Honest Review Hub

How to Train Yourself to Sleep on Your Side

By Adeel Farooq, Founder & Editor · Researched by the Honest Review Hub Research Team · Evidence-checked · Last updated: 2026-11-21

Quick answer

You can train yourself to sleep on your side, and the trials show how: a physical prompt that makes rolling onto your back uncomfortable, used every night for weeks. In a randomised trial, a vibration device cut back-sleeping from 47% of the night to 17% in two months. The tennis ball sewn into a shirt works too, but people abandon it about twice as often. Side sleeping matters because breathing interruptions are roughly twice as frequent on the back as on the side — and for about half of people with mild apnea, side sleeping alone brings the count into the normal range. Expect weeks of practice, and expect to drift back sometimes.

Applies to: back sleepers who snore, anyone with positional sleep apnea, and stomach sleepers moving to their side.

Sleep position feels like something you are born with — you are a back sleeper, or a side sleeper, or a stomach sleeper, and that is that. The trials say otherwise. Sleep position is a habit, and habits change when the environment changes: a vibration on the chest when you roll over, a tennis ball in the back of a shirt, a body pillow that blocks the roll. Nobody decides their position at 3 a.m. — the setup decides it — so this page is about the setup, the trials behind it, and the honest numbers on how well it holds.

Why bother? Because position is the cheapest lever in sleep medicine. That is the prize: not a gadget, but a position that halves the problem for free, every night, if you can stay in it.

One warning before the plan: side sleeping has a cost, and this page states it where the training is prescribed rather than letting you discover it later. Side sleeping loads the shoulder, and shoulder pain from side sleeping is the commonest reason people abandon the position. If your shoulder hurts, the training is not failing — the setup is incomplete. Our shoulder pain from side sleeping guide covers the setup that protects it.

What works best? A chest-worn vibration device beats the tennis ball on every measure the trials report: in a head-to-head randomised trial, treatment success was 68.0% with the device against 42.9% with the tennis ball, and effective compliance was 75.9% against 42.3% (Eijsvogel et al., 2015). But the device costs money and the tennis ball costs nothing, and both prevented back sleeping to a median of 0% of the night. If you want the free method, use the tennis ball — and know that comfort is what decides whether you keep using it, so make the ball comfortable before you blame yourself for quitting.

A sleeper lying on their back with the airway narrowed, compared with side sleeping

On the back, gravity pulls the tongue and soft palate into the airway. On the side, they fall sideways instead. That is the whole mechanism — and the whole prize. Photo: illustrative.

The verdict: position can be trained

The verdict is two facts held together. First, side sleeping helps breathing a lot: a 2017 meta-analysis of seven studies of the new vibration devices found a mean reduction of 11.3 breathing interruptions an hour — a 54% reduction — and a 33.6 percentage-point reduction in the share of the night spent on the back, an 84% reduction, with large effect sizes on both (Ravesloot et al., 2017). Second, keeping the position is the hard part: in the longest trial, the discontinuation rate was 28.8% after two months and 49.4% after six months (Laub et al., 2017). The training works while it is used, and half the users stop using it within six months. A plan that ignores the second fact is a plan that fails in month three.

The definition matters for knowing whether the training is worth your time. Positional sleep apnea is commonly defined as breathing interruptions at least 50% lower off the back than on it — and by that definition, an average of 56% of people with sleep apnea have it (Ravesloot et al., 2013). The stricter definition — interruptions halved and under 5 an hour off the back — is met by 49.5% of people with mild apnea, 19.4% with moderate, and 6.5% with severe, in a study of patients at two US sleep centres (Mador et al., 2005). The milder the apnea, the more likely side sleeping alone fixes it; the more severe, the more likely you need more than position. Your sleep study report usually lists the count separately for back and non-back sleep — if the back number is much higher, you are the person this training is for.

