1
Treating reading disability
without reading: evaluating
alternative intervention
approaches
Dorothy Bishop
University of Oxford
http://www.psy.ox.ac.uk/oscci/
TALK AT BRITISH DYSLEXIA ASSOCIATION, MARCH 2008
2
Conventional approaches to
dyslexia
• Most children have problems with
“phonological awareness”
• Interventions focus on training children to
identify sounds in words and relate these
to letters
3
Problems with conventional
approach
• Nicolson et al 2003 summarising previous
work on phonological interventions:
– Intensive and prolonged phonological
intervention can be effective in improving
reading accuracy
– Reading fluency remains a problem
– Methods that are effective for most children,
don’t work for all: A hard core of children
remain very hard to treat
4
Neuroscience studies of
developmental disorders
The Holy Grail
Develop a theory of the disorder that
not only explains why it occurs, but also
motivates effective intervention
5
Goals of this talk
• Identify some questions for parents
considering a new treatments
• Illustrate with Dore method and fish oil
6
Questions to ask of a new
treatment
1. Is the theory scientifically plausible?
2. What does the intervention claim to improve?
3. Are costs reasonable relative to benefits?
4. Does evidence for efficacy goes beyond
testimonials? – have studies been done with
groups for whom treatment is recommended?
5. Is there evidence that gains are due to
treatment rather than maturation, practice,
placebo, etc.?
7
Dore method: what is it?
• Method for curing
cerebellar problems
developed by Wynford
Dore to help his
dyslexic daughter
• Individualised program
of exercises, done for
around 10 mins, 2 x
per day, assessed
every 6 weeks
8
Dore method: the exercises
• Hundreds of exercises, e.g.:
– standing on a cushion on one leg and
throwing a beanbag from one hand to
another for one minute
– hopping on one leg in large circle,
clockwise then anticlockwise
– sitting upright in a chair, turning head from
side to side, pausing to focus on chosen
point
– balancing on a wobble board
(Examples only: full details confidential because
commercially sensitive)
9
Dore method: the theory
•Dyslexia and other
learning difficulties arise
when the cerebellum
fails to develop normally
•Cerebellar impairments
differ from person to
person but can be
diagnosed by specific
tests of mental and
physical co-ordination
10
Cerebellar theory of dyslexia
• Not proven, but some evidence for it
• Brain imaging and neuroanatomical
studies offer some support
• Theory that dyslexia involves failure to
automatise skills is plausible
• Associated deficits in motor co-ordination
in a subset of people with dyslexia
11
Previous research on
effectiveness of motor training
• Training can improve
performance on motor
tasks, e.g. juggling
• In rats, exercise can
reverse cerebellar deficits
caused by prenatal alcohol
or zero gravity
• But no evidence that motor
training enhances
development of non-motor
skills
12
The Theory: evaluation
• Notion that training motor skills will have
effect on other skills:
• “This hypothesis required something of a leap of
faith, in that it is generally believed that the
cerebellum comprises a very large number of
independent ‘cerebro-cortical microzones’, and
so it is not clear why training on one sort of task
should generalize to unrelated tasks”
(Reynolds et al, 2003, p 53)
Reynolds, D., Nicolson, R. I., & Hambly, H. (2003). Evaluation of an exercise-based
treatment for children with reading difficulties. Dyslexia, 9(1), 48–71.
https://doi.org/10.1002/dys.235
13
The Theory: evaluation
• If training focusing on one region of cerebellum
had general effects on all cerebellar functions,
then
– activities like juggling and skateboarding should
benefit everyone
– sportsmen and women should have low risk of
dyslexia
Paul Nixon
Kenny Logan
Duncan Goodhew
14
Questions
1. Is the theory scientifically plausible?
• Notion that cerebellum may be implicated
in dyslexia is plausible though not proven
• Notion that motor exercises will have
beneficial effect on regions of cerebellum
concerned with learning is considerably
less plausible
15
Q2: What does the intervention
claim to improve?
Dore intervention recommended for:
• Dyslexia
• Dyspraxia
• Attention Deficit Hyperactivity Disorder
• Asperger syndrome
16
Dore (2006): Symptoms which
respond dramatically to treatment
– poor handwriting
– low self-esteem
– poor spelling
– poor short-term
memory
– poor organisational
skills
– sporting skills
– poor concentration
– clumsiness
– frustration and anger
– phobias
– faddiness with food
– oversensitivity to touch
– poor social skills
– lack of empathy
– lack of eye contact
– poor posture
– travel sickness
– poor sleep patterns
Dore (2006), chapter 7
17
Costs in relation to benefits
Cost of treatment is around £1700-£2000:
“ Surely it is a price worth paying in the attempt to
transform the life of your child so that they are
able to enjoy school, to develop social skills, to
develop good sporting skills, to have good
prospects in life?”
