Where should the baby sleep? Bed-sharing, room-sharing, and separate rooms

The question

Many parents end up sharing a bed with their baby — sometimes by plan, sometimes because it's 3am and the baby will only sleep on a warm body. Others keep the baby in their room in a cot, and some have the baby in a separate room from early on. Official advice in the US says never bed-share; advice in the UK is more nuanced; and the recommended answer for "how long in our room?" changed as recently as 2022. What does the evidence actually say about the three options — baby in your bed, baby in your room on a separate surface, baby in a separate room — and which circumstances make each arrangement dramatically more or less dangerous?

Short answer

The safest documented arrangement is the baby in your room on a separate surface: guidelines summarise this as cutting SIDS risk substantially compared with the baby sleeping alone in a separate room, and UK guidance asks for it for the first six months [1][2][3][22]. Bed-sharing on an adult bed is linked to higher SIDS risk — but the risk is not one number; it depends enormously on the circumstances. The danger is greatest with a few specific hazards: a parent who smokes, has drunk alcohol or taken sedating drugs, or a baby sleeping on a sofa or armchair rather than a bed. Without those hazards, one large analysis still finds a meaningful increase for babies under 3 months (a modelled estimate of about 0.2 vs 0.08 per 1,000), while other large analyses find little or no added risk — and the field has not settled the disagreement [4][6].

Guidelines split on what to tell parents about bed-sharing: US guidance (AAP) advises against it under all circumstances, while UK guidance (NICE, UNICEF Baby Friendly) takes a harm-reduction approach — discussing safer practices with every family, because many parents end up bed-sharing whether they plan to or not [1][2][13]. For parents themselves, the evidence is weaker (rating C): sharing the sleep space is linked to more broken maternal sleep, and low mood and sleep arrangements seem to influence each other in both directions.

(Overall evidence rating: B. The bed-sharing hazard evidence is B; the exact size of the room-sharing protection and how long to keep it up are marked C in the text below; the residual risk of non-hazardous bed-sharing is contested — rated C.)

The three arrangements at a glance

Note on the baseline: the ~0.3 per 1,000 figure is the UK population average across all sleep arrangements [3], not the measured rate for any single arrangement — no study has counted deaths per thousand in each arrangement group directly.

What the strongest evidence says

A few definitions first. SIDS (sudden infant death syndrome) is the sudden, unexplained death of a baby under one year after a full investigation. SUID (sudden unexpected infant death) is the broader US term: SIDS plus deaths from unknown causes plus accidental suffocation or strangulation in bed. Bed-sharing means the baby sleeps on the same surface as an adult; room-sharing means the baby sleeps in the parents' room on a separate surface — a cot, crib, Moses basket, bassinet, or bedside sleeper that meets safety standards, close to the parents' bed [1][2]; sofa-sharing means falling asleep with the baby on a sofa or armchair — researchers treat this as a different and far more dangerous category [1][2].

How common is the outcome we're talking about? In the UK, SIDS affects roughly 0.3 in every 1,000 live births (about 0.03% of births) — 164 unexplained deaths under one year in England and Wales in 2023 (0.28 per 1,000), and 188 across the UK (0.29 per 1,000) [3]. In the US, the broader SUID category runs about 1 in 1,000 live births [5]. These are rare events — which is exactly why absolute numbers matter more than scary-sounding multiples.

Bed-sharing: the risk is driven by hazards, and the hazards are enormous

Bed-sharing roughly triples SIDS risk overall — but "overall" hides the real story. The largest individual-level analysis ever done combined five case-control datasets (1,472 SIDS cases, 4,679 controls across the UK, Europe and Australasia) and found an adjusted odds ratio of 2.7 for bed-sharing versus not bed-sharing [6]. A separate UK analysis of two case-control studies (400 SIDS cases) found a multivariable odds ratio of 3.9 for co-sleeping overall [7]. Both are case-control studies — strong observational designs, but observational: they can show association, not prove the bed itself caused the death.

