28  Do you need a check-up every six months?

Almost everyone in Britain believes they should see a dentist every six months. It is one of the most successful pieces of health messaging ever devised, and nobody can quite say where it came from. The best available accounts trace it to mid-twentieth-century convention and to marketing, not to research.

It has now been tested. This chapter is about what happened, and about the odd fact that the guideline has not caught up with it.

The advice

Assign a recall interval ranging from 3 to 24 months, based on oral health needs and disease risk

Strength of recommendation: Conditional

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

For children the range is 3 to 12 months.

Notice what this actually says: it does not say six months. It says risk-based, anywhere from three months to two years. The guideline has already moved away from the six-month rule; it is the profession and the public who have not.

What the guideline says its evidence is

Recommendation based on moderate certainty evidence that recall interval can be varied on individual basis without negative effects for adults who regularly attend dentist. Evidence not available for ‘hard-to-reach’ adults or those with more complex presentations. Recommendation in line with NICE (CG19).

— Delivering Better Oral Health, chapter 13, table 6 (1)

Two things to note before going on, because I got both wrong in an earlier draft and the chapter’s argument changed when I fixed them.

First, that statement is the adult one. The children’s rows, for the 3 to 12 month range, carry a different sentence: “Recommendation based on very low certainty evidence. Recommendation in line with NICE (CG19).” I had quoted the children’s evidence against the adult recommendation, which made the guideline look far more pessimistic than it is. The rest of this chapter is what remains once the right sentence is in place.

Second, the adult statement is not a bare deferral to an old guideline. It states the trial’s own finding, that the interval can be varied without negative effects; it names the population that finding applies to, regular attenders; and it names the population it does not, hard-to-reach adults and complex presentations. NICE CG19 is Dental checks: intervals between oral health reviews (2), published in 2004, but the row cites it alongside the 2020 Cochrane review of the INTERVAL trial, not instead of it.

Following the citation, and finding it is out of date

The Cochrane review DBOH cites alongside NICE is Fee and colleagues, updated in 2020 (3). It includes two studies with data from 1,736 participants. The larger is the INTERVAL trial, conducted in UK general dental practices in adults who were regular attenders, comparing 6-month, 24-month and risk-based recall, with outcomes at four years.

Here is what it found.

Comparison Outcome Effect 95% CI n Certainty
Risk-based vs 6-month tooth surfaces with any caries MD 0.15 −0.77 to 1.08 1,478 high
Risk-based vs 6-month % sites with gingival bleeding MD 0.78% −1.17% to 2.73% 1,472 high
Risk-based vs 6-month quality of life (OHIP-14) MD −0.35 −1.02 to 0.32 1,551 high
Risk-based vs 6-month moderate-to-extensive caries RR 1.04 0.99 to 1.09 1,478 moderate
24-month vs 6-month tooth surfaces with any caries MD −0.60 −2.54 to 1.34 271 moderate
24-month vs 6-month quality of life (OHIP-14) MD −0.24 −1.55 to 1.07 305 high
Table 28.1: Recall intervals in adults, four-year follow-up (3)

Read the certainty column, and then read the sample sizes next to it, because the two halves of this table are not the same strength of finding.

The risk-based versus six-month rows are high certainty on about 1,500 people. That comparison is as solid as anything in this book.

The 24-month versus six-month rows are not the same. Only patients whose own dentist had already judged them suitable for a two-year interval were eligible for that randomization, so the comparison rests on 271 participants drawn from a pre-selected group, and Fee and colleagues rate the caries outcome moderate rather than high, downgrading for imprecision: the interval includes little to no difference but also a difference larger than the minimal important difference. I had this row marked high in an earlier draft, which overstated it.

This is, along with Chapter 19, one of only two places in this entire book where I have found high-certainty evidence, and both of them are findings of no important difference. Adults who attended every six months did not end up with fewer decayed surfaces, less gum bleeding, or better oral-health-related quality of life than adults on risk-based or two-yearly recall. The intervals are narrow and centered near zero, which as Chapter 19 explains is a real answer rather than an empty literature.

The gap that is left

Set against that table, DBOH rates the adult row moderate. Fee and colleagues rate the risk-based comparison high for three patient-relevant outcomes. So there is a gap, but it is one rating wide rather than the chasm I first described, and the guideline gives its reason in the same sentence.

The reason is the second half of the statement: “Evidence not available for ‘hard-to-reach’ adults or those with more complex presentations.” That is a real limit and DBOH is right to record it. INTERVAL recruited adults who had seen a dentist within the previous two years, in general practice, with mouths in reasonable order. The trial establishes that lengthening the interval does not harm those people. It says nothing about the people who worry a public health body most, and a recommendation written for the whole adult population has to carry the weaker of the two.