Side sleeping is also the best position for most adults on the spine evidence, with a qualification. A 2019 scoping review of four studies found side-lying generally protective against waking spinal pain and stiffness, while noting there were not enough high-quality studies for firm recommendations (Cary et al., 2019). A 2025 systematic review of six observational studies found side-lying the most common posture, helping when alignment is supported and hurting when it is not (Saini et al., 2025). The qualification is the setup: side sleeping helps when the pillow fills the shoulder-to-ear gap and the spine stays straight, and hurts when it does not. That is why this page spends as long on the setup as on the training.

What the position trials prove

Cartwright's 1984 study is the foundation everything else stands on. Thirty male patients evaluated for sleep apnea by all-night polysomnography were compared for breathing interruptions during side versus back sleep; for the 24 who occupied both positions during the night, the count was twice as high on their backs as on their sides, and the difference was inversely related to obesity (Cartwright, 1984). Five patients who met the diagnostic criteria for sleep apnea over the whole night fell within normal limits while on their side. The study is small, all-male, and forty years old — and its finding has been confirmed by every positional study since.

The device trials are the modern evidence. Laub and colleagues randomised 101 patients with positional sleep apnea — 52 to the Sleep Position Trainer, a chest-worn vibration device, and 49 to no treatment — for two months, then followed all patients as a cohort for six months with the device (Laub et al., 2017). In the device group, mean breathing interruptions fell from 18 an hour to 10 (P<0.001) while the control group did not change, and mean supine sleep fell from 47% of the night to 17% (P<0.001). Compliance in the first two months was 75.5%, and daytime sleepiness improved after six months. The effect was maintained in those who kept using it at six months; the discontinuation rate was 28.8% at two months and 49.4% at six.

Eijsvogel and colleagues ran the head-to-head that matters most to a reader choosing a method: 55 patients with mild to moderate positional apnea, 29 to the Sleep Position Trainer and 26 to the tennis ball technique, for one month (Eijsvogel et al., 2015). Both therapies prevented supine sleep to a median of 0% of the night. Treatment success — breathing interruptions under 5 an hour — was reached by 68.0% of the device group and 42.9% of the tennis ball group. Sleep quality improved more with the device, and effective compliance was 75.9% for the device against 42.3% for the tennis ball (P=0.01). The tennis ball works. People just stop using it about twice as often.

The Cochrane review puts the whole field in perspective. Srijithesh and colleagues included eight randomised trials with 323 participants, comparing positional therapy against CPAP and against inactive control (Srijithesh et al., 2019). Against inactive control, positional therapy cut breathing interruptions by 7.38 events an hour and daytime sleepiness by 1.58 points on the Epworth scale. Against CPAP, CPAP produced the greater reduction in interruptions — by a mean difference of 6.4 events an hour (95% CI 3.00 to 9.79, low-certainty evidence) — but people used positional therapy 2.5 hours longer per night than CPAP (moderate-certainty evidence), and there was no difference in sleepiness, quality of life or quality of sleep between the two. Positional therapy is less powerful than CPAP and easier to live with — a decision for a sleep clinician, not for this page.

The long-term follow-up is the last piece. De Ruiter and colleagues randomised 99 patients with mild to moderate positional apnea to the Sleep Position Trainer or an oral appliance and followed them for 12 months; 58 completed the study, 29 in each group (de Ruiter et al., 2018). Median breathing interruptions in the device group fell from 13.2 an hour at baseline to 7.1 after 12 months (P<0.001), with no significant difference from the oral appliance group, and the effect was maintained at 12 months in those who kept using the device. Long-term adherence was similar in the two groups. The pattern across all the trials is the same: the method works, the effect lasts while the method lasts, and the method lasts only if it is comfortable enough to keep.

The 4-week training plan

This plan applies the trial principles — a physical prompt, applied every night, for weeks — to a reader who may not have sleep apnea at all. It is a practical plan, not a trial protocol, and it is labelled as such. If you have diagnosed sleep apnea, run the plan past your sleep clinician first: positional therapy is a treatment, and the Cochrane review's comparison with CPAP is a clinical decision.

