Dream Feed Schedule: What the Research Shows and How to Time It

The 1993 study almost everyone is quietly quoting

Most confident claims I found about the late-evening feed trace back to one small trial. In 1993, Teresa Pinilla and Leann Birch published “Help me make it through the night” in Pediatrics. Twenty-six first-time parents and their newborns were randomly split into two groups of 13. The treatment parents were asked to offer what the researchers called a “focal feed” between 10pm and midnight every night. That is the same thing most parents now call a dream feed.

What tends to get lost is that the focal feed was only one of three instructions. The same parents were also told to gradually lengthen the gaps between middle-of-the-night feeds by trying reswaddling, a diaper change or walking first, and to make day and night look and feel as different as possible. All families kept 72-hour diaries every week until their babies were eight weeks old.

By three weeks, the treatment babies had significantly longer sleep episodes at night. By eight weeks, the abstract reports that 100% of treatment infants were sleeping through the night, compared with 23% of control infants. Notice the definition: sleeping from midnight to 5am. Five hours. The treatment babies also fed less at night but made up for it by taking more milk in the early morning, and total 24-hour milk intake did not differ between the groups.

Two caveats. First, because the late feed was bundled with interval-stretching and day-night cues, the study cannot tell you which ingredient did the work. Second, the sample was tiny. A Cochrane review that searched for controlled trials of this kind of programme for exclusively breastfed babies found exactly one trial to include, and concluded there is “some evidence” that first-time parents using a structured programme can teach their babies to sleep through the night by eight weeks.

A larger UK trial is the more sobering read. St James-Roberts and colleagues randomly assigned 610 mothers and newborns to a behavioural programme, an educational intervention, or routine care. Their programme also included a focal feed between 10pm and midnight. The finding on that element was blunt: it “was not implemented.” Parents did not adopt it. The part they did adopt, settling babies in the cot and keeping night interaction minimal, was linked to a modest rise in babies sleeping five or more hours at night by 12 weeks.

Finally, a 2022 observational study. Quante and colleagues followed 313 infants and found that bottle feeding at bedtime at one month was associated with a longer longest stretch of night sleep at six months. That is a feed at bedtime, not a late dream feed, and an association, not proof of cause. It is consistent with the idea, and it does not settle it.

Work backward from your bedtime, not hers

Most examples I found online start from the baby's bedtime. I think that is backwards. The whole purpose of a late top-up is to protect your first block of sleep, so the anchor that matters is when you intend to be asleep, and the feed is placed just before it. The arithmetic: take the time you want your head on the pillow, subtract the time a sleepy feed and resettle usually takes you, and that is your target. If a drowsy feed plus a burp and a quiet return to the crib takes around 30 minutes in your house, a 10:30pm lights-out means starting at about 10pm.

The table below runs that sum for a few made-up evenings. Treat every clock time in it as an illustration. No authoritative body publishes a dream feed time, and the baby bedtimes shown are not recommendations either. The middle column is the only part drawn from guidance, and it comes from the AAP's feeding page on HealthyChildren.org, which describes how far apart feeds tend to be at different ages. That spacing tells you whether a late feed is a separate event or just the next regular feed with a new name.

Read the first row closely. In the early weeks, the AAP describes breastfed newborns nursing about every two hours and bottle-fed newborns eating every two to three hours. At that pace a baby put down at 7:30pm will usually be due again somewhere in the evening anyway, so a dream feed schedule at this age mostly means nudging a feed that was going to happen into a window that suits you. The idea starts to become a real choice later, once the gaps between feeds have stretched far enough that a 10pm feed is something you are deciding to offer rather than something her hunger was going to demand.

Notice too what the table leaves out. There is no row for “the age to start,” because nothing I could find in the AAP, NHS or NICHD material names one. There is also no promised outcome column. The Pinilla trial measured its outcomes in a group of 13 babies who were also getting two other interventions, and the St James-Roberts parents largely did not sustain the feed. If the sum produces a time that consistently wakes her fully, or if she starts waking at 2am as well as being fed at 10pm, that is the arithmetic telling you this particular tool does not fit your baby.