Dore (2006) p. 171
“Money-back guarantee”
But only if “no physiological change” – i.e. child
who improves on balance/eye tracking won’t get
refund, even if dyslexia/ADHD etc unchanged
18
Question 4
Have studies been done with groups for
whom treatment is recommended?
19
Published study on Dore
intervention
• Two papers in Dyslexia reporting different
phases
20
2003 study at Balsall Green school
• 296 children from 3 school yrs
• Selected 35 “at risk” on basis of Dyslexia
Screening Test : strong risk in 34%, mild in 21%
• Divided randomly into untreated and treated
groups
• Previous diagnoses:
– treated: 4 dyslexic, 1 dyspraxic
– control: 2 dyslexic, 1 dyspraxic, 1 ADHD
• Mean ‘reading age’ at start:
– -10.6 months in treated; -4.4 months in control
21
Q5. Is there evidence that gains
are due to treatment?
22
How to measure reading etc.
0
10
20
30
40
50
60
70
5 6 7 8 9 10 11 12 13 14 15
age (yr)
words
read
A 9-year-old reads 20 words
Reading Age is 8 years: Sounds bad: 12 months behind age level
23
Reading age misleading: does not take into
account spread of scores at a given age;
non-linear relation with chronological age
A 9-year-old reads 20 words. Error bars show middle 50% of children
Within normal range for children of this age
0
10
20
30
40
50
60
70
80
5 6 7 8 9 10 11 12 13 14 15
age (yr)
words
read
24
Better to measure reading in terms of
statistical abnormality at that age
0
10
20
30
40
50
60
70
80
5 6 7 8 9 10 11 12 13 14 15
age (yr)
words
read
Percentile: % of children of this age obtaining this score or lower;
Also z-score or standard score: different way of expressing same idea
25
Uninteresting reasons why
scores may improve - 1
• Maturation
–Children change with age
–Shoe size may go up after treatment, but
does not mean that treatment made feet
bigger
• A particular problem if “reading age”
used, but should not be an issue if
age-adjusted scores used
26
Uninteresting reasons why
scores may improve - 2
• Practice effects
• Child does test better 2nd time around
because they have done it before
• Numerous examples in research literature:
e.g. Oxford-Durham fish-oil study: motor
skills improved significantly in untreated as
well as treated children on retest
27
Uninteresting reasons why
scores may improve - 3
• Regression to the mean
–Statistical artefact whereby someone
selected for extreme score at time 1 will
on average have less extreme score at
time 2
28
Regression to the mean
-8
-6
-4
-2
0
2
4
6
8
1 2
test occasion
score
-4
-3
-2
-1
0
1
2
3
1 2
test occasion
average
score
Correlation between time 1 and time 2 = .11
29
Regression to the mean
Correlation between time 1 and time 2 = .99
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
1 2
test occasion
score
-4
-3
-2
-1
0
1
2
3
1 2
test occasion
average
score
30
Regression to the mean
Correlation between time 1 and time 2 = .76
-3
-2
-1
0
1
2
3
4
1 2
test occasion
score
-4
-3
-2
-1
0
1
2
3
1 2
test occasion
average
score
31
Uninteresting reasons why
scores may improve - 4
• Other intervention/placebo effect
–Child may be having other help or may
respond to increased attention
32
These unwanted sources of
change can be identified if we
have a CONTROL GROUP
• Untreated matched group given same pre-
and post-test will control for:
– Maturation
– Practice effects
– Regression to the mean
– Effects of other intervention
33
Alternative treatment control
group
• Crucial to see if improvement due to:
• Placebo/expectation effects
–Child, parent, teachers all expect and
want to see gains
–Child gets more attention, boosted
confidence, etc.
34
Need for controls
Parallels in physical medicine suggest
dangerous to assume you know
answer in advance – e.g. congestive
heart failure treatment
35
Results: total on dyslexia
screening
• High score indicates more risk
• NB score include bead-threading/posture
• Treated: mean fell from 0.74 to 0.34
– “strong risk” fell from 33% to 11%
• Control: mean fell from 0.72 to 0.44
– “strong risk” fell from 35% to 24%
Everyone gets better, even if not treated
36
DST results
Treated Group
0 2 4 6 8 10
Read
Spell
Nons pass reading
One min writing
RAN
Phon segment
Backward digits
Verbal Fl.
Semantic Fl
Bead thread
Post Stability
mean decile
time 2 time 1
usual cutoff for
dyslexia
*
*
*
*
*
37
Significant group differences in gain on
reading, bead threading, semantic fluency
Control group
0 2 4 6 8 10
Read
Spell
Nons pass reading
One min writing
RAN
Phon segment
Backward digits
Verbal Fl.