When researchers split bed-sharers by circumstance, the picture changes completely [7][8]:

• Sofa or armchair: sleeping with a baby on a sofa or chair carries an odds ratio of roughly 18–21 compared with not sharing sleep — the single most dangerous common arrangement [4][7]. The AAP's technical report describes the risk of death on a couch, armchair or cushion as 22- to 67-fold higher [2].
• Alcohol (more than ~2 units) or sedating drugs: bed-sharing next to an adult who drank more than two units of alcohol in the previous 24 hours carries an odds ratio of roughly 18 [7]. One smaller UK study found an odds ratio above 40 for parental alcohol use, though with very wide confidence intervals [8].
• Smoking: in one large UK analysis, bed-sharing with a smoking parent carried an odds ratio of roughly 9 for babies under 3 months (and was not statistically significant for older babies) [7]. An 8-year Scottish study found maternal smoking raised SIDS risk about 4-fold on its own but about 25-fold combined with bed-sharing — the risks multiply [9]. In that study, 87% of bed-sharing SIDS cases had mothers who smoked during pregnancy; the two risks multiply each other [9].
• Young, preterm, or low-birth-weight babies: babies under about 3–4 months, born before 37 weeks, or weighing under 2.5 kg are more vulnerable in shared sleep [1][7].
• Non-parental bed-sharers: sharing the bed with other children or pets carries higher risk than sharing with a parent [10][13].

Without those hazards, the added risk looks small — but this is where researchers disagree. A 2022 evidence review and meta-analysis by Blair and colleagues found a multivariable odds ratio of 1.08 for bed-sharing in the absence of hazardous circumstances — essentially no detectable increase, and not statistically significant [4]. But Carpenter's 2013 analysis, using different methods and a different comparison group, found a five-fold increased risk even in the lowest-risk group (breastfed babies under 3 months of non-smoking, sober parents), translating to an estimated absolute risk of 0.08 per 1,000 for room-sharing versus 0.23 per 1,000 for bed-sharing [6]. Blair's team argues Carpenter's choice of an ultra-low-risk comparison group magnified the apparent difference and made it non-generalisable [4]; Carpenter and colleagues, in a published PLOS comment on Blair's 2014 analysis, argue the reverse — that Blair's non-hazardous estimates were biased downward by how the comparison group was built [11]. Honest summary: the risk without hazards is either zero or small, and the field has not settled which. (The residual-risk sub-question is rated C.)

Bed-sharing is associated with longer breastfeeding. In a UK study of 678 breastfeeding mothers, the median duration of any breastfeeding was 14 weeks for mothers who rarely bed-shared, 24 weeks for intermittent bed-sharers, and over 26 weeks for frequent bed-sharers [12]. This is observational — mothers most motivated to breastfeed are also the most likely to bed-share, so cause and effect are tangled — but guideline committees have concluded that blanket "never bed-share" messaging would most likely shorten breastfeeding for some families [13].

Room-sharing: the evidence-backed middle ground

The headline figure comes from case-control studies, not trials. Two landmark UK studies anchor the literature. A 1999 English study of 325 SIDS deaths and 1,300 controls found that sleeping in a separate room carried roughly ten times the odds of SIDS compared with room-sharing (multivariate OR 10.49) [20]. A 2005 Scottish study of 123 SIDS deaths and 263 controls found a smaller effect: separate-room sleeping carried about three times the odds, though the confidence interval was wide and only just excluded no effect (OR 3.26) [9]. Both are case-control studies: they compare sleep locations among babies who died with those of living babies. They can show an association, not prove the room itself caused the difference — and plausible confounders (parental smoking, socioeconomic circumstances, breastfeeding, bedding, infant health, and why families chose each arrangement) were only partly adjustable [20][9].

So what does "up to 50%" actually mean? The AAP's 2022 policy statement cites several such observational studies and concludes that room-sharing on a separate surface decreases SIDS risk "by as much as 50%" [1]. The technical evidence base behind that statement reviews the underlying case-control literature and reaches the same cautious conclusion: protection is suggested through the first year, but the data cannot pin down an exact magnitude [2]. Treat the phrasing carefully: "as much as" is doing real work. It is the top end of a range of observational estimates, not a measured result from a trial, and you cannot convert these case-control proportions into a population absolute-risk figure. The honest reading is: studies consistently point toward protection, and the true causal effect is somewhere between modest and large — nobody can pin it down precisely. (The size of the room-sharing protection is rated C.)