If her bedtime moves, recalculate from yours: babies who go down at 6:30pm and 8pm can share a dream feed time if their parents go to bed at the same hour. The baby sleep schedules hub has full-day examples by age.

What do the AAP and the NHS say about feeding a sleeping baby?

Neither the AAP nor the NHS has a page about dream feeds that I could find. Both have guidance on feeding and night feeds, and it shapes how to think about a planned late feed.

The AAP's starting position is responsive feeding. Its HealthyChildren.org feeding page says that, in general, babies should be fed whenever they seem hungry, looking to the baby rather than the clock. It lists cues such as rooting and hands to the mouth, and notes that crying is a late sign of hunger. It also makes an exception: for babies born prematurely or with certain medical conditions, scheduled feedings advised by the pediatrician are best, and a baby who is having trouble gaining weight should not go too long between feeds, even if that means waking the baby. A dream feed is the opposite of feeding on cue in one sense: you choose the time, and she is not asking. For a baby whose weight is being watched, that is a question for the pediatrician.

The NHS is compatible. Its helping your baby to sleep page says newborns wake at night because they need feeding, that by three to six months babies need fewer night feeds and some may sleep five to eight hours or longer at night, “but not all,” and that for some babies aged six months to a year night feeds may no longer be necessary. Its night-time habits line up with a dream feed: keep the lights low, talk little and quietly, put your baby down as soon as they have been fed and changed.

The AAP's getting your baby to sleep page advises, for babies four months and older, putting babies to bed drowsy, not asleep, so they learn to fall asleep on their own, while saying you can attend to night needs like feeding. If your aim is for her to settle herself, be honest about whether the late feed helps or hinders that. The how to get a baby to sleep through the night guide goes further into that side.

The safety part is not mine to write, so I will point you to it. After any feed, the AAP's safe sleep guidance is to place your baby on her back on a firm, flat surface in her own sleep space. If you bring her into your bed to feed, both the AAP and NICHD Safe to Sleep say to put her back in her own sleep area before you go to sleep, or as soon as you wake if you dozed off. NICHD's safe sleep environment page singles out couches and armchairs as particularly dangerous when an adult falls asleep on them while feeding. At 10:30pm, read that one twice.

Stopping: there is no official age

Nothing on the AAP, NHS or NICHD pages I read names an age to drop a dream feed, and I have not found a peer-reviewed study that tests one. The Pinilla trial ended at eight weeks. St James-Roberts followed families to nine months, but parents there did not implement the focal feed. So any month you see attached to stopping is somebody's opinion, often a commercial sleep consultant's, which helps explain why two websites can disagree by three months.

What the official sources do give you is the backdrop. The NHS says that for some babies aged six months to a year night feeds may no longer be necessary, and the word doing the work in that sentence is “some.” The AAP's feeding guidance describes feeds spacing out gradually as babies grow, and says that by around six months a baby may take 6 to 8 ounces at each of 4 or 5 feedings in 24 hours, with solid foods starting around then too. Neither source turns those observations into an instruction about the late-evening feed. If you want to know whether your baby still needs calories at night, that is a weight-and-growth question, and the pediatrician can answer it with the growth chart in front of them.

Without a rule, the signals are practical rather than age-based. The feed stops being sleepy and starts waking her up properly. She takes very little of it. The first waking of the night stays at the same time whether or not she has had the feed, which suggests it is not protecting anything. Or she begins sleeping through to a time you can live with and the feed becomes the only reason anyone is up at 10:30pm. None of those needs an age to recognise.

How people stop varies too, and here I can only describe, not recommend. Some families drop it in one evening. Others make it gradually smaller or earlier over several nights before stopping. No source I found compares the two, so I am not going to pretend one is better. If she is formula fed, a smaller bottle is easy to measure; if she is breastfed, a shorter feed is the equivalent, though it is harder to know exactly how much she took. If anything about her feeding or weight is being monitored, ask before you change it.

It is easy to treat “still has a dream feed” as a sign of being behind. It is not. The research behind the idea is thin in both directions, so there is no standard to be behind. The when do babies sleep through the night page explains why the ages you see quoted vary so much in the first place: studies use different definitions of the phrase.