Semantic Fl
Bead thread
Post Stability
mean decile
time 2 time 1
Treated Group
0 2 4 6 8 10
Read
Spell
Nons pass reading
One min writing
RAN
Phon segment
Backward digits
Verbal Fl.
Semantic Fl
Bead thread
Post Stability
mean decile
time 2 time 1
38
Control group subsequently
given the treatment
• Results published in Dyslexia journal in
November 2006
• Control group now known as group D
(delayed intervention)
Reynolds, D., & Nicolson, R. I. (2007). Follow-up of an exercise-based
treatment for children with reading difficulties. Dyslexia, 13(2), 78–96.
https://doi.org/10.1002/dys.331
39
Group I
0 2 4 6 8 10
Read
Spell
Nons pass reading
One min writing
RAN
Phon segment
Backward digits
Verbal Fl.
Semantic Fl
Bead thread
Post Stability
mean decile
time 1
time 4
GroupD
0 2 4 6 8 10
Read
Spell
Nons pass reading
One min writing
RAN
Phon segment
Backward digits
Verbal Fl.
Semantic Fl
Bead thread
Post Stability
mean decile
time 1
time 4
Note:
lack of “stunning” progress on literacy tests
Results on dyslexia screening test, time 1 and time 4
40
Interim conclusion
• Dyslexia screening test data
– Little evidence of improvement on literacy
tests associated with intervention
• How, then can Dore (2006) conclude that
the results were “stunning” and that:
– comprehension age: increased almost 5 x
– reading age, increased 3 x
– writing, increased by “an extraordinary 17 x”
41
Data from school-administered
tests, treated group only
NFER reading test
84
96
108
120
132
144
-15 -3 9 21
time relative to grp I start
mean
reading
age
(mo)
DDAT groups
mean
national average
% improvement
calculated by dividing
pink line by orange
line
Conclude “reading age
increased 3 times”
Meaningless without
control data!!!
42
Data from school-administered
tests, treated group only
NFER reading test
84
96
108
120
132
144
-15 -3 9 21
time relative to grp I start
mean
reading
age
(mo)
DDAT groups
mean
national average
% improvement
calculated by dividing
pink line by orange
line
Conclude “reading age
increased 3 times”
Meaningless without
control data!!!
Goes on to do similar
calculations on SATS
scores, which are based
on teacher ratings
43
Q5. Is there evidence that gains
are due to treatment?
• Improvement looks dramatic only on
measures where there is no control data
• On reading measures where control group
available, no evidence of sustained gains
in the treated group
44
Fish oil
45
The theory
• Certain highly unsaturated fatty acids
(HUFAs) important in brain development
and neuronal signal transduction
• Could be problematic if:
– Dietary insufficiency
– Genetic abnormality in phospholipid
metabolism
• Administration of HUFAs may improve
synaptic transmission
46
Evidence of abnormal fatty acid
levels in dyslexia
• Clinical signs of fatty acid deficiency found
in adults with dyslexia (Taylor et al, 2000)
• Clinical signs of FAD correlate with
severity of dyslexia in children (Richardson
et al, 2000)
• Taylor, K. E., Higgins, C. J., Calvin, C. M., Hall, J. A., Easton, T., McDaid, A. M.,
& Richardson, A. J. (2000). Dyslexia in adults is associated with clinical signs of
fatty acid deficiency. Prostaglandins, Leukotrienes and Essential Fatty Acids
(PLEFA), 63(1), 75–78. https://doi.org/10.1054/plef.2000.0195
• Richardson, A. J., Calvin, C. M., Clisby, C., Schoenheimer, D. R., Montgomery,
P., Hall, J. A., Hebb, G., Westwood, E., Talcott, J. B., & Stein, J. F. (2000). Fatty
acid deficiency signs predict the severity of reading and related difficulties in
dyslexic children. Prostaglandins, Leukotrienes, and Essential Fatty Acids,
63(1–2), 69–74. https://doi.org/10.1054/plef.2000.0194
47
Evidence from treatment trials
• Significant reduction in ADHD symptoms in
children with comorbid dyslexia/ADHD cf.
placebo (Richardson & Puri, 2002)
– reading not assessed (!!??)
• Cf. no improvement of ADHD symptoms vs.
placebo in 2 studies of ADHD, though studies
vary in fatty acid, sample, etc
– Hirayama et al. 2004; Voigt et al., 2001
• Richardson, A. J., & Puri, B. K. (2002). Progress in Neuro-Psychopharmacology &
Biological Psychiatry, 26(2), 233–239. https://doi.org/10.1016/s0278-
5846(01)00254-8
• Hirayama, S., Hamazaki, T., & Terasawa, K. (2004).. European Journal of Clinical
Nutrition, 58(3), Article 3. https://doi.org/10.1038/sj.ejcn.1601830
• Voigt, R. G., Llorente, A. M., Jensen, C. L., Fraley, J. K., Berretta, M. C., & Heird, W.
