There is a moment when this article really matters, and it is usually eleven at night. You have the basket open on a page, with a donut-shaped pillow in it that promises to "prevent plagiocephaly", and on another tab the website of a centre that makes custom helmets. In between, your baby's head is flat on one side and the feeling is that every week that passes is a week lost.
Here are the two answers straight away, then I explain them with the numbers.
On the pillow: the evidence is thin, it does not concern your home cot, and the sleep safety recommendations say not to put soft objects in the cot. On the helmet: there is one serious randomised trial, and it found no differences compared with doing nothing.
This is not an article against those who sell these things. It is an article that places side by side what is promised and what has been measured, so that you can decide.
If instead you are looking for the overall picture (what flat head is, why it happens, how it is assessed), that is all in Flat head in the newborn: what the studies really say. Here we talk only about products.
First of all: how common it is, and how serious
You need the right scale, otherwise every product looks urgent.
In a Canadian cohort of 440 healthy babies born at term, assessed between 7 and 12 weeks, 46.6% had some form of plagiocephaly. Nearly one in two. And the detail that changes everything: 78.3% of them had a mild form.
Then there is the real fear, the one nobody says out loud: will he be left with something? One study followed 235 children with flat head or brachycephaly from infancy up to an average age of 9 years, measuring IQ, school performance and motor skills. The result: adding the severity of the deformation and the development measured at 7 and 18 months to the model did not meaningfully improve the prediction of how they would do at school. What mattered were family characteristics and early risk factors, not how flat the head was.
This does not mean ignoring the problem. It means you are deciding on a question of shape, with long timescales and wide margins, not on an emergency.
The anti-plagiocephaly pillow: what has actually been measured
When you search for "flat head pillow" you find dozens of products and very little research. The studies that exist are essentially three, and it is worth looking at each one, because none of the three describes the situation in which you would use it.
One. A German randomised study on 50 babies aged 5 months or younger with deformation already present compared for six weeks a bed pillow against stretching exercises. Both groups improved: on the skull asymmetry index, 2.09% with the exercises and 3.01% with the pillow in plagiocephaly; in mixed forms the pillow did somewhat better. The problem with this study is what it did not have: a group that did nothing. Part of the improvement comes anyway on its own (in the helmet trial, 23% of the babies left to natural course recovered completely), so a comparison between two treatments does not tell you how much of it comes from the pillow.
Two. An Indian randomised study tested a pillow filled with mustard seeds in 56 premature babies born at 32 weeks or less and under 1,500 grams, admitted to a neonatal intensive care unit. At four weeks the asymmetry was clearly lower in the pillow group (mean index 3.16 against 7.85). It is a real result, but look at the context: very small babies, under continuous monitoring, positioned by health staff. It is not a full-term baby sleeping alone in his cot.
Three. A Japanese study followed the use of the pillow after discharge up to six months of corrected age, finding less deformation and better cognitive scores. Except that there were 19 babies in total, eleven with the pillow and eight without, not assigned at random. With these numbers, and without randomisation, you generate a hypothesis: nothing is proven.
Summing up in one line: two studies on premature babies in a protected setting and one comparison without a control group. If a seller tells you it is "clinically proven", this is what lies behind it.
The part you will not find on the product page
Here the discussion changes nature, because we are no longer talking about effectiveness but about safety.
Paediatric safe sleep recommendations have said the same thing for years: the baby goes into the cot on his back on a firm mattress, and that is all. No pillows, no bumper pads, no duvets, no soft toys, no positioning devices.
This is not generic caution, it is numbers. In the population study conducted in the United States on 260 cot deaths and 260 matched controls, the last sleep environment told a lot: the use of a pillow was associated with a 2.5 times higher risk (95% confidence interval: 1.5-4.2), a soft surface with a 5.1 times higher risk (3.1-8.3), prone position with 2.4 times and the face or head covered by bedding with 2.5 times. In the model that took all the factors together, pillow and soft surface remained significant.
And regarding devices sold precisely to keep a baby in one position: a joint report by the CDC, FDA and Consumer Product Safety Commission examined 13 suffocation deaths over 13 years associated with their use. All the babies except one were 3 months old or younger; most had been put to sleep on their side, and many were found face down. The same report notes something that weighs heavily here: some of those devices had been authorised for reflux or plagiocephaly, but many others were freely marketed with promises of preventing cot death.
It needs to be said clearly what these data say and what they do not. They do not say that the pillow you are looking at will kill your child. They say that this category of objects has no evidence of safety in unsupervised sleep, that soft objects near the head are associated with a higher risk, and that you are weighing a small and uncertain benefit against a small but serious risk. And that a free alternative, which adds nothing to the cot, you already have.