Week 1 — Build the nest. Before any prompt, make side sleeping comfortable, or the prompt will drive you back to your back. Body pillow to hug, pillow between the knees, and a pillow at the right height for side sleeping — about 10 to 12 cm of supported height in the one study that measured side sleepers individually (Tian et al., 2025). Fall asleep on your side in the nest for seven nights. If you wake on your back, roll back to your side without scolding yourself — the trials count position, not willpower, and one back-sleeping night is a data point, not a failure.

Week 2 — Add the prompt. Sew a tennis ball into a pocket on the back of a snug T-shirt, or use a small backpack worn backwards, so rolling onto your back is uncomfortable. This is the tennis ball technique as the trials used it: it works, and its weakness is comfort — compliance was 42.3% against the device's 75.9% (Eijsvogel et al., 2015). Make the ball as comfortable as you can: a soft ball, a pocket that holds it between the shoulder blades rather than at the neck, a shirt that does not ride up. If the ball hurts, you will abandon it — and the trial says you will abandon it about twice as often as a device.

Week 3 — Consider a device. If the tennis ball is not holding, or you want the method with the better trial record, a chest-worn vibration device is the next step. The trials show the device cuts back sleeping from 47% of the night to 17% in two months (Laub et al., 2017), and the effect is maintained at 12 months in those who keep using it (de Ruiter et al., 2018). This page names no brand and recommends none — the trials tested a class of device, and the buying decision is yours. If you have sleep apnea, this is the point to talk to your clinician rather than buying alone.

Week 4 — Consolidate and measure. Keep the prompt for a full month, then start removing it: the tennis ball first, the body pillow last. Keep a morning log through the whole four weeks and beyond: position fallen asleep in, position woken in, snoring, headache. The log is the instrument — the trials measured position objectively, and your log is the closest you can get at home. If after four weeks you are still waking on your back most nights, the prompt was not the problem — the setup was uncomfortable, or your apnea is not positional. That is information, not failure, and it belongs in a conversation with a sleep clinician.

The side-sleeping setup: pillow, arm and mattress

The pillow. On your side, the pillow must fill the gap from the mattress to your ear, across the shoulder — and hold that height all night. Tian's side sleepers needed 9.74 to 11.76 cm of individualised height for the curve closest to natural standing, while 11.76 to 13.78 cm scored highest for comfort (Tian et al., 2025). A single flat pillow rarely reaches that for a broad-shouldered sleeper. The deciding factor is whether the stack holds: a stack that starts at 12 cm and collapses to 7 cm by morning is worse than a single firm pillow that holds 10 cm all night.

The bottom arm. The question readers ask most about the setup has no trial answer, so this is practical advice, labelled as such: do not sleep with your bottom arm under your head or under the pillow. The arm goes numb, the shoulder compresses, and you roll onto your back to escape it — which is exactly the training failing. Instead, stretch the bottom arm out in front of you, slightly bent, with the hand near your face, or tuck it between the body pillow and your chest. If your arm goes numb every night despite that, see the care section: numbness is a symptom, not a setup detail.

The mattress. The mattress matters more than most side-sleep trainees expect, because a soft mattress lets the shoulder sink, which shrinks the gap the pillow must fill. The best trial of mattress firmness — 313 adults with chronic low back pain, randomised to firm or medium-firm mattresses for 90 days — found medium-firm better for pain in bed, pain on rising and disability (Kovacs et al., 2003), and a 2015 systematic review of 24 controlled trials concluded a medium-firm, ideally self-adjusted mattress is optimal for comfort, sleep quality and spinal alignment (Radwan et al., 2015). Neither study tested side-sleep training — but a medium-firm mattress is the evidence-backed starting point, and a new mattress is worth re-running the height check for.