C. (2001) The Journal of Pediatrics, 139(2), 189–196.
https://doi.org/10.1067/mpd.2001.116050
48
Evidence from treatment trials
• Oxford-Durham study on children with
developmental coordination disorder; Treated
show significantly more improvement in literacy
and ADHD symptoms: (Richardson &
Montgomery, 2005)
• Requests to see data to identify children with
dyslexia from this sample get no response
Richardson, A. J., & Montgomery, P. (2005). The Oxford-Durham study: A
randomized, controlled trial of dietary supplementation with fatty acids in children
with developmental coordination disorder. Pediatrics, 115(5), 1360–1366.
https://doi.org/10.1542/peds.2004-2164
49
Controlled trial of fish oil in
dyslexic adults
• Cyhlarova et al, 2007 report baseline
results - no differences in membrane fatty
acid levels between dyslexic and control
adults, though ratio of types of fatty acid
differs
• Requests for information on progress of
this treatment trial get no response
Cyhlarova, E., Bell, J. G., Dick, J. R., MacKinlay, E. E., Stein, J. F., & Richardson,
A. J. (2007). Membrane fatty acids, reading and spelling in dyslexic and non-
dyslexic adults. European Neuropsychopharmacology, 17(2), 116–121.
https://doi.org/10.1016/j.euroneuro.2006.07.003
50
Q1. Is the theory scientifically
plausible?
• Membrane phospholipid deficiency:
speculative theory developed to account
for schizophrenia, extended to
neurodevelopmental disorders
• Most plausible when applied to children
who show physical symptoms suggestive
of essential fatty acid deficiency
51
Q2. What does the intervention
claim to improve?
• Very generally applied to
neurodevelopmental disorders, including
ADHD, dyslexia, dyspraxia and autism
52
Q3. Are costs reasonable?
• MorEPA £19.50 for 60 capsules (1 per
day)
• Treatment may need to be ‘long term’
• £118 per year
53
Q4. Does evidence for efficacy
goes beyond testimonials? –
have studies been done with
groups for whom treatment is
recommended?
• Several clinical trials but only one
specifically on children with dyslexia (and
ADHD) - did not look at reading outcomes
• Study of developmental coordination
disorder included measures of reading as
part of outcome assessment
54
Q5. Is there evidence that gains
are due to treatment?
• Inclusion of control group makes it
possible to take into account practice,
maturation, etc.
• However, some effects likely to be chance
when many comparisons done; need for
replication in new sample using better
measures of reading outcomes
55
Barriers to objective evaluation
• Failure to recognise important effects of :
– expectations
– practice
– maturation
– statistical artefact
• Human tendency to be impressed by
testimonials
56
Testimonials more compelling
than statistics
• E.g. quote from a parent, re. MMR vaccine and
autism:
“The only evidence which I’ve seen cited is
epidemiological. Now to me that’s statistics. Now
statistics either turn you on or they don’t. I don’t have
time for statistics when I see a sick child in front of me.”
• Response from “guardian badscience” blog:
“You should be thankful that somebody has had time for
statistics, or else you’d be treating that sick child by
bleeding her.”
57
N.B. Testimonials problematic because
(a) selective
(b) often at odds with objective evaluation
58
Barriers to objective evaluation
• Failure to recognise important effects of :
– expectations
– practice
– maturation
– statistical artefacts
• Human tendency to be impressed by
testimonials
• Human tendency to think something that has
taken time/effort/money was worthwhile
59
Trial of Sunflower therapy
Bull (2007)
• Both treated and control children improved
their test scores over time to similar extent
• Higher academic self-esteem in those
undergoing treatment
• 57% of parents thought Sunflower therapy
was effective in treating learning difficulties
Bull, L. (2007). Sunflower therapy for children with specific learning difficulties
(dyslexia): A randomised, controlled trial. Complementary Therapies in Clinical
Practice, 13(1), 15–24. https://doi.org/10.1016/j.ctcp.2006.07.003
60
Barriers to objective evaluation
• Failure to recognise important effects of :
– expectations
– practice
– maturation
– statistical artefacts
• Human tendency to be impressed by
testimonials
• Human tendency to think something that has
taken time/effort/money was worthwhile
• Human tendency to be impressed by
neuroscientific explanations
61
“The seductive allure of
neuroscience explanations”
Weisberg et al. 2008. J. Cognitive Neuroscience 20: 470-7
mean
rating of
satisfaction
with
explanation

Exercise/fish oil intervention for dyslexia

  • 1.