This evidence will never get stronger than this — and that's worth understanding. You cannot run a randomised trial of sleep location: no ethics committee would let researchers assign newborns to sleep alone in a separate room. So the field is permanently limited to observational studies, with all the confounding that implies. Guideline committees (the AAP, the Lullaby Trust, UNICEF Baby Friendly) land in the same place anyway — same room, separate surface, for at least the first six months — not because the evidence is airtight, but because the direction is consistent across studies, countries, and decades, and the downside of room-sharing is inconvenience rather than danger [1][3][22].

How long? The guidance says six months, anchored to when deaths happen. The AAP recommends room-sharing "ideally for at least the first 6 months", adding that room-sharing without bed-sharing appears protective through the first year but "there is no specific evidence for when it might be safe to move an infant to a separate room before 1 year of age" — the first six months matter most because that is when deaths concentrate [1]. The UK data underline this: 87% of SIDS deaths happen in the first six months, with the peak at 1–2 months of age [3]. The Lullaby Trust accordingly encourages families to continue room-sharing up to six months [3], and UNICEF Baby Friendly advises keeping the baby in the parents' bedroom at night "for at least the first six months" [22]. Note the guidance changed: the older AAP wording said "ideally a year"; the 2022 update shortened it to six months because the evidence for the 6–12 month window is weaker [1]. (The how-long guidance is rated C: six months is anchored to the age distribution of deaths, not a tested threshold.)

One behaviour that crosses all three arrangements: room-sharing can quietly slide into bed-sharing at 3am. In a US study, room-sharing families were more likely to end up with the baby in the adult bed overnight [21]. Since bed-sharing carries a very different risk profile, this is exactly the moment the harm-reduction checklist below is for.

The sofa warning

Falling asleep with the baby on a sofa or armchair is the single most dangerous common sleep arrangement — more dangerous than either room-sharing or bed-sharing [1][2]. The risk of death on a couch, armchair or cushion is described by the AAP as 22- to 67-fold higher [2], and case-control studies give odds ratios of roughly 18–21 versus not sharing sleep [4][7]. This is also why guidelines matter in what they say about beds: the UK's UNICEF Baby Friendly Initiative explicitly warns against overstating bed risks in a way that drives exhausted parents to feed on the sofa instead [14]. If you catch yourself dozing off somewhere that isn't a prepared bed — especially a sofa or armchair — move the baby to their own sleep surface, or to a prepared bed as the less dangerous option [2]. If exhaustion tempts you to feed on the sofa "just for a minute", set up a chair or the bedside with a plan to return the baby to their own surface [1].

If you bed-share — by plan or by 3am accident — reduce the hazards

Guidelines agree on the hierarchy: a separate sleep surface in your room is safest; an adult bed with hazards avoided is next; a sofa or armchair is the most dangerous place to fall asleep with a baby — more dangerous than the bed [1][2][13].

• Use a firm, flat mattress with no gaps against the wall; keep pillows, duvets, blankets and soft toys away from the baby; baby sleeps on their back, lightly dressed [1][13].
• Never bed-share if you or your partner smokes (even outside the bedroom), if the mother smoked in pregnancy, or if either of you has drunk alcohol or taken anything sedating — including medicines that cause drowsiness [1][13].
• Don't bed-share if your baby was born before 37 weeks or under 2.5 kg [1][13].
• Keep other children and pets out of the bed; bed-sharing with anyone other than a parent carries higher risk [10][13].
• If you feed in bed, plan the return trip: put the baby back on their separate surface before you fall asleep. Notably, bringing a baby into bed briefly for feeding or comfort and then returning them to their own sleep space was not found to be a risk factor [6].
• Reassess every night: tiredness, alcohol, and new mattresses change the calculation.

