whom received atomoxetine).

The drug trial lasted for 16-weeks. Changes in reading skills were measured by:

WJIII Word Attack and Word Identification tests

WJIII Spelling

WJIII Reading Comprehension and Reading Vocabulary

Comprehensive Test of Phonological Processing (CTOPP)

Gray Oral Reading Tests-4 (GORT-4)

Test of Word Reading Efficiency (TOWRE)

It is unclear why but on many variables the Placebo group had higher pre-treatment reading/spelling skills than the Atomoxetine group. For example, within the Dyslexia only group the Placebo sub-group had an average standard score on the Word Attack subtest of 88.05 while the Atomoxetine group had an average score of 84.11. It was 83.86 vs 80.47 for Word Identification, 83.32 vs 77.32 for Spelling. This may seem a minor point but consider the effect of regression to the mean The more extreme a score is at pre-test the more likely it is to “bounce’ back towards average at post-testing. Thus, the Atomoxetine group was more likely to “benefit” from regression effects than the Placebo group.

A summary of the interesting bits of the results:

  • Children with dyslexia who were treated with atomoxetine made greater gains than children with dyslexia given the placebo treatment on WJIII Word Attack, WJIII Basic Reading Cluster and WJIII Reading Vocabulary.
  • Children with dyslexia + ADHD who were treated with atomoxetine made greater gains than children with dyslexia + ADHD given the placebo treatment on the CTOPP Elision test (a peripheral/distal reading sub-skill).

So what?

  1. At first glance it seems odd that children with dyslexia would respond better to a medication designed for ADHD than children who actually had ADHD. However, there are reasonable explanations. First, medication isn’t a panacea. It doesn’t cure ADHD. It improves symptoms. It is possible that the ADHD + dyslexia remained more impaired than the dyslexia group even when given medication. Second, children with dyslexia often have sub-threshold symptoms of ADHD that may in fact respond better to medication than the more severe symptoms seen in children actually diagnosed with ADHD.
  2. It is possible that at least some of the effects were due to greater regression to the mean in the Atomoxetine vs Placebo group (see above).
  3. It is possible that the medication simply improved test behaviour rather than reading ability per se.

The data are arguably the beginning of a research journey. They provide some preliminary support for the idea that psycho-stimulant medication can improve academic skills even in children without ADHD. However, I am sceptical about this. I can see how medication might improve test-taking behaviour in some children. I can also see how it might “smooth out” the inconsistency seen in children with attention problems and with dyslexia. However, medication doesn’t teach. It doesn’t matter how good your attention is; if you don’t know it you don’t know it.

I am more interested in how medication affects response to reading intervention. The graph below shows data from a single case seen in my clinic. The case is of a male with ADHD + mixed dyslexia. He was receiving reading intervention delivered 4/week. See here and here for descriptions of the program.


The data points represent a weekly nonword reading test. The test items were constructed using the grapheme-phoneme conversion rules taught in the intervention program. If the student learns ~4-5 new GPCs weekly they should be able to read ~5 extra nonwords each week. The “flat lines” represent data from 2 x baseline periods and 2 x treatment periods in which the participant was receiving reading intervention only. One can see that not much progress was being made. It wasn’t that the boy wasn’t learning new things. It was more than his “recall” was inconsistent and his test-taking behaviour was poor.

Upwards growth in test scores were seen almost immediately upon beginning to take a stimulant medication. Later data, not shown in the graph, showed that removing the reading intervention so the only treatment was medication resulted in a return to a flat line. That is, the medication didn’t teach skills.

We have seen this pattern in several cases now; although the results of other cases have not been as dramatic as this first case.  (I should also point out that the skeptic in me thinks that these data are too perfect and I need to see them replicated before I truly “believe” them).

Our very preliminary conclusions are that attention is necessary for new learning to take place and that the medication helps set the conditions for learning to happen (the batter into which the teaching is stirred if you will). However, medication will probably not make you a good reader sans the teaching.

At this point there is no justification for trialling medication in students who just have dyslexia. However, we hope to see a larger trial of the response of students with ADHD + dyslexia to reading intervention versus medication + reading intervention.

For help with dyslexia, ADHD, autism spectrum disorders and other developmental and learning disorders in the Gold Coast and Tweed regions contact the Understanding Minds Clinic.


Comments Off