Commentary

This recently published systematic review by Arora et al. adhered to the standard methodological processes used by Cochrane and it included electronic searches of some databases as recently as mid-March 2017. The authors placed no restrictions upon language or publication status. Children and adolescents aged 3-19 years attending school from any country were eligible for inclusion. Randomised controlled trials (RCTs) were included whether the unit of randomisation was individual children or a group (eg a school or class). Data extraction followed the Cochrane Handbook for Systematic Reviews of Interventions.1 An assessment of risk of bias in the included studies resulted in a low, unclear or high risk of bias judgement across a number of domains.2

From the 2,238 records initially identified, 25 full-text articles were assessed for eligibility. Six RCT studies ultimately satisfied the inclusion criteria and were incorporated in the quantitative synthesis (meta-analysis). The six studies varied considerably in relation to how test-positive children were identified, followed up and referred. The follow-up period of the included trials was less than two years so the authors were only able to report the short-term effects of school dental screening. A further limitation is that the main outcome measure was reported attendance or registration with a dentist. Only one study included the prevalence of dental caries per child and associated clinical variables.3

Differences in the screening interventions meant that only four studies could be included in a meta-analysis assessing ‘traditional screening versus no screening’ with dental attendance as the outcome. In this case, an inconclusive result was found with very low-certainty evidence. The high heterogeneity may be partly explained by the fact that one study was an individual level RCT and the other three were cluster RCTs.

Criteria-based screening (the referring dentist using pre-established criteria) showed a 7% relative increase in dental attendance compared to no screening, but the confidence interval ranged from a 1% decrease to a 16% increase and there was low certainty of the evidence underlying the effect estimate. There was no evidence of a difference between criteria-based screening compared to traditional screening.

Within this systematic review, one of the included papers studied the effectiveness of different types of referral letter used following screening, but it is worth highlighting that this paper focused upon their effectiveness following orthodontic screening.4 A specific (personalised) screening letter was preferred by participants to a non-specific version. The risk ratio (RR) was 1.39 (95% CI; 1.09-1.77) indicating a 39% relative increase in attendance to a general dentist in the specific referral group versus the non-specific group, but there was low certainty of the underlying evidence. Elsewhere within this review, Hebbal (2005) found that traditional school screening with additional motivation (eg oral health education) compared to traditional school screening alone, led to a 208% relative increase in dental attendance RR 3.08 (95% CI; 2.57-3.71) again, with low certainty evidence.5 However, the improved attendance in the group with additional motivation cannot necessarily be associated to the effect of screening per se.

There are a number of limitations with the studies included in this systematic review and identified by the authors. 1) The included trials were all short-term and it is not known if the benefits of screening continue beyond this. 2) ‘Dental attendance’ as an outcome measure does not provide evidence that better oral health was achieved in those referred. 3) There is a need for standardisation of terminology (for example, when authors discuss using ‘specific’ referral letters and ‘criteria-based’ screening) to make interventions more transparent for study comparisons and for readers. 4) None of the RCTs reported data on any adverse effects of school dental screening. 5) None of the RCTs assessed the cost-effectiveness of the interventions used. 6) The conduct and reporting of clinical trials should adhere to CONSORT group guidelines.6

It is reassuring though not surprising that the key results of the meta-analysis by Arora et al. published in late 2017 are in agreement with the earlier systematic review by Joury et al. published online in late 2016.7 (My commentary on the Joury et al. paper appeared in EBD in October 2017).8 There is no evidence of improvement in dental attendance or reduction in dental caries between ‘screening’ and ‘no screening’ groups, despite a slight difference between the studies included in each systematic review. Both systematic reviews found low to very low certainty in the underlying evidence. Beyond this, it is not clear if improved dental attendance in those screened positive translates to improved oral health.