What it means for the parents

So far this topic has been all about the baby. But you're in the bed — or the next room — too, and the research has things to say about you as well. One big caveat up front: scientists have studied parents' sleep and mood far less than they've studied SIDS, none of these studies randomly assigned families to sleep arrangements (which would be unethical), and most of them blur bed-sharing and room-sharing together. So read this as clues, not verdicts. (Parental-outcomes evidence rating: C.)

Your sleep will probably be more broken if the baby sleeps in your room — bed or not. Two studies that tracked mothers' sleep with wrist monitors (rather than just asking them) found the same pattern: mothers sharing a sleep space with their baby had more fragmented nights — more wakings, longer stretches awake — than mothers whose babies slept in a separate room, while the babies' own monitor-measured sleep differed little [15][16]. In the study that followed families to 18 months, room-sharing mothers still showed poorer actigraphic sleep — lower sleep percentage, shorter longest sleep periods, more wakings [16]. In a US study of first-time mothers, mothers' questionnaire reports showed a bigger gap: at 9 months, early independent sleepers were reported to sleep about 40 minutes longer per night, with the longest unbroken stretch 100 minutes longer, than room-sharers [21]. But these are questionnaire reports, and the actigraphy finding — babies' monitor-measured sleep barely differed while the mothers' sleep took the hit — suggests the reports partly measure what mothers notice [15][21]. Two honest complications. First, in the wrist-monitor studies "co-sleeping" mostly meant room-sharing, so they can't tell you precisely how bed-sharing compares with a bedside cot [15][16]. Second, one study found that mothers who already slept badly during pregnancy were the ones most likely to end up co-sleeping — so some of this may be about who chooses to co-share, not what co-sharing does [15].

Low mood and sleep arrangements travel together — but nobody knows which leads. In a 15-month cohort of 428 predominantly low-income, non-Hispanic Black mother–infant pairs, bed-sharing at the start predicted higher maternal depressive symptoms at 15 months, even after adjustment — but not the reverse [18]. This is still observational, though, and other research finds the arrow pointing the other way: mothers already struggling with mood or sleep are more likely to bring the baby into their bed or room in the first place [15][17]. A 2025 US population study of over 100,000 mothers adds one more signal: about 1 in 9 mothers (11.7%) reported postpartum-depression symptoms, and room-sharing was linked to slightly higher odds of those symptoms, but only among more advantaged mothers (married, educated beyond high school, privately insured) — adjusted odds ratios of 1.11 to 1.15 — and because it was a one-time survey, it can't show whether room-sharing came before the low mood or the other way round [23]. It may well run in both directions — and a third factor (a difficult baby, no support, sheer exhaustion) may be driving both the co-sleeping and the low mood.

It can strain the partnership, and invite judgment. In one year-long study, mothers who were still sharing sleep (room or bed) with their baby after 6 months reported worse marital adjustment and more negative coparenting than mothers whose babies slept alone — but fathers reported no such link, and the differences were already visible at 1 month, which points to selection as much as cause [17]. Mothers who co-slept past 6 months also reported feeling more criticised for their choice — even after accounting for how much they themselves liked the arrangement [19].

For fathers and partners specifically: we mostly don't know. Fathers' sleep and health by sleep arrangement have barely been studied. In one study that tracked fathers' sleep as well as mothers', fathers' sleep was not related to the infant's sleep arrangement — but the mothers in that study did nearly all the night care, so it may look different in families who split nights [17]. The same-sample longitudinal study found that in room-sharing families, fathers were less involved in night-time care than in families where the baby slept separately [16]. If you want genuinely shared nights, proximity alone won't create them — you have to plan them.

A word on timing the move. The temptation to move the baby out early often peaks around 4–6 months — exactly when parental exhaustion is worst and when the data say protection still matters (87% of deaths in the first six months [3]). There is no evidence for a "safe" earlier month, so if you do move the baby before six months, know that you're moving ahead of the evidence, not with it [1]. After six months the picture changes: the observational evidence for continued protection is thinner, and with no tested safe month before age one, the decision becomes a family judgment call [1][2].