    1 Treating reading disability withoutreading: evaluating alternative intervention approaches Dorothy Bishop University of Oxford http://www.psy.ox.ac.uk/oscci/ TALK AT BRITISH DYSLEXIA ASSOCIATION, MARCH 2008
  • 2.
    2 Conventional approaches to dyslexia •Most children have problems with “phonological awareness” • Interventions focus on training children to identify sounds in words and relate these to letters
  • 3.
    3 Problems with conventional approach •Nicolson et al 2003 summarising previous work on phonological interventions: – Intensive and prolonged phonological intervention can be effective in improving reading accuracy – Reading fluency remains a problem – Methods that are effective for most children, don’t work for all: A hard core of children remain very hard to treat
  • 4.
    4 Neuroscience studies of developmentaldisorders The Holy Grail Develop a theory of the disorder that not only explains why it occurs, but also motivates effective intervention
  • 5.
    5 Goals of thistalk • Identify some questions for parents considering a new treatments • Illustrate with Dore method and fish oil
  • 6.
    6 Questions to askof a new treatment 1. Is the theory scientifically plausible? 2. What does the intervention claim to improve? 3. Are costs reasonable relative to benefits? 4. Does evidence for efficacy goes beyond testimonials? – have studies been done with groups for whom treatment is recommended? 5. Is there evidence that gains are due to treatment rather than maturation, practice, placebo, etc.?
  • 7.
    7 Dore method: whatis it? • Method for curing cerebellar problems developed by Wynford Dore to help his dyslexic daughter • Individualised program of exercises, done for around 10 mins, 2 x per day, assessed every 6 weeks
  • 8.
    8 Dore method: theexercises • Hundreds of exercises, e.g.: – standing on a cushion on one leg and throwing a beanbag from one hand to another for one minute – hopping on one leg in large circle, clockwise then anticlockwise – sitting upright in a chair, turning head from side to side, pausing to focus on chosen point – balancing on a wobble board (Examples only: full details confidential because commercially sensitive)
  • 9.
    9 Dore method: thetheory •Dyslexia and other learning difficulties arise when the cerebellum fails to develop normally •Cerebellar impairments differ from person to person but can be diagnosed by specific tests of mental and physical co-ordination
  • 10.
    10 Cerebellar theory ofdyslexia • Not proven, but some evidence for it • Brain imaging and neuroanatomical studies offer some support • Theory that dyslexia involves failure to automatise skills is plausible • Associated deficits in motor co-ordination in a subset of people with dyslexia
  • 11.
    11 Previous research on effectivenessof motor training • Training can improve performance on motor tasks, e.g. juggling • In rats, exercise can reverse cerebellar deficits caused by prenatal alcohol or zero gravity • But no evidence that motor training enhances development of non-motor skills
  • 12.
    12 The Theory: evaluation •Notion that training motor skills will have effect on other skills: • “This hypothesis required something of a leap of faith, in that it is generally believed that the cerebellum comprises a very large number of independent ‘cerebro-cortical microzones’, and so it is not clear why training on one sort of task should generalize to unrelated tasks” (Reynolds et al, 2003, p 53) Reynolds, D., Nicolson, R. I., & Hambly, H. (2003). Evaluation of an exercise-based treatment for children with reading difficulties. Dyslexia, 9(1), 48–71. https://doi.org/10.1002/dys.235
  • 13.
    13 The Theory: evaluation •If training focusing on one region of cerebellum had general effects on all cerebellar functions, then – activities like juggling and skateboarding should benefit everyone – sportsmen and women should have low risk of dyslexia Paul Nixon Kenny Logan Duncan Goodhew
  • 14.
    14 Questions 1. Is thetheory scientifically plausible? • Notion that cerebellum may be implicated in dyslexia is plausible though not proven • Notion that motor exercises will have beneficial effect on regions of cerebellum concerned with learning is considerably less plausible
  • 15.
    15 Q2: What doesthe intervention claim to improve? Dore intervention recommended for: • Dyslexia • Dyspraxia • Attention Deficit Hyperactivity Disorder • Asperger syndrome
  • 16.
    16 Dore (2006): Symptomswhich respond dramatically to treatment – poor handwriting – low self-esteem – poor spelling – poor short-term memory – poor organisational skills – sporting skills – poor concentration – clumsiness – frustration and anger – phobias – faddiness with food – oversensitivity to touch – poor social skills – lack of empathy – lack of eye contact – poor posture – travel sickness – poor sleep patterns Dore (2006), chapter 7
  • 17.