Left or right? Either side works for training, and the trials do not favour one. The left side is sometimes recommended for reflux, and the right side is sometimes recommended for the heart, but neither claim is tested in the position-training trials — choose the side that is comfortable, and switch sides through the night if one shoulder tires. Comfort decides adherence, and adherence decides everything.

Myths about side-sleep training

Myth: "The tennis ball is the best method."
Evidence: The tennis ball works — both methods in the head-to-head trial prevented back sleeping to a median of 0% — but the vibration device beat it on success (68.0% against 42.9%) and on compliance (75.9% against 42.3%) (Eijsvogel et al., 2015). Best is the method you keep using.
Myth: "Side sleeping is best for everyone."
Evidence: Side-lying was generally protective against waking spinal pain in a scoping review of four studies (Cary et al., 2019) — but side sleeping loads the shoulder, and for some people that cost outweighs the benefit. Best depends on your body.
Myth: "You must stay on one side all night."
Evidence: The trials measure time off the back, not time on one side. Switching sides through the night is fine — and kind to the shoulders.
Myth: "A device cures sleep apnea."
Evidence: Positional therapy is less effective than CPAP for cutting breathing interruptions — CPAP better by 6.4 events an hour — though people used it 2.5 hours longer a night (Srijithesh et al., 2019). For moderate or severe apnea, position alone is usually not enough.

What the evidence cannot tell you

The limits are the trials' limits. Every position-training trial studied people with sleep apnea, not healthy back sleepers who want to change position for comfort (Laub et al., 2017). The follow-ups are short — one to two months in most trials, twelve months in one (de Ruiter et al., 2018) — and the discontinuation rates are high: 49.4% by six months in the longest trial. No trial has tested a body pillow alone as a retraining method, and no trial has tested retraining in stomach sleepers moving to the side. The Cochrane review's evidence is low-certainty for the CPAP comparison and the trials are small — eight trials, 323 participants in total (Srijithesh et al., 2019).

What would settle it is a trial nobody has run: randomise healthy back sleepers to a body pillow, a tennis ball, a vibration device, or nothing, and measure sleep position objectively and morning symptoms for six months. Until then, the morning log is the reader's own trial — position fallen asleep in, position woken in, snoring, headache, for four weeks and beyond. If a page tells you a method is proven to retrain side sleeping in everyone, it is selling a certainty the research does not support.

When to get care

⚠ Stop self-managing and get care promptly if: your partner reports that you stop breathing during sleep, or you wake gasping or choking; you have loud snoring with pauses, morning headaches, or daytime sleepiness that affects driving or work; or you have numbness, tingling or weakness in an arm or hand that does not resolve when you change position. These are the signs of sleep apnea or nerve involvement, and positional training is not a substitute for diagnosis.

See a clinician sooner if snoring and daytime sleepiness persist despite side sleeping, if shoulder pain lasts more than a few weeks, or if morning headaches keep coming back. A sleep study is the usual next step for suspected apnea: it measures breathing interruptions separately for back and non-back sleep, which is exactly the information that says whether positional therapy can work for you. Bring your morning log — position, snoring, headaches — because it is the most useful thing you can hand a sleep clinician.

Key takeaways

A side sleeper with a body pillow, spine in a straight line from ears to hips

The nest: body pillow to hug, pillow between the knees, and the right height under the head. Comfort is what makes the training stick. Photo: illustrative.