The bottom line for you: the safest arrangement for the baby (baby in your room, on a separate surface) may cost you some sleep, especially in the first 6 months — and for some families that cost shows up in mood and partnership friction. That's a real trade-off, not a reason to ignore the safety evidence. If you're finding the nights are grinding you down, that's worth raising with your midwife, health visitor, or GP just as much as any question about the baby's sleep.

What remains uncertain

• The low-risk residual question (Carpenter vs. Blair) is the central unresolved dispute for bed-sharing. It cannot be resolved by a randomised trial — randomly assigning babies to bed-share would be unethical — so it will stay observational, with all the confounding that implies.
• The true size of the room-sharing protection. Estimates range from "up to 50%" (AAP summary) to ten-fold odds in one 1990s English study [20] — the field cannot pin down a causal number, and no randomised trial is ethically possible.
• The safe month to move. There is no evidence for a specific month before age one at which moving the baby to a separate room becomes safe; six months is anchored to the age distribution of deaths, not a tested threshold [1].
• Why room-sharing protects. Proposed mechanisms — lighter shared arousal, faster parental response, fewer hazardous bedding situations — are plausible but unproven; residual confounding could explain part of the effect.
• Whether the baby actually sleeps worse in the parents' room. Questionnaire studies say yes [21]; actigraphy says the difference may be mostly in what parents notice [15].
• Whether breastfeeding itself protects during bed-sharing is contested: some analyses suggest it does, others find the apparent protection is really the absence of smoking and alcohol among breastfeeding mothers [6][4]. Whether room-sharing on a separate surface has any effect on breastfeeding duration is unclear [21].
• Absolute risks by hazard stratum are not measured, only modelled for one subgroup. The 0.08 → 0.23 per 1,000 figures come from Carpenter 2013's statistical model [6]; no study reports directly observed absolute SIDS rates for smoking, alcohol, or sofa-sharing strata, so those are presented as relative risks only.
• We lack good data on how common bed-sharing actually is (needed for population-level risk), on unintentional versus planned bed-sharing, on mattress firmness within bed-sharing, on bed-sharing with fathers versus mothers, and on twins.
• The parental side is thinly studied. One large US population survey links room-sharing to slightly higher postpartum-depression symptoms among more advantaged mothers, but it is cross-sectional and cannot establish which came first [23]. Fathers' sleep and wellbeing by arrangement, relationship strain, and work outcomes are barely studied.

Benefits and risks in absolute terms

Reading odds ratios without base rates is how "five times the risk!" becomes misleading. Here are the strata, with the UK SIDS baseline of roughly 0.3 per 1,000 live births as the anchor [3]. Only Carpenter 2013 published modelled absolute risks (one low-risk subgroup); for hazard strata, no directly observed absolute rates exist, so those rows give relative risks only — converting case-control odds ratios into per-1,000 figures would require a baseline exposure rate we don't have. And no study gives an absolute "X in 1,000" figure for room-sharing versus separate-room sleeping: the studies are case-control — they compare the mix of sleep locations among babies who died with the mix among living babies — so they produce odds ratios, not death rates per thousand. Any per-thousand comparison would require assumptions the data don't support, so we won't invent one.

What we can add in absolute terms is timing: 87% of SIDS deaths happen in the first six months [3], and the highest-risk age is 1–2 months [3] — which is also when the sleep disruption of room-sharing tends to be worst for parents.

For perspective: the oft-quoted "five-fold risk" for low-risk young infants translates to roughly 2 extra deaths per 10,000 babies — a real increase from a very small base, not a common outcome. Carpenter's team estimated that around 88% of the SIDS deaths that occurred while bed-sharing would not have occurred had the baby been placed on their back in a cot by the parents' bed [6].