    17 Costs in relationto benefits Cost of treatment is around £1700-£2000: “ Surely it is a price worth paying in the attempt to transform the life of your child so that they are able to enjoy school, to develop social skills, to develop good sporting skills, to have good prospects in life?” Dore (2006) p. 171 “Money-back guarantee” But only if “no physiological change” – i.e. child who improves on balance/eye tracking won’t get refund, even if dyslexia/ADHD etc unchanged
  • 18.
    18 Question 4 Have studiesbeen done with groups for whom treatment is recommended?
  • 19.
    19 Published study onDore intervention • Two papers in Dyslexia reporting different phases
  • 20.
    20 2003 study atBalsall Green school • 296 children from 3 school yrs • Selected 35 “at risk” on basis of Dyslexia Screening Test : strong risk in 34%, mild in 21% • Divided randomly into untreated and treated groups • Previous diagnoses: – treated: 4 dyslexic, 1 dyspraxic – control: 2 dyslexic, 1 dyspraxic, 1 ADHD • Mean ‘reading age’ at start: – -10.6 months in treated; -4.4 months in control
  • 21.
    21 Q5. Is thereevidence that gains are due to treatment?
  • 22.
    22 How to measurereading etc. 0 10 20 30 40 50 60 70 5 6 7 8 9 10 11 12 13 14 15 age (yr) words read A 9-year-old reads 20 words Reading Age is 8 years: Sounds bad: 12 months behind age level
  • 23.
    23 Reading age misleading:does not take into account spread of scores at a given age; non-linear relation with chronological age A 9-year-old reads 20 words. Error bars show middle 50% of children Within normal range for children of this age 0 10 20 30 40 50 60 70 80 5 6 7 8 9 10 11 12 13 14 15 age (yr) words read
  • 24.
    24 Better to measurereading in terms of statistical abnormality at that age 0 10 20 30 40 50 60 70 80 5 6 7 8 9 10 11 12 13 14 15 age (yr) words read Percentile: % of children of this age obtaining this score or lower; Also z-score or standard score: different way of expressing same idea
  • 25.
    25 Uninteresting reasons why scoresmay improve - 1 • Maturation –Children change with age –Shoe size may go up after treatment, but does not mean that treatment made feet bigger • A particular problem if “reading age” used, but should not be an issue if age-adjusted scores used
  • 26.
    26 Uninteresting reasons why scoresmay improve - 2 • Practice effects • Child does test better 2nd time around because they have done it before • Numerous examples in research literature: e.g. Oxford-Durham fish-oil study: motor skills improved significantly in untreated as well as treated children on retest
  • 27.
    27 Uninteresting reasons why scoresmay improve - 3 • Regression to the mean –Statistical artefact whereby someone selected for extreme score at time 1 will on average have less extreme score at time 2
  • 28.
    28 Regression to themean -8 -6 -4 -2 0 2 4 6 8 1 2 test occasion score -4 -3 -2 -1 0 1 2 3 1 2 test occasion average score Correlation between time 1 and time 2 = .11
  • 29.
    29 Regression to themean Correlation between time 1 and time 2 = .99 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 1 2 test occasion score -4 -3 -2 -1 0 1 2 3 1 2 test occasion average score
  • 30.
    30 Regression to themean Correlation between time 1 and time 2 = .76 -3 -2 -1 0 1 2 3 4 1 2 test occasion score -4 -3 -2 -1 0 1 2 3 1 2 test occasion average score
  • 31.
    31 Uninteresting reasons why scoresmay improve - 4 • Other intervention/placebo effect –Child may be having other help or may respond to increased attention
  • 32.
    32 These unwanted sourcesof change can be identified if we have a CONTROL GROUP • Untreated matched group given same pre- and post-test will control for: – Maturation – Practice effects – Regression to the mean – Effects of other intervention
  • 33.
    33 Alternative treatment control group •Crucial to see if improvement due to: • Placebo/expectation effects –Child, parent, teachers all expect and want to see gains –Child gets more attention, boosted confidence, etc.
  • 34.
    34 Need for controls Parallelsin physical medicine suggest dangerous to assume you know answer in advance – e.g. congestive heart failure treatment
  • 35.
    35 Results: total ondyslexia screening • High score indicates more risk • NB score include bead-threading/posture • Treated: mean fell from 0.74 to 0.34 – “strong risk” fell from 33% to 11% • Control: mean fell from 0.72 to 0.44 – “strong risk” fell from 35% to 24% Everyone gets better, even if not treated
  • 36.
    36 DST results Treated Group 02 4 6 8 10 Read Spell Nons pass reading One min writing RAN Phon segment Backward digits Verbal Fl. Semantic Fl Bead thread Post Stability mean decile time 2 time 1 usual cutoff for dyslexia * * * * *
  • 37.