Questions people actually ask

Does the tennis ball technique work for snoring?
Yes, but not as well as a vibration device. In the head-to-head trial, both methods prevented back sleeping to a median of 0% of the night, but treatment success was 68.0% with the device against 42.9% with the tennis ball, and effective compliance was 75.9% against 42.3% (Eijsvogel et al., 2015). The ball is free and it works — people just stop using it about twice as often, so make it comfortable before you blame yourself for quitting.
How long does it take to learn to sleep on your side?
No trial gives a learning time for healthy sleepers — every position trial studied people with sleep apnea, not people retraining for comfort. The trials ran one to two months, with the device effect established at two months and half the users stopped by six months (Laub et al., 2017). Expect weeks of practice, plan for months, and keep the morning log: position fallen asleep in, position woken in, snoring, headache.
Is there a positional device I can use instead of a tennis ball?
Yes — chest-worn vibration devices are the tested alternative, and they beat the tennis ball on success and compliance (Eijsvogel et al., 2015). In the randomised trial, the device cut breathing interruptions from 18 to 10 an hour and back sleeping from 47% to 17% of the night in two months (Laub et al., 2017), and the effect was maintained at 12 months in those who kept using it (de Ruiter et al., 2018). This page names no brand; if you have sleep apnea, talk to your clinician before buying.
What pillow setup helps me stay on my side?
A body pillow to hug, a pillow between the knees, and a head pillow tall enough to fill the shoulder-to-ear gap — an individualised height of 9.74 to 11.76 cm put the neck closest to its natural standing curve in the one study that measured side sleepers individually (Tian et al., 2025). A stack that starts at 12 cm and collapses to 7 cm by morning fails the morning check; a single firm pillow that holds 10 cm all night beats it. See our two pillows guide for the height check.
Where should I put my bottom arm when side sleeping?
In front of you, slightly bent, with the hand near your face — or tucked between the body pillow and your chest. Do not sleep with your bottom arm under your head or under the pillow: the arm goes numb, the shoulder compresses, and you roll onto your back to escape it, which is exactly the training failing. No trial has tested arm positions, so this is practical advice, labelled as such — and numbness that persists despite changing position is a reason to see a clinician.
Is side sleeping best for most adults?
For breathing and for the spine, usually yes — with a shoulder-sized qualification. A scoping review that screened 4,186 articles found side-lying generally protective against waking spinal pain and stiffness (Cary et al., 2019), and about 56% of people with sleep apnea have the position-dependent kind (Ravesloot et al., 2013). But side sleeping loads the shoulder, and for some people that cost outweighs the benefit — which is why this page spends as long on the setup as on the training.
What mattress firmness is best for side sleepers?
Medium-firm is the evidence-backed starting point, though no trial tested it for side-sleep training. In the best mattress trial, 313 adults with chronic low back pain did better on medium-firm than firm for pain in bed, pain on rising and disability (Kovacs et al., 2003), and a review of 24 controlled trials concluded a medium-firm, ideally self-adjusted mattress is optimal for comfort, sleep quality and spinal alignment (Radwan et al., 2015). A soft mattress lets the shoulder sink, which shrinks the gap the pillow must fill — so re-run the height check after any mattress change.