Practical considerations

• Same room for every sleep, day and night, for the first six months — SIDS risk applies to naps too [3][22].
• Separate surface, close to your bed: a cot, crib, Moses basket, bassinet, or bedside sleeper that meets current safety standards, within arm's reach for feeding and checking [1].
• Keep the sleep surface clear and flat. Room-sharing protects; the surface still has to be safe — firm, flat, and free of pillows, duvets, bumpers, and toys, with the baby on their back [1].
• Same rules for every caregiver. Whoever puts the baby down — partner, grandparent, night nanny — the arrangement is the same: same room, separate surface, back-sleeping position [1][22].
• Expect the trade-off and plan for it: if broken sleep is harming your health or mood, arrange shifts, accept help, and sleep when the baby sleeps where you can.
• Don't trade room-sharing for sofa-sharing. If exhaustion tempts you to feed the baby on the sofa "just for a minute", set up a chair or the bedside with a plan to return the baby to their own surface — falling asleep on a sofa or armchair with a baby is the single most dangerous common arrangement [1].
• After six months, the evidence for continued protection is thinner — the AAP notes observational protection may extend through the first year, but no evidence establishes a specific safe month to move before age one — so the decision increasingly becomes a family judgment call rather than a safety imperative [1][2].

When to talk to your doctor, midwife, or pediatrician

• Your baby was born preterm (before 37 weeks) or at low birth weight — ask specifically about sleep arrangements and any extra precautions.
• You or your partner smokes, or you take sedating medication (including prescribed) and aren't sure how it affects night-time responsiveness.
• Your baby has breathing difficulties, is unusually hard to rouse, has pauses in breathing, or has reflux that worries you — seek prompt medical advice for breathing concerns.
• Night waking is accompanied by poor weight gain, feeding difficulties, or frequent vomiting — these can signal medical issues, not just habit.
• You are falling asleep while holding or feeding the baby and can't prevent it — ask your midwife or health visitor for strategies; unplanned sleep on sofas or armchairs is the highest-risk scenario.
• Broken sleep is seriously affecting your mood, your relationship, or your ability to function safely (for example, driving) — this deserves professional support in its own right.
• You are anxious about SIDS or about your current sleep setup — health visitors and GPs have these conversations routinely and without judgment.