    37 Significant group differencesin gain on reading, bead threading, semantic fluency Control group 0 2 4 6 8 10 Read Spell Nons pass reading One min writing RAN Phon segment Backward digits Verbal Fl. Semantic Fl Bead thread Post Stability mean decile time 2 time 1 Treated Group 0 2 4 6 8 10 Read Spell Nons pass reading One min writing RAN Phon segment Backward digits Verbal Fl. Semantic Fl Bead thread Post Stability mean decile time 2 time 1
  • 38.
    38 Control group subsequently giventhe treatment • Results published in Dyslexia journal in November 2006 • Control group now known as group D (delayed intervention) Reynolds, D., & Nicolson, R. I. (2007). Follow-up of an exercise-based treatment for children with reading difficulties. Dyslexia, 13(2), 78–96. https://doi.org/10.1002/dys.331
  • 39.
    39 Group I 0 24 6 8 10 Read Spell Nons pass reading One min writing RAN Phon segment Backward digits Verbal Fl. Semantic Fl Bead thread Post Stability mean decile time 1 time 4 GroupD 0 2 4 6 8 10 Read Spell Nons pass reading One min writing RAN Phon segment Backward digits Verbal Fl. Semantic Fl Bead thread Post Stability mean decile time 1 time 4 Note: lack of “stunning” progress on literacy tests Results on dyslexia screening test, time 1 and time 4
  • 40.
    40 Interim conclusion • Dyslexiascreening test data – Little evidence of improvement on literacy tests associated with intervention • How, then can Dore (2006) conclude that the results were “stunning” and that: – comprehension age: increased almost 5 x – reading age, increased 3 x – writing, increased by “an extraordinary 17 x”
  • 41.
    41 Data from school-administered tests,treated group only NFER reading test 84 96 108 120 132 144 -15 -3 9 21 time relative to grp I start mean reading age (mo) DDAT groups mean national average % improvement calculated by dividing pink line by orange line Conclude “reading age increased 3 times” Meaningless without control data!!!
  • 42.
    42 Data from school-administered tests,treated group only NFER reading test 84 96 108 120 132 144 -15 -3 9 21 time relative to grp I start mean reading age (mo) DDAT groups mean national average % improvement calculated by dividing pink line by orange line Conclude “reading age increased 3 times” Meaningless without control data!!! Goes on to do similar calculations on SATS scores, which are based on teacher ratings
  • 43.
    43 Q5. Is thereevidence that gains are due to treatment? • Improvement looks dramatic only on measures where there is no control data • On reading measures where control group available, no evidence of sustained gains in the treated group
  • 44.
  • 45.
    45 The theory • Certainhighly unsaturated fatty acids (HUFAs) important in brain development and neuronal signal transduction • Could be problematic if: – Dietary insufficiency – Genetic abnormality in phospholipid metabolism • Administration of HUFAs may improve synaptic transmission
  • 46.
    46 Evidence of abnormalfatty acid levels in dyslexia • Clinical signs of fatty acid deficiency found in adults with dyslexia (Taylor et al, 2000) • Clinical signs of FAD correlate with severity of dyslexia in children (Richardson et al, 2000) • Taylor, K. E., Higgins, C. J., Calvin, C. M., Hall, J. A., Easton, T., McDaid, A. M., & Richardson, A. J. (2000). Dyslexia in adults is associated with clinical signs of fatty acid deficiency. Prostaglandins, Leukotrienes and Essential Fatty Acids (PLEFA), 63(1), 75–78. https://doi.org/10.1054/plef.2000.0195 • Richardson, A. J., Calvin, C. M., Clisby, C., Schoenheimer, D. R., Montgomery, P., Hall, J. A., Hebb, G., Westwood, E., Talcott, J. B., & Stein, J. F. (2000). Fatty acid deficiency signs predict the severity of reading and related difficulties in dyslexic children. Prostaglandins, Leukotrienes, and Essential Fatty Acids, 63(1–2), 69–74. https://doi.org/10.1054/plef.2000.0194
  • 47.
    47 Evidence from treatmenttrials • Significant reduction in ADHD symptoms in children with comorbid dyslexia/ADHD cf. placebo (Richardson & Puri, 2002) – reading not assessed (!!??) • Cf. no improvement of ADHD symptoms vs. placebo in 2 studies of ADHD, though studies vary in fatty acid, sample, etc – Hirayama et al. 2004; Voigt et al., 2001 • Richardson, A. J., & Puri, B. K. (2002). Progress in Neuro-Psychopharmacology & Biological Psychiatry, 26(2), 233–239. https://doi.org/10.1016/s0278- 5846(01)00254-8 • Hirayama, S., Hamazaki, T., & Terasawa, K. (2004).. European Journal of Clinical Nutrition, 58(3), Article 3. https://doi.org/10.1038/sj.ejcn.1601830 • Voigt, R. G., Llorente, A. M., Jensen, C. L., Fraley, J. K., Berretta, M. C., & Heird, W. C. (2001) The Journal of Pediatrics, 139(2), 189–196. https://doi.org/10.1067/mpd.2001.116050
  • 48.
