What Sensory Integration Therapy Studies Actually Measured

What the research found

Sensory integration therapy rests on three senses the theory treats as foundational: the vestibular sense of balance and movement, the tactile sense of touch, and proprioception, which registers where the body is in space. A therapist arranges play that supplies controlled amounts of that input, usually one child at a time, on the premise that a child who organises the input more efficiently will find ordinary tasks less effortful. Across the Korean studies gathered in one recent review, a course typically ran twice a week in sessions of forty minutes, for somewhere between eight and sixty sessions in total.

That review, published in the World Journal of Clinical Cases in 2024 by Seri Oh, Narae Lee and colleagues at Kangwon National University and several other Korean institutions, searched three Korean-language databases for work appearing between 2001 and October 2020. Of 6,362 records, 24 studies met the criteria: each compared a treated group against a comparison group and reported enough data to pool. Three were randomised controlled trials. The remaining twenty-one assigned children to groups without randomisation, which leaves open the possibility that the groups differed before any therapy began.

Pooled across those studies, the effect on social skill was 1.22, and on adaptive behaviour, meaning the everyday self-care and routine-following that occupational therapists record, 1.15. Sensory processing reached 0.85. Gross motor skill reached 0.26 and fine motor skill 0.25. By the convention the authors themselves cite, 0.8 and above counts as a large effect and 0.2 or below as a small one, which places the therapy’s weakest results in the motor domains.

Sorted by diagnosis rather than by outcome, the pooled effect was 1.50 for cerebral palsy, 1.35 for autism and 1.06 for ADHD, falling to 0.48 for developmental disability and 0.14 for intellectual disability, where the confidence interval crossed zero and no effect can be claimed. Every number in the review is an average of studies that disagreed with one another by 92 to 97 percent. Those first three figures rest on three studies, two studies and two studies respectively. Forty-nine children in total contributed to the social skill estimate.

What it means for you and your child

In most of these studies the comparison group was simply not receiving sensory integration therapy, which answers a narrower question than the one families are asking. The authors say as much: where sensory integration has been compared against other interventions rather than against nothing, its reported advantage is relatively low. An effect measured this way cannot tell a parent whether the same hours spent on physiotherapy or unstructured play would have achieved as much.

The individual studies were small, with treated groups of between three and thirty-two children, and the review does not report whether the people scoring the outcomes knew which children had received the therapy. That matters most for the measures where the therapy looked strongest, since social skill and adaptive behaviour are rated by adults rather than counted by a machine. Small, unrandomised, unblinded studies do not produce worthless numbers, but they produce unstable ones, and the instability tends to inflate a result before it deflates one, which is why the largest figures here deserve the least weight.

Geography constrains the review further. Every study came from a Korean database and many were master’s or doctoral theses, a limitation the authors state before calling for randomised trials across several countries. They also note that no standard protocol for the therapy exists, which means the label covers a range of practice rather than a fixed procedure, and a clinic elsewhere may be doing something meaningfully different under the same name.

Sessions of forty minutes produced a pooled effect of 0.88 while sessions of thirty minutes produced roughly zero, but that comparison runs between studies rather than within them, so session length travelled alongside the therapist, the diagnosis and everything else that varied. It licenses no particular scheduling decision.

Its largest pooled effects concern social skill, adaptive behaviour and sensory processing, each drawn from a handful of small studies; its smallest concern motor coordination. A clinic able to name which of those it expects to change, and how that change will be measured across a defined number of sessions, is offering something a family can check against the research.

Sensory integration therapy has been studied many times and rigorously only rarely, and the largest numbers in this review rest on the smallest groups of children. That is a description of uncertainty rather than of failure. If your child is partway through a course of it, the accurate account is that nobody yet knows how much of the credit belongs to the therapy, and a clinic willing to say so is the more trustworthy one.