References

1. Moon RY et al. Sleep-Related Infant Deaths: Updated 2022 Recommendations for Reducing Infant Deaths in the Sleep Environment. Pediatrics. 2022. Policy statement — bed sharing not recommended under any circumstances; room-sharing advised. https://publications.aap.org/pediatrics/article/150/1/e2022057990/188304/Sleep-Related-Infant-Deaths-Updated-2022 — [Guideline]
2. Moon RY et al. Evidence Base for 2022 Updated Recommendations for a Safe Infant Sleeping Environment. Pediatrics. 2022. Technical report — room-sharing reduces SIDS by up to 50%; sofa/armchair/cushion death risk described as 22- to 67-fold higher. https://publications.aap.org/pediatrics/article/150/1/e2022057991/188305/Evidence-Base-for-2022-Updated-Recommendations-for — [Guideline]
3. The Lullaby Trust. SIDS and SUDC facts and figures (ONS data, 2023): 164 unexplained deaths under one year in England and Wales (0.28 per 1,000 live births); 188 across the UK (0.29 per 1,000). https://www.lullabytrust.org.uk/wp-content/uploads/2025/02/SIDS_and_SUDC_Facts-and-Figures.pdf — [Population data]
4. Blair PS, Ball HL, Pease AS, Fleming PJ. Bed-sharing and SIDS: an evidence-based approach. Archives of Disease in Childhood. 2022. Meta-analysis: non-hazardous bed-sharing multivariable OR 1.08 (95% CI 0.58–2.01), not significant; critique of Carpenter 2013 reference-group methods. https://doi.org/10.1136/archdischild-2021-323469 — [B]
5. CDC. Data and Statistics for SUID and SIDS: ~3,700 SUID in the US in 2022; SUID rate ~100.7 per 100,000 live births (~1 per 1,000). https://www.cdc.gov/sudden-infant-death/data-research/data/ — [Population data]
6. Carpenter R et al. Bed sharing when parents do not smoke: is there a risk of SIDS? An individual level analysis of five major case–control studies. BMJ Open. 2013. n=1,472 cases / 4,679 controls; bed-sharing adjusted OR 2.7; lowest-risk subgroup AOR 5.1 with modelled absolute risk 0.08 → 0.23 per 1,000. https://bmjopen.bmj.com/content/3/5/e002299 — [B]
7. Blair PS, Sidebotham P, Pease AS, Fleming PJ. Bed-sharing in the absence of hazardous circumstances: is there a risk of sudden infant death syndrome? An analysis from two case-control studies conducted in the UK. PLOS One. 2014;9(9):e107799. doi:10.1371/journal.pone.0107799 https://pmc.ncbi.nlm.nih.gov/articles/PMC4169572/ — [B] (combined analysis of 2 UK population case-control studies; annotation: n=400 cases / 1,386 controls; co-sleeping overall OR 3.9 [2.7–5.6]; sofa OR 18.3 [7.1–47.4]; alcohol >2 units OR 18.3 [7.7–43.5]; smoking OR 8.9 [5.3–15.1] under 3 months, 1.4 [0.7–2.8] not significant for older infants; non-hazardous bed-sharing OR 1.1 [0.6–2.0] overall, 1.6 [0.96–2.7] under 3 months (not significant), 0.1 [0.01–0.5] for older infants)
8. Blair PS et al. Hazardous cosleeping environments and risk factors amenable to change: case-control study of SIDS in south west England. BMJ. 2009. n=80 cases / 87 controls; alcohol OR 41.6, shared sleep surface OR 21.8, maternal smoking in pregnancy OR 13.4. Wide confidence intervals — small study, direction consistent with larger ones. https://pmc.ncbi.nlm.nih.gov/articles/PMC2762037/ — [C]
9. Tappin D et al. An 8 year study of risk factors for SIDS: bed-sharing versus non-bed-sharing. Archives of Disease in Childhood. 2005. 87% of bed-sharing SIDS cases had mothers who smoked in pregnancy; smoking OR 3.74 alone vs 24.78 combined with bed-sharing. https://pmc.ncbi.nlm.nih.gov/articles/PMC2065975/ — [B]
10. Vennemann MM et al. Modifiable risk factors for SIDS in Germany: results of GeSID. Acta Paediatrica. 2005. Bed-sharing without maternal smoking OR 2.20 [0.99–4.91] (borderline); bed-sharing with maternal smoking OR 6.44 [2.62–15.81]; sharing with non-maternal others riskier than with mother. Via NICE NG194 evidence review: https://www.nice.org.uk/guidance/ng194/evidence/n-cosleeping-risk-factors-pdf-326764485978 — [B]
11. Published PLOS comment by Carpenter and colleagues on Blair et al.'s 2014 analysis (ref 7): argues the non-hazardous bed-sharing odds ratios were underestimated because the non-co-sleeping comparison group itself contained many smoking/alcohol-exposed parents. The authors' approach adjusted for these factors in multivariable models; the dispute is unresolved in the literature. https://journals.plos.org/plosone/article/comment?id=10.1371/annotation/1ea5c75a-8b0c-46cb-88ee-32b64e861da6 — [Commentary]
12. Ball HL et al. Bed-sharing by breastfeeding mothers: who bed-shares and what is the relationship with breastfeeding duration? Acta Paediatrica. 2016. Median any-breastfeeding: 14 weeks (rare bed-sharers) vs 24 weeks (intermediate) vs >26 weeks (frequent); observational, confounding by breastfeeding intent. https://doi.org/10.1111/apa.13354 — [C]