If you really want a pillow of that kind, the only use that holds up is while awake and supervised, never for the night. Which, however, is exactly the moment when tummy time would do much more.
The helmet: what the trial says, and what everything else says
On the helmet the literature is broader, and it is split into two worlds that do not talk to each other.
The world of randomised studies. There is practically one: the HEADS trial, published in the BMJ in 2014. Eighty-four babies aged 5-6 months with moderate to severe deformation, born after the 36th week, without muscular torticollis or craniosynostosis, randomly assigned to six months of helmet or natural course, with a final blinded assessment at 24 months. The result: change in skull shape equal in the two groups, both for plagiocephaly (mean difference −0.2; 95% CI −1.6 to 1.2; p=0.80) and for brachycephaly (p=0.81), and full recovery in 26% with the helmet and 23% without. In the protocol the helmet had to be worn 23 hours a day from 6 to 12 months, and the side effects were reported by all the parents: skin irritation in 96%, difficulty hugging the baby in 77%, bad smell in 76%, sweating in 71%, pain in 33%. The total costs per baby: €1,401 in the helmet group against €157 in the group left to natural course. The authors conclude that the helmet should not be recommended as standard treatment for healthy babies.
The world of observational studies. Here the helmet almost always wins. A 2025 meta-analysis of 13 studies and 1,189 babies found a clearly greater reduction in asymmetry with the helmet, even compared with natural course alone. But the same authors put on the brakes: very high heterogeneity between studies (I² 90%), effects demonstrated in the short term, and the explicit request for further high-quality randomised studies before applying these results in practice. The guideline of the Congress of Neurological Surgeons says essentially the same thing: there is a substantial body of non-randomised evidence in favour of a more marked and faster improvement with the helmet, especially in severe forms and if applied early, but the criteria for deciding who and when remain uncertain.
And then there is the review which, in my view, explains why these two worlds do not meet. In 2024 a group went through 444 articles on the helmet, keeping 60: almost all retrospective case series, that is the lowest level of evidence. Randomised studies with blinded assessment: one. And the tools used to measure the deformation were so different from one another as to make it, literally, impossible to formulate generalisable recommendations.
So, honestly: the helmet is not a scam and it is not the first move. If someone proposes it to you at three months as the first thing, without having looked at the neck and without having given you a month of work on positions, that is the moment to ask for a second opinion. If instead we are talking about a severe deformation, in a baby already being followed, in which positional work has not shifted anything, it is a legitimate discussion to have with a specialist.
A review of 25 studies and 7,594 babies adds the practical piece: when the helmet is used, what predicts the result is the age at which it starts, the consistency in wearing it and the starting severity. It is worth knowing before you begin: if the helmet is worn little, one of the conditions for it to work is missing, while the side effects all remain.
What does work instead, and costs nothing
The unfair comparison is this: products are bought in thirty seconds, habits take weeks. But the numbers are on the side of habits.
In a Dutch randomised study, 380 babies were assessed at 7 weeks: 68 had a positional preference and the 65 enrolled were randomly assigned to a four-month paediatric physiotherapy programme or to usual care. In the treated group the risk of severe plagiocephaly fell by 46% at 6 months (relative risk 0.54; 95% CI 0.30-0.98) and by 57% at 12 months (0.43; 0.22-0.85). The number of babies to treat to avoid one severe case was around four (3.85 at 6 months, 3.13 at 12 months): very few, for an intervention made of positioning and movement, with nothing to buy. And the mechanism shows in the data: at 6 months the parents in the treated group fed, positioned and handled the baby more symmetrically.
Even a systematic review from 2025, which nonetheless recommends the helmet in moderate and severe forms, notes that for non-synostotic cranial asymmetry the best results come from physical therapy combined with work with the people who care for the baby.
In practice, the list is short and you already know it:
- tummy time while awake and supervised, a little at a time and many times a day: the full guide, including what to do if he cries, is in Tummy time: how to actually do it;
- alternating the side the head rests on when he sleeps, without ever changing the position on his back;
- alternating the arm you hold him with and the side you feed him on;
- turning the cot relative to the room, because babies turn towards the light and the door, always to the same side;
- reducing the hours in bouncers, car seats and seats outside of car journeys.
The question that comes before the product
There is a reason why so many parents buy the pillow and then nothing changes: they have not solved the why.