Whether that justifies moving high down to moderate for the population the trial did cover is a judgment I would make differently, and it is a narrow disagreement rather than a finding of neglect. What I would want is the distinction stated: high certainty for regular attenders, no evidence for anyone else, rather than a single moderate covering both.

The children’s rows are a separate matter and the very low rating there is plainly right: the only paediatric evidence is one small Norwegian study.

Who was in these trials? Adults who were regular attenders, defined as having seen a dentist at least once in the previous two years, in UK general practice. Children in the other study were also regular attenders, in a Norwegian public dental service.

So the finding is: among people who already go to the dentist and have reasonably looked-after mouths, going twice a year rather than less often does not improve outcomes over four years.

It says nothing about people who do not attend, who have untreated disease, or who are at high risk. For them the whole question is different, and the answer is probably that they need to be seen more, not less.

What was not measured. Neither trial assessed adverse effects of different recall intervals, and the review says so. The review does not say that four years is too short; that caveat is mine, and I should not have attributed it. If anything Fee and colleagues lean the other way, concluding that “further studies comparing dental recall intervals for adults in primary care seem unnecessary, given the moderate to high certainty of the evidence for the outcomes of this review.” They do call for trials in children, and for work on potential harms such as overtreatment.

PubMed, 21 August 2026, for updates to CD004346. The 2020 version (.pub5) remains current.

I have the full text of the Cochrane review, its characteristics of included studies, the INTERVAL main trial report and the NIHR HTA monograph, all supplied during writing. The HTA monograph contains the cost analysis, which I have not audited.

Full record in appraisals/searches/ch28-recall.md.

Verdict

Certainty of evidence

High that, for adults who attend regularly, risk-based recall produces no important difference from 6-month recall in caries, gingival bleeding or oral-health-related quality of life over four years (n ≈ 1,500). Moderate for the 24-month comparison, which rests on 271 participants their own dentist had already judged suitable for a two-year interval. Very low for children, from one small study.

DBOH rates the adult row moderate. For the risk-based comparison in regular attenders I would say high; for the population as a whole, where the trial is silent, DBOH’s caution is defensible.

Directness to the advice as worded
Excellent for the six-month question, poor for the risk-based question. The trial compared exactly the intervals people argue about, with patient-relevant outcomes, over four years, in general practice. What it cannot tell you is how to assign an individual to a point on a 3-to-24 month range, which is what the recommendation actually asks clinicians to do.
Is the strength label defensible?

Conditional is right, and the certainty rating is defensible but blunt. This is the one place in the book where my complaint runs the other way: not that a label overstates its evidence, but that a single moderate is asked to cover a population in which the evidence is high for one part and absent for another. The remedy is the same as everywhere else in this book, which is to say which part the rating belongs to.

What I would still press on is the effect downstream. Nothing in the tables a patient or a practice ever sees records that six-monthly recall has been tested against risk-based recall in a four-year randomized trial and found to make no important difference for regular attenders. That is a strong, reassuring, publicly useful finding, and it is not reaching anyone.

What would change my mind
Evidence that four years is too short and differences emerge later. Evidence in irregular attenders or high-risk groups, where I would expect the answer to reverse. And any trial actually testing risk-based assignment tools against each other, which would address the question DBOH’s rating seems to be about.

What this means for you

If you are an adult who goes to the dentist reasonably regularly and your mouth is in decent shape, there is high-certainty evidence that a schedule your dentist sets based on your actual risk works as well as going every six months. Over four years it produced no measurable difference in decay, bleeding gums, or how your mouth felt.

The national guideline already says this: 3 to 24 months, based on need. Six months is not in it.

Three caveats, and the third is the one I nearly left out.

If you have untreated decay or gum disease, or you rarely attend, none of this applies to you; the trials were in people who were already doing fine. In England, whether you can get an appointment is a more pressing constraint than how often you should want one.

And the same trial that found no clinical difference also found that patients “greatly value and are willing to pay for frequent dental check-ups” (4). That is not a nuisance finding to be argued away. Under GRADE, what people value is one of the four inputs to a recommendation, on the same footing as the effect estimate. So the honest statement is narrower than “you do not need to go every six months.” It is that the clinical case for a six-month default in regular attenders has not survived testing, while the reasons a person might still want to go that often, reassurance among them, are untouched by any of this and are a legitimate basis for choosing.