13. NICE. Postnatal care guideline NG194: evidence review N — co-sleeping risk factors (2021). Committee advice: bed-sharing strongly advised against with smoking, >2 units alcohol, sedating drugs/medication, preterm/LBW infant, or non-parental sharers; advises discussing safer practices with all parents. https://www.nice.org.uk/guidance/ng194/evidence/n-cosleeping-risk-factors-pdf-326764485978 — [Guideline]
14. UNICEF UK Baby Friendly Initiative: position that overstating bed-sharing risks can drive exhausted parents to the more dangerous sofa; recommends honest, individualised safer-sleep conversations. Via Public Health Agency NI safer-sleeping guidance: https://www.publichealth.hscni.net/sites/default/files/2019-04/Promoting%20Safer%20Sleeping%20for%20Infants%20Guidance%20document%20Final.pdf — [Guideline/expert position]
15. Volkovich E et al., Sleep Medicine 2015. doi:10.1016/j.sleep.2015.08.016 — [C] (longitudinal observational, 153 families, actigraphy + diaries to 6 months; "co-sleeping" mostly meant room-sharing; co-sleeping mothers had more fragmented sleep — more awakenings, longer wake after sleep onset — than solitary-sleeping mothers while infant actigraphic sleep did not differ by group; worse maternal sleep in pregnancy predicted co-sleeping at 3 months, so causation may run both ways)
16. Volkovich E et al., Sleep 2018;41(2):zsx207. doi:10.1093/sleep/zsx207 — [C] (longitudinal observational, 5-night actigraphy + diaries at 3, 6, 12, and 18 months — n=146, 141, 135, 130; persistent room-sharing mothers had lower actigraphic sleep percentage, shorter longest sleep periods, and more wakings; infant actigraphic differences were limited to longest sleep period while mothers reported more infant wakings; room-sharing families had higher maternal separation anxiety and lower paternal overall/night-time caregiving involvement; authors state no causal effects can be inferred)
17. Teti DM et al., Developmental Psychology 2016;52(8):1169–1181. doi:10.1037/dev0000148 — [C] (prospective longitudinal, 280 families, 1–12 months, questionnaires; "co-sleeping" included room- and bed-sharing; mothers still co-sleeping beyond 6 months had lower marital adjustment and more negative coparenting than mothers of solitary-sleeping infants — differences evident as early as 1 month, with selection vs causation unresolved; fathers' reports showed no link with sleep arrangements, and fathers' sleep was not associated with infant sleep-arrangement patterns)
18. Nulty et al., Sleep Health 2022. PMID 34922857, PMCID PMC8821130 — [C] (longitudinal cohort, 428 predominantly low-income, non-Hispanic Black mother–infant dyads, pregnancy through 15 months; bed-sharing at baseline predicted higher maternal depressive symptoms at 15-month follow-up after adjustment — the reverse direction was not found; observational secondary analysis, so unmeasured strain could explain the link)
19. Shimizu M, Teti DM, Infant and Child Development 2018. PMID 29930484 — [C] (103 mothers followed through the first year; mothers who co-slept — room or bed — beyond 6 months reported significantly higher social criticism about their choice than mothers who moved to solitary sleep by 6 months, F(1,101)=12.62, p<.001, even after accounting for how much they liked the arrangement; self-report)
20. Blair PS et al. Babies sleeping with parents: case-control study of factors influencing the risk of the sudden infant death syndrome. BMJ. 1999;319:1457–1462. (C — large case-control, but very wide CI and era-specific confounding)
21. Paul IM et al. Mother-infant room-sharing and sleep outcomes in the INSIGHT study. Pediatrics. 2017;140(1):e20170122. (C — prospective but sleep location observed, not randomised; questionnaire-reported sleep)
22. UNICEF UK Baby Friendly Initiative. Caring for your baby at night: a health professionals' guide (2022 reissue). (D — professional guidance, not a study)
23. Patibandla S, Haile ZH. Association Between Infant-Mother Room-Sharing and Symptoms of Postpartum Depression: A Population-Based Study. Matern Child Health J. 2025. PMCID: PMC12006247. (C — cross-sectional population survey; adjusted for many covariates, but cannot establish whether room-sharing preceded symptoms or the reverse)

Changelog

• 2026-09-14: Topic created by merging bed-sharing-safety and room-sharing-separate-surface (both approved 2026-09-09) into a single three-arrangement topic; overall rating B; the room-sharing magnitude and how-long sub-claims are explicitly marked C in the text, and the residual non-hazardous bed-sharing risk stays C as before. Replaced the stale "200–250 deaths/yr in England and Wales" figure with the 2023 ONS figures (164 deaths, 0.28/1,000) per the 2026-09-09 review resolution. Full decision history of both predecessor topics is carried verbatim in review.md.