    48 Evidence from treatmenttrials • Oxford-Durham study on children with developmental coordination disorder; Treated show significantly more improvement in literacy and ADHD symptoms: (Richardson & Montgomery, 2005) • Requests to see data to identify children with dyslexia from this sample get no response Richardson, A. J., & Montgomery, P. (2005). The Oxford-Durham study: A randomized, controlled trial of dietary supplementation with fatty acids in children with developmental coordination disorder. Pediatrics, 115(5), 1360–1366. https://doi.org/10.1542/peds.2004-2164
  • 49.
    49 Controlled trial offish oil in dyslexic adults • Cyhlarova et al, 2007 report baseline results - no differences in membrane fatty acid levels between dyslexic and control adults, though ratio of types of fatty acid differs • Requests for information on progress of this treatment trial get no response Cyhlarova, E., Bell, J. G., Dick, J. R., MacKinlay, E. E., Stein, J. F., & Richardson, A. J. (2007). Membrane fatty acids, reading and spelling in dyslexic and non- dyslexic adults. European Neuropsychopharmacology, 17(2), 116–121. https://doi.org/10.1016/j.euroneuro.2006.07.003
  • 50.
    50 Q1. Is thetheory scientifically plausible? • Membrane phospholipid deficiency: speculative theory developed to account for schizophrenia, extended to neurodevelopmental disorders • Most plausible when applied to children who show physical symptoms suggestive of essential fatty acid deficiency
  • 51.
    51 Q2. What doesthe intervention claim to improve? • Very generally applied to neurodevelopmental disorders, including ADHD, dyslexia, dyspraxia and autism
  • 52.
    52 Q3. Are costsreasonable? • MorEPA £19.50 for 60 capsules (1 per day) • Treatment may need to be ‘long term’ • £118 per year
  • 53.
    53 Q4. Does evidencefor efficacy goes beyond testimonials? – have studies been done with groups for whom treatment is recommended? • Several clinical trials but only one specifically on children with dyslexia (and ADHD) - did not look at reading outcomes • Study of developmental coordination disorder included measures of reading as part of outcome assessment
  • 54.
    54 Q5. Is thereevidence that gains are due to treatment? • Inclusion of control group makes it possible to take into account practice, maturation, etc. • However, some effects likely to be chance when many comparisons done; need for replication in new sample using better measures of reading outcomes
  • 55.
    55 Barriers to objectiveevaluation • Failure to recognise important effects of : – expectations – practice – maturation – statistical artefact • Human tendency to be impressed by testimonials
  • 56.
    56 Testimonials more compelling thanstatistics • E.g. quote from a parent, re. MMR vaccine and autism: “The only evidence which I’ve seen cited is epidemiological. Now to me that’s statistics. Now statistics either turn you on or they don’t. I don’t have time for statistics when I see a sick child in front of me.” • Response from “guardian badscience” blog: “You should be thankful that somebody has had time for statistics, or else you’d be treating that sick child by bleeding her.”
  • 57.
    57 N.B. Testimonials problematicbecause (a) selective (b) often at odds with objective evaluation
  • 58.
    58 Barriers to objectiveevaluation • Failure to recognise important effects of : – expectations – practice – maturation – statistical artefacts • Human tendency to be impressed by testimonials • Human tendency to think something that has taken time/effort/money was worthwhile
  • 59.
    59 Trial of Sunflowertherapy Bull (2007) • Both treated and control children improved their test scores over time to similar extent • Higher academic self-esteem in those undergoing treatment • 57% of parents thought Sunflower therapy was effective in treating learning difficulties Bull, L. (2007). Sunflower therapy for children with specific learning difficulties (dyslexia): A randomised, controlled trial. Complementary Therapies in Clinical Practice, 13(1), 15–24. https://doi.org/10.1016/j.ctcp.2006.07.003
  • 60.
    60 Barriers to objectiveevaluation • Failure to recognise important effects of : – expectations – practice – maturation – statistical artefacts • Human tendency to be impressed by testimonials • Human tendency to think something that has taken time/effort/money was worthwhile • Human tendency to be impressed by neuroscientific explanations
  • 61.
    61 “The seductive allureof neuroscience explanations” Weisberg et al. 2008. J. Cognitive Neuroscience 20: 470-7 mean rating of satisfaction with explanation