References

  1. Cartwright, R. D. (1984). Effect of sleep position on sleep apnea severity. Sleep, 7(2), 110–114. PubMed: 6740055 / DOI: 10.1093/sleep/7.2.110
  2. Mador, M. J., Kufel, T. J., Magalang, U. J., Rajesh, S. K., Watwe, V., & Grant, B. J. B. (2005). Prevalence of positional sleep apnea in patients undergoing polysomnography. Chest, 128(4), 2130–2137. PubMed: 16236865 / DOI: 10.1378/chest.128.4.2130
  3. Ravesloot, M. J. L., van Maanen, J. P., Dun, L., & de Vries, N. (2013). The undervalued potential of positional therapy in position-dependent snoring and obstructive sleep apnea — a review of the literature. Sleep and Breathing, 17(1), 39–49. PubMed: 22441662 / PMC3575552
  4. Ravesloot, M. J. L., White, D., Heinzer, R., Oksenberg, A., & Pépin, J. L. (2017). Efficacy of the new generation of devices for positional therapy for patients with positional obstructive sleep apnea: a systematic review of the literature and meta-analysis. Journal of Clinical Sleep Medicine, 13(6), 813–824. PMC5443742
  5. Laub, R. R., Tønnesen, P., & Jennum, P. J. (2017). A sleep position trainer for positional sleep apnea: a randomized, controlled trial. Journal of Sleep Research, 26(5), 641–650. PubMed: 28370716 / DOI: 10.1111/jsr.12530
  6. Eijsvogel, M. M., et al. (2015). Sleep position trainer versus tennis ball technique in positional obstructive sleep apnea syndrome. Journal of Clinical Sleep Medicine, 11(2), 139–147. PubMed: 25515276 / PMC4298771
  7. de Ruiter, M. H. T., et al. (2018). Durability of treatment effects of the sleep position trainer versus oral appliance therapy in positional OSA: 12-month follow-up of a randomized controlled trial. Sleep and Breathing, 22(2), 441–450. PubMed: 28913630 / PMC5918490
  8. Srijithesh, P. R., Aghoram, R., Goel, A., & Dhanya, J. (2019). Positional therapy for obstructive sleep apnoea. Cochrane Database of Systematic Reviews, 5, CD010990. Cochrane CD010990
  9. Cary, D., Briffa, K., & McKenna, L. (2019). Identifying relationships between sleep posture and non-specific spinal symptoms in adults: a scoping review. BMJ Open, 9, e027633. PubMed: 31256029 / PMC6609073
  10. Saini, Y., Rai, A., & Sen, S. (2025). Relationship between sleep posture and low back pain: a systematic review. Musculoskeletal Care, 23(2), e70114. PubMed: 40338112 / DOI: 10.1002/msc.70114
  11. Tian, S., et al. (2025). The individualized optimal pillow height and neck support design for side sleepers. Medical & Biological Engineering & Computing, 63(2), 535–544. PubMed: 39412632 / DOI: 10.1007/s11517-024-03204-x
  12. Kovacs, F. M., et al. (2003). Effect of firmness of mattress on chronic non-specific low-back pain: randomised, double-blind, controlled, multicentre trial. The Lancet, 362(9396), 1599–1604. PubMed: 14630439 / DOI: 10.1016/S0140-6736(03)14792-7
  13. Radwan, A., et al. (2015). Effect of different mattress designs on promoting sleep quality, pain reduction, and spinal alignment in adults with or without back pain; systematic review of controlled trials. Sleep Health, 1(4), 257–267. PubMed: 29073401 / DOI: 10.1016/j.sleh.2015.08.001

Last updated: 2026-11-21 by Adeel Farooq

Published 21 November 2026 — new page.

Not medical advice: The content on this site is for education only and isn't medical advice. Nothing here diagnoses or treats any condition. Talk to a qualified clinician about any health concern — especially breathing problems, numbness, or pain that doesn't improve.

How we researched this page

We mapped the reader questions assigned to this guide, then traced every number to its primary source, read first-hand on PubMed or PubMed Central: the Cartwright 1984 back-versus-side study (PMID 6740055), the Mador 2005 positional-apnea prevalence study (PMID 16236865), the Ravesloot 2013 review (PMID 22441662), the Ravesloot 2017 meta-analysis (PMC5443742), the Laub 2017 position-trainer trial (PMID 28370716), the Eijsvogel 2015 tennis-ball trial (PMID 25515276), the de Ruiter 2018 12-month follow-up (PMID 28913630), the Srijithesh 2019 Cochrane review (CD010990), the Cary 2019 scoping review (PMID 31256029), the Saini 2025 back-pain review (PMID 40338112), the Tian 2025 side-sleeper study (PMID 39412632), the Kovacs 2003 mattress trial (PMID 14630439) and the Radwan 2015 mattress review (PMID 29073401). No manufacturer claims were used as evidence. Full method: how we research.

AF

Adeel Farooq, Founder & Editor. We publish what's wrong with products too — including ones we review closely. Editorial policy