If your child always turns his head to the same side, very often it is not a habit but a limitation of neck movement. As long as that remains, you can change pillow every month: he will always go back to resting on the same spot, because it is the only one where he is comfortable. How much the neck matters is shown well by a Spanish randomised study on 34 babies with plagiocephaly under 28 weeks of age: adding manual work on the cervical spine to the programme for the parents increased active rotation by almost 30 degrees, against 6 degrees for the programme alone.
Honesty about this data, because it concerns us: the manual part is done with your hands on the baby, and at a distance it cannot be done. What can be done very well by video call is the other half: understanding which side he turns to, measuring how much, watching how you hold him and how you put him down, and giving you the right exercises to repeat every day. If the suspicion is a real torticollis, the whole issue is here: Congenital torticollis in the newborn: why timing matters more than anything.
The three questions to ask yourself before paying
- Has anyone looked at my child's neck? If the answer is no, any product is working on the effect and not on the cause.
- Have I really changed the habits for at least a month? Not "I tried": I mean alternating sides, reduced time in the bouncer, tummy time several times a day, every day.
- Is the person proposing this to me selling it? This is not an accusation, it is information you need to have. Always ask to see the full written quote, check-ups included.
When the product is not the answer and you need the doctor
Some things are not resolved either with a pillow or with a helmet, and they need to be seen by the paediatrician without waiting:
- a hard bony ridge along a line of the skull;
- a deformity that worsens instead of stabilising after 4-6 months;
- a head circumference that moves out of its growth curve;
- loss of skills the baby had already acquired;
- a marked facial asymmetry or an eye visibly lower than the other.
These are not cases for physiotherapy or for devices: they are cases for a medical diagnosis.
How we work
If you are here with the basket still open, the most useful thing you can do before buying is to have your baby looked at by someone who is not selling you anything.
With Ri-Hub Junior you start with a free twenty-minute introductory consultation by video call with a physiotherapist on the Italian professional register who works with children and teenagers: we look together at how the head turns, how you hold him and where he always rests, and we tell you whether what you are about to buy makes sense in your case or not. For children up to 10 years old sessions last 50 minutes and cost 59,90 €; from 11 years onwards the standard session is 29,90 €.
And if behind it there is a complex neurological condition, the right place is not us: those pathways need in-person assessments and specialised centres, and we tell you straight away rather than letting you lose weeks.
Sources
- Helmet therapy in infants with positional skull deformation: randomised controlled trial (HEADS): BMJ, 2014
- Helmet Therapy for Positional Plagiocephaly: A Systematic Review of the Tools Used to Diagnose, Offer Treatment Recommendations, and Assess Treatment Outcomes: Pediatric Neurology, 2024
- Helmet versus non-helmet treatment in infants with positional cranial deformation: A systematic review and meta-analysis: Journal of Clinical Neuroscience, 2025
- Use of orthotic helmets in children with positional plagiocephaly and brachycephaly: a systematic review: Child's Nervous System, 2025
- Congress of Neurological Surgeons Systematic Review and Evidence-Based Guideline on the Role of Cranial Molding Orthosis (Helmet) Therapy: Neurosurgery, 2016
- Cranial remolding orthosis for children with deformational skull deformities: factors affecting success and duration of treatment: World Neurosurgery X, 2024
- A prospective randomized trial on preventative methods for positional head deformity: physiotherapy versus a positioning pillow: Journal of Pediatrics, 2013
- Mustard Seed Pillow for Prevention of Deformational Plagiocephaly in ≤32 Weeks' Gestational Age Infants: An Open Label Randomized Trial: Indian Pediatrics, 2024
- Effects of continued positioning pillow use until a corrected age of six months on cranial deformation and neurodevelopment in preterm infants: Early Human Development, 2020
- Suffocation deaths associated with use of infant sleep positioners: United States, 1997-2011: MMWR, CDC, 2012
- Sleep environment and the risk of sudden infant death syndrome in an urban population: the Chicago Infant Mortality Study: Pediatrics, 2003
- Sudden infant death syndrome: American Family Physician, 2015
- Effect of pediatric physical therapy on deformational plagiocephaly in children with positional preference: a randomized controlled trial: Archives of Pediatrics & Adolescent Medicine, 2008
- Effectiveness of pediatric integrative manual therapy in cervical movement limitation in infants with positional plagiocephaly: a randomized controlled trial: Italian Journal of Pediatrics, 2021
- The incidence of positional plagiocephaly: a cohort study: Pediatrics, 2013
- Prediction of school-age IQ, academic achievement, and motor skills in children with positional plagiocephaly: Paediatrics & Child Health, 2021
Last source check: September 2026.
