Assessment & Research

Cognitive and familial risk evidence converged: A data-driven identification of distinct and homogeneous subtypes within the heterogeneous sample of reading disabled children.

Willems et al. (2016) · Research in developmental disabilities 2016
★ The Verdict

Reading-disabled kids split into four data-driven subtypes—use the profiles to pick precise, deficit-matched lessons.

✓ Read this if BCBAs doing academic assessments or writing reading goals for elementary students.
✗ Skip if Clinicians who only handle severe autism or exclusively preschool language cases.

01Research in Context

01

What this study did

Willems et al. (2016) fed lots of test scores and family history data into a computer. The kids already had reading disability diagnoses.

The computer looked for natural groupings. It found four clear clusters, each with its own pattern of cognitive strengths, weaknesses, and family risk.

02

What they found

One cluster showed strong family ties to dyslexia. The other three had different mixes of memory, language, or processing problems.

The result: reading-disabled children are not one big blob. They are at least four smaller, more uniform blobs that may need different teaching plans.

03

How this fits with other research

McIntyre et al. (2017) used the same math on students with high-functioning autism. They also found four reading profiles, but linked them to autism symptom severity instead of family history.

Gardner et al. (2009) and Sparaci et al. (2015) did similar cluster work inside autism and PDD-NOS. All studies prove one point: big labels hide small, teachable subgroups.

Costa et al. (2013) showed family history alone predicts reading failure. Gonny’s team folds that risk into the cluster recipe, giving you a dual lens: genes plus current skills.

04

Why it matters

Stop treating every poor reader the same. Run brief tests in phonology, memory, and family history. Match the pattern to one of the four clusters. Pick interventions that fit that cluster’s deficit, not the broad “dyslexia” tag. You could save months of trial-and-error lessons.

Free CEUs

Want CEUs on This Topic?

The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.

Join Free →
→ Action — try this Monday

Add a quick family dyslexia checklist to your intake packet and group learners by the four cognitive patterns before you write treatment targets.

02At a glance

Intervention
not applicable
Design
other
Population
other
Finding
not reported

03Original abstract

The evident degree of heterogeneity observed in reading disabled children has puzzled reading researchers for decades. Recent advances in the genetic underpinnings of reading disability have indicated that the heritable, familial risk for dyslexia is a major risk factor. The present data-driven, classification attempt aims to revisit the possibility of identifying distinct cognitive deficit profiles in a large sample of second to fourth grade reading disabled children. In this sample, we investigated whether genetic and environmental risk factors are able to distinguish between poor reader subtypes. In this profile, we included reading-related measures of phonemic awareness, letter-speech sound processing and rapid naming, known as candidate vulnerability markers associated with dyslexia and familial risk for dyslexia, as well as general cognitive abilities (non-verbal IQ and vocabulary). Clustering was based on a 200 multi-start K-means approach. Results revealed four emerging subtypes of which the first subtype showed no cognitive deficits underlying their poor reading skills (Reading-only impaired poor readers). The other three subtypes shared a core phonological deficit (PA) with a variable and discriminative expression across the other underlying vulnerability markers. More specific, type 2 showed low to poor performance across all reading-related and general cognitive abilities (general poor readers), type 3 showed a specific letter-speech sound mapping deficit next to a PA deficit (PA-LS specific poor readers) and type 4 showed a specific rapid naming deficit complementing their phonological weakness (PA-RAN specific poor readers). The first three poor reader profiles were more characterized by variable environmental risk factor, while the fourth, PA-RAN poor reader subtype showed a significantly strong familial risk for dyslexia. Overall, when we zoom in on the heterogeneous phenomenon of reading disability, unique and distinct cognitive subtypes can be identified, distinguishing between those poor readers more influences by the role of genes and those more influenced by environmental risk factors. Taking into account this diversity of distinct cognitive subtypes, instead of looking at the reading disabled sample as a whole, will help tailor future diagnostic and intervention efforts more specifically to the needs of children with such a specific deficit and risk pattern, as well as providing a more promising way forward for genetic studies of dyslexia.

Research in developmental disabilities, 2016 · doi:10.1016/j.ridd.2015.12.018