A proprietary programme already resident in the settings that trust it
Sophie Brinsmead and Colin Woodcock, an educational psychologist at Hampshire Educational Psychology Service and an educational psychologist and academic tutor at the University of Southampton, have produced a critique and systematic review of Attention Autism — the intervention developed by the speech and language therapist Gina Davies, built around what its materials describe as an irresistible invitation to learn.
The programme runs as group sessions over six weeks, in four sequential stages, with the stated aim that children will end able to sustain attention for twenty minutes. Stage one presents a bucket of motivating toys the children watch an adult use but cannot touch. Stage two builds a sequence to an exciting finish — a sandcastle constructed and then crushed. Stage three uses interactive games in which only some children take a turn, developing shifting attention. Stage four moves the child from adult-led to independent activity.
It is proprietary. Delivering it requires attending training from a licensed trainer, at around £295, within a five-level structure, and advanced practitioner status requires nine days of training plus one-to-one coaching. The authors note there are no published figures on how widely it is used, but practitioner reports and its frequency in early years settings in their own local authorities suggest it has been adopted extensively. They attempted to contact licensed trainers, and Davies herself, for additional materials, and were unsuccessful.
The context makes the question urgent rather than academic. Recorded autism as a special educational need rose by more than 130% across the four UK nations across the last decade, and autism is now the most commonly identified area of need among children with an Education, Health and Care Plan. A large number of settings are looking for something to do, they have a limited budget for training, and they are choosing between programmes largely on the basis of what appears to be evidence-based. Research on early intervention providers has found that the perceived evidence base for an intervention significantly predicts whether it gets used — meaning the field is running partly on reputation rather than data.
So the question the review asks is narrow and fair: how much evidence is there that Attention Autism improves joint attention in children up to eleven?
Five studies, and the fidelity that was never there to be tested
The search returned 257 records. After duplicates and screening, twenty-nine full texts were read; exactly one met the inclusion criteria. Hand-searching reference lists and two further databases produced four more. The total evidence base for a programme in widespread use across UK early years settings is five studies.
Of those five, two were published in peer-reviewed journals and three are unpublished master’s or doctoral theses. One of the peer-reviewed studies was co-authored by Gina Davies, the programme’s creator; another study lists her in the acknowledgements. The authors appraised all five with a quality tool scored out of 48. The results were 26, 22, 16, 8 and 8. Only one study included a control group. Only one analysed its results statistically. One assessed the programme through 98 parental questionnaires, of which 98% reported positive changes — a genuine finding about parent perception, and not a measure of joint attention.
Then there is the fidelity problem, which is the heart of it. Attention Autism must be delivered by someone trained through the official pathway. In two studies, the researcher — not an advanced practitioner — trained the school staff. Two others were delivered by multidisciplinary teams including clinical psychologists, occupational therapists and speech and language therapists, a staffing model no ordinary school could replicate. Only one study described adults receiving formal training, and even there it was unclear whether this was the official training.
Delivery varied as much as staffing. The programme specifies four stages across six weeks building to twenty minutes; three of the five studies delivered only the first three stages, none documented reaching twenty minutes, and none reported how long was spent on each stage. In one study the intervention was compressed into five hours across five consecutive days, interspersed with forest school, messy play, lunch and snack breaks — which, as the reviewers point out, leaves open whether any gains came from the forest school, an activity independently associated with improved concentration.
The cumulative position is not that Attention Autism has been tested and failed. It is that it has never actually been tested. Every study reported positive results; every study also departed from the protocol, most substantially. When the intervention delivered is not the intervention as designed, a positive result cannot be attributed to the design — and the reviewers are appropriately careful to say that further, more rigorous research is needed rather than that the programme does not work. That distinction matters, and anyone citing this review to dismiss the programme outright would be overreading it.
The ingredient that survived the failure to deliver the protocol
But the reviewers make one observation that is sharper than their formal conclusion, and it is the most useful thing in the paper. Improvements in joint attention were reported despite low fidelity — across studies that trained staff differently, ran different numbers of stages, over different durations, with different professionals. Something was working that did not depend on the protocol being followed.
They point at what it might be. Novel objects promote joint attention on their own. Small-scale research has found that objects which genuinely interest a child promote joint attention. And an object of interest to the child is the one component present in every study regardless of how far the delivery drifted. The reviewers put the implication carefully: it may be this, rather than the intervention as a whole, that produced the reported gains.
If that is right, the active ingredient is not proprietary. It is not licensed, it does not cost £295, and it does not require nine days of training to access. It is the principle that a child’s attention follows their interest — which is monotropism, the autistic-led account of attention as deep, interest-gated and singular. Attention Autism can be read as an operationalisation of that principle, packaged and sold. And a school that cannot afford the training can still act on the principle tomorrow, because the principle is free.
One bucket, two incompatible purposes — compliance or company
There is a fork here that the review surfaces and that matters more than the effectiveness question, because it determines what you are actually doing to a child.
The reviewers note that the programme’s theoretical underpinnings are not explicitly stated, and then set out the candidates. One is social motivation theory: the proposal that social interaction is less intrinsically motivating for some autistic children, so a motivating object is paired with social interaction to raise the reward value — and they are explicit that this is in line with behaviourist principles of learning, citing the Koegel lineage. The other is monotropism, which frames the same activity as basing the work on a child’s strengths and interests, meeting them inside their own attention rather than recruiting it toward compliance.
The same bucket serves both. In the first reading, the object is a lever: interest is instrumentalised to make a social demand more palatable, and success is the child attending to the adult. In the second, the object is a meeting point: the adult enters the child’s interest, and success is shared attention that the child actually wants. Identical materials, opposite purposes — and the child can tell the difference even when the observer cannot.
This is why the fidelity finding and the ingredient finding matter beyond methodology. If the effective component is the child’s genuine interest, then the monotropic reading is the one the evidence points toward, and the behaviourist scaffolding around it is doing less work than the training implies. And if a setting is running the programme in the compliance register — measuring success by whether the child sat still and attended when directed — it is running something behaviourist with a bucket, and the twenty-minute target becomes a compliance metric rather than a marker of engagement. The test is straightforward and worth applying honestly: is the adult following the child’s attention, or capturing it?
None of this means throwing the programme out. Moore’s study, the strongest of the five, used a control group and appropriate statistics and did find significant gains in both initiating and responding to joint attention. It means being honest that a widely-adopted licensed programme rests on five studies, three unpublished, none delivered as designed, and that the part most likely to be doing the work is a principle any adult can apply without paying for it.
Citations
Brinsmead, S. & Woodcock, C. (2026) — Attention Autism: An academic critique — pre-publication version, University of Southampton / Hampshire Educational Psychology Service
Murray, D., Lesser, M. & Lawson, W. (2005) — Attention, monotropism and the diagnostic criteria for autism — Autism
Chevallier, C., Kohls, G., Troiani, V. et al. (2012) — The social motivation theory of autism — Trends in Cognitive Sciences
Paynter, J., Luskin-Saxby, S., Keen, D. et al. (2020) — Brief report: Perceived evidence and use of autism intervention strategies in early intervention providers — Journal of Autism and Developmental Disorders
