Discussion
Our study shows that the number of Cochrane systematic reviews in paediatric dentistry and oral health has increased during the last few years. This situation should improve the basis for clinical decision-making because systematic reviews are considered essential sources of evidence for guideline development. The methodological quality of most of our reviews was high, corresponding to the high quality standards of the Cochrane Collaboration. Nevertheless, the likelihood of publication bias was not frequently assessed. This is an important factor to take into account in the conduct of a meta-analysis and in the interpretation of results.
Cochrane reviews should not define recommendations for practice because this requires assumptions about the relative importance of benefits and harms of an intervention and judgements that are beyond the scope of a systematic review. However, Cochrane review authors always propose implications for practice. Our study demonstrated that most of the reviews (43%) and all interventions supported by research evidence focused on the prevention of dental caries. For children and adolescents, topical fluoride treatments (with toothpaste, gel or varnish) were found effective for permanent and deciduous teeth and sealants for occlusal tooth surfaces of permanent molars. The predominance of this topic seems justified because it is the most important from a public health policy viewpoint. Early childhood caries is the most frequent chronic disease affecting young children and is 5 times more common than asthma. The selected reviews also concerned orthodontic treatment and oral surgery. However, for clinicians, several secondary research gaps are the management of oro-dental trauma or conservative treatments. Actually, the latter involve materials that may be harmful because of some toxicity.
Many of our reviews (78%) produced inconclusive evidence. The most common reasons for failure to provide reliable information to guide clinical decisions are the small numbers of RCTs and patients per meta-analysis. According to a cross-sectional descriptive analysis about characteristics of meta-analyses in the Cochrane Database of Systematic Reviews, the median number of RCTs included in meta-analyses was 3 (Q1–Q3 2–6) and the median number of patients was 91 (Q1–Q3 44–210). Our findings are consistent with these figures and emphasize that more high-quality primary research may be frequently needed to reach conclusiveness. However, none of the selected reviews was empty; that is, randomized evidence always existed and was included in the review, even when inconclusive. Another explanation for the inconclusiveness may be the inability to perform data synthesis. Diversity in outcomes measured across RCTs within a review may substantially limit the ability to perform meta-analyses and may explain the lack of recommendations. Many meta-analyses frequently exclude a large number of RCTs because outcomes are too different between studies. The standardization of outcomes was initiated by the OMERACT group and is expanding with the COMET Initiative. In the field of dentistry, some studies have defined core outcome sets to help solve this problem, such as in implantology and for the evaluation of pulp treatments in primary teeth. Finally, all systematic reviews should be considered as informative because they may allow for identifying well-informed uncertainties about the effects of treatments.
Previous methodological surveys assessed the conduct quality of systematic reviews in the field of dentistry. In a study of 109 systematic reviews published in major orthodontic journals, 26 were published in the Cochrane Database of Systematic Reviews. In all, 21% of the selected reviews satisfied 9 or more of the 11 AMSTAR criteria. However, to our best knowledge, no methodological survey concerned specifically pediatric oral health.
Our study has some limitations. Indeed, we considered only Cochrane systematic reviews in our study, but many non-Cochrane systematic reviews have also assessed interventions in the paediatric oral health field. Nevertheless, Cochrane systematic reviews are the highest standard in evidence-based health care. Moreover, Cochrane reviews have a standard structure, which always includes implications for practice. Another potential limitation is that we assessed whether the experimental intervention should be used in practice, should not be used in practice or should be used only in research based on the Implications-for-practice section only and we did not critically judge the review evidence ourselves. However, Cochrane review authors describe clinical implications only after describing the quality of evidence and the balance of benefits and harms.