15  The quieter tier

Most of this book has been about Strong recommendations, because that is where the gap between label and evidence does the most damage. But of the 91 recommendations in Delivering Better Oral Health, only 28 are Strong. The other two thirds are Conditional or Good practice, and they are worth a chapter of their own, because they are where the guideline is at its most honest and where nobody is looking.

This chapter takes three of them together.

Supervised toothbrushing

Parent or carer to assist and supervise toothbrushing if required

Strength of recommendation: Good practice

— Delivering Better Oral Health, chapter 2, table 1e (1)

DBOH cites dos Santos and colleagues’ systematic review of supervised toothbrushing (2). Its conclusion, in the authors’ words:

There is no conclusive [evidence] …

— dos Santos and colleagues (2)

The review searched 2,046 records and found four controlled trials. It could not perform a meta-analysis, because of clinical heterogeneity among the studies and differences in how caries incidence was reported. Of the four trials, two found statistically significant differences favoring supervised brushing, but “information about the magnitude and/or the precision of the effect estimate was lacking”, and in one trial the clustering effect was not taken into account.

That last detail is worth dwelling on. A trial that randomizes schools but analyses children as though they were independent will produce confidence intervals that are too narrow and a p-value that is too small. It is one of the most common statistical errors in this literature, and the review caught it.

So: four trials, no poolable estimate, effect sizes not reported, one analysis with a known error. The right label for that is Good practice, and that is the label it has. The guideline is not overclaiming; it is describing a genuine gap.

The thing worth noticing is which gap. Supervised brushing is one of the most widely implemented public health interventions in children’s dentistry. Childsmile in Scotland is built on it. And nobody has established its independent effect, as distinct from the effect of the fluoride toothpaste it delivers. The review’s stated aim was exactly this: “The anticaries effect of supervised toothbrushing, irrespective of the effect of fluoride toothpaste, has not been clearly determined yet.”

Sugar at bedtime

Avoid sugar-containing foods and drinks at bedtime when saliva flow is reduced and buffering capacity is lost

Strength of recommendation: Conditional (Good practice for under-3s)

— Delivering Better Oral Health, chapter 2, tables 1a to 1f (1)

Recommendation based on very low certainty evidence for increased risk of dental caries associated with bedtime consumption of food and drinks containing free sugars (in children aged 3 years and older).

— Delivering Better Oral Health, chapter 13 (1)

The citation is Baghlaf and colleagues (3). Eighteen studies met the inclusion criteria: 13 cross-sectional, 4 cohort, 1 case-control. Six were rated good quality, eight fair, four poor. And:

It was not possible to conduct a meta-analysis, because of the considerable variations [in the studies].

— Baghlaf and colleagues (3)

No pooled estimate, mostly cross-sectional data, very low certainty. Conditional is the correct label and DBOH applies it.

I find this recommendation the most persuasive of the three despite having the weakest formal evidence, and it is worth saying why, because it illustrates something Chapter 2 argued in the abstract.

The mechanism here is the same one as Chapter 6, and it is the same one that makes the Chapter 14 night-feeding signal plausible: salivary flow falls during sleep, so anything left on the teeth stays there, unbuffered and uncleared, for hours. Three separate recommendations in this guideline lean on that single physiological fact. None of them has been tested directly. They are all the same untested assumption wearing different clothes, and their apparent independent agreement is an illusion.

That is not a reason to disbelieve it. It is a reason to notice that a guideline can look better-corroborated than it is, when several of its recommendations rest on one shared premise.

Sugar-free medicines

Use sugar-free versions of medicines if possible

Strength of recommendation: Good practice

— Delivering Better Oral Health, chapter 2, tables 1a to 1g (1)

Chapter 13 offers no evidence statement for this at all. It is one of the 39 recommendations in the guideline whose evidence statement names no certainty level.

And that is fine. This is what the Good practice category is for: advice that is obviously sensible, costs nothing, harms nobody, and that no one is ever going to randomize. A child on long-term liquid medication is bathing their teeth in sugar several times a day. You do not need a trial.

I include it to make a point about proportion. This book has spent a lot of pages on recommendations whose labels outrun their evidence. Here is one with no evidence at all, correctly labeled, and I have nothing to say against it. The category works when it is used.

PubMed, 21 August 2026, for updates to the three reviews cited. None found: dos Santos 2018 and Baghlaf 2018 both remain current, and no systematic review of sugar-free medicines and caries was identified.

I have full text for dos Santos 2018 and Baghlaf 2018 including supplements, supplied during writing.

Full record in appraisals/searches/ch15-quieter-tier.md.

Verdict

Certainty of evidence
Supervised brushing: very low. Four trials, no meta-analysis possible, effect sizes unreported, one analysis ignoring clustering. Bedtime sugar: very low. Eighteen studies, mostly cross-sectional, no meta-analysis possible. Sugar-free medicines: no evidence, and none needed.
Directness to the advice as worded
Reasonable throughout. These reviews are about the interventions the advice describes, in the right people. The problem is quantity and quality of evidence, not relevance, which makes this part of the guideline a cleaner case than most of Part II.
Is the strength label defensible?

Yes, for all three, and this is the guideline working as designed. Very low certainty gets Conditional. No rateable evidence gets Good practice. Nothing is dressed up.

The observation that survives is not a criticism of any single row. It is that the bedtime recommendation, the last-thing-at-night recommendation and the night-feeding caveat all rest on one untested physiological premise about salivary flow, and read as three independent supports when they are one.

What would change my mind
For supervised brushing: a cluster trial, correctly analyzed, isolating supervision from toothpaste supply. Given how much public money rests on this intervention, its absence is remarkable. For bedtime sugar: any prospective study measuring timing of intake rather than reconstructing it from a questionnaire.

What this means for you

Supervise your child’s brushing until they are old enough to do it properly. Nobody has proved this helps independently of getting fluoride onto the teeth, but getting fluoride onto the teeth is the point, and a six-year-old left alone generally does not.

Do not send children to bed on sugar. The formal evidence is very low certainty and the mechanism is untested, but it is free, and the reasoning is the same one that underpins several other pieces of advice in this book.

Ask for sugar-free medicine if your child is on something long-term. This is the easiest recommendation in the guideline to follow and there is no argument against it.