Ophthalmic Abnormalities and Reading Impairment: Alexandra L. Creavin, MBChBa, Raghu Lingam, MBChB, MRCPH, PhDb, Colin Steer, PhDa, Cathy Williams, MBBS, FRCOphth, PhDa Pediatrics, Vol 135,6, 2015
This study was based on a survey of nearly six thousand children taking part in the Bristol "Children of the Nineties" study. It defined severe reading impairment as more than two standard deviations below the mean on the Neale Analysis of Reading Ability (Second Revised British Edition) and achieving below the expected L4 in the UK's national SAT test for 11 year olds. 3 per cent of the sample met these criteria. A further 8 per cent were 1 standard deviation below the mean. Children with IQ below 70 and those who were blind were excluded. The study involved "secondary analysis of previously collected data".
The authors note in the first paragraph that
Specific learning disorder with reading impairment (dyslexia) affects 3% to 20% of school-aged children (∼375 000 UK children1) depending on the definition used.
The problem, however, is that they have no evidence that these weak readers were dyslexic. All they know is that they had low test scores and did not have severely impaired intelligence. They say that
The strong consensus of the scientific community is that reading problems, including dyslexia, are not caused by vision abnormalities.This consensus has not yet been tested in a large population cohort.
There are two problems here. The first is that, while we need to be able to see in order to read, no one I know thinks that reading problems are caused by vision abnormalities, at least as their primary cause. In the very early stages, though, the time at which vision problems are detected by normal eye testing may be an issue, and one that is not investigated here. Reading places new demands on the vision system of four and five year olds, and it may be some time before glasses are prescribed - it's difficult to carry out a sight test with people who have not yet learned the alphabet.
The second is that the authors can't identify the pupils with severe or moderate reading impairment as dyslexic, because they have not looked at individual cases at all. There can be many sources of reading problems, including limited early language development, interruptions to hearing in early childhood, and teaching that is not matched closely enough to learning needs. Neither have the researchers even looked at whether the children had been assessed for any benefit from tinting. This leaves them in the position of making critical comments on a phemonemon that they have not even investigated.
They note "a lack of robust epidemiological evidence to suggest that these (tinting) therapies are effective in improving outcomes for those with dyslexia". This is based on an editorial - not a research article - in the British Medical Journal. The editorial is a review of previous studies, all of which had flaws that might well have invalidated them individually. The biggest flaw in any investigation of tinting is that you can't carry out a blind trial on a tint. A partially sighted trial is not scientific, and the method of using two tints, the preferred one and one close enough to it that differences are not immediately obvious, does not test the benefit of tinting in itself. Screening tests very often show a benefit from several tints, but more from one than any other. So, the observed phenomenon of improvements in reading from tinting can't be investigated using the same means as, say, a new drug. The only solution is to return to case study evidence, based on direct observation, and this is as valid now as it was for Joseph Lister. Professor Arnold Wilkins' reply to the BMJ editorial, citing further case study evidence that is ignored for no reason, is here.
This article is very close to an abuse of academic authority. The BBC, to its credit, interviewed the lead author alongside a parent whose child's reading speed had more than doubled as a result of tinting. It is far too simple for people to put such direct evidence into a dustbin labelled "placebo" simply because they refuse to investigate it.
One finding from this study that confirms current experience is that children with severe reading impairment are a little more likely to have problems with their eyes working together - stereoscopy - than others. This reinforces the idea that children with reading problems need to have their eyes tested. An eye test will not necessarily reveal sensitivity to light, however, because it does not last long enough, and does not test the use of the eyes in the context of reading text. Visual stress is not a matter that can be investigated by established ophthalmic techniques, and it certainly can cause reading difficulties, by making the task very uncomfortable, causing headaches or worse. These factors are similar to those associated with dyslexia, though that is a topic in its own right. I know of no teachers or optometrists who see dyslexia and visual stress as the same thing, and it is wrong to suggest otherwise.