Overall, we found that there is high consensus among institutions for recommending single preventive interventions, for which there is moderate to high evidence that such interventions are effective. However, for these single interventions, the lack of data or the low certainty of the evidence regarding doses, frequency or duration of supplementation results into less clear recommendations and ultimately lesser consensus between institutions. Indeed, while no institution is arguing that vitamin K is effective for preventing haemorrhagic disease of the newborn, national policies are recommending different regimens with regard to route, dose and frequency. This is likely to be explained by the limited data from studies that would directly compare different regimens. In this case, we acknowledge that prospective and comparative trials are challenging to be conducted in the future, due to the low incidence of the outcome and above all due to the severity of the outcome.
For SIDS and unintentional injuries, assessing the effectiveness of preventive interventions is difficult due to the nature of the outcome (death or severe morbidity) and the multiple causes involved. For obvious ethical reasons, no RCT can be performed to assess the impact of selected risk factors on the outcome. Evidence is therefore mainly based on retrospective case control studies and surveillance data. This leads to difficulties in differentiating associations from causality between risk factors and outcomes. Furthermore, the multi-causality of the outcome (several risk factors are associated with the outcome) adds difficulties in establishing the contribution of each risk factor. In addition to the association between risk factors and the outcome (SIDS or unintentional injuries), there is overall limited evidence assessing the effect of the recommendations through knowledge, attitude and practices.
The assessment of screening tools seemed to be more controversial than for the single or multiple interventions, with a higher degree of disagreement between institutions. This is likely to be explained by the fact that more aspects must be considered when evaluating the balance between benefits and harms of universal screening than of single interventions. For interventions, the main factors to assess and balance are the effectiveness of the intervention to prevent the undesirable outcome and potential adverse effects associated with the intervention. In contrast, for screening, there is a need to assess the effectiveness of the screening tool to detect the condition, the potential harms derived from the screening tool and treatment, and the treatment effectiveness. In addition, outcomes following early detection and treatment versus the time of presenting signs or symptoms must be considered. Furthermore, the significance and consequences of false positive and false negative cases vary between settings with different prevalence of the disease and with different health care system. As the complete set of criteria for a good screening programme that was adopted by the WHO needs to be assessed, economic and other health system related factors need consideration too [12].
Apart from the immediate findings, the approach we used for this evidence synthesis is valuable in several other ways.
First, providing both aspects (summaries of existing recommendations and evidence) transparently and independently informs clinicians, researchers and policymakers about important research gaps. Where are formal systematic reviews required and where is field research needed? When we reviewed preventive interventions, it became clear that field research through new trials is needed most. Examples are trials to elucidate the optimal duration of vitamin D supplementation for preventing rickets, studies to determine the benefits of early detection and treatment of children with ASD compared with those identified with developmental concern, or research to find a suitable screening tool for iron deficiency anaemia.
Second, the review proved to be a unique opportunity to summarize what is already known across ten controversial and highly important preventive interventions for child health in a single document. It is a crucial starting point when a group or institution intends to develop or update recommendations. Indeed, this approach not only allows the identification of research gaps, but also areas for which recent systematic reviews have already been conducted and areas for which no further evidence might be required. In addition, we noticed that sometimes several institutions committed their own systematic reviews in order to inform the development or update of a recommendation, leading to several groups conducting independent systematic reviews in parallel and addressing the same research question. This represents a potentially avoidable duplication of work, wasting resources.
Third, our approach may facilitate ownership of the prevention chapters in the pocket book by institutions such as paediatric societies. We systematically looked at existing recommendations from credible European institutions, national and international, but also non-European institutions, for generalizable and applicable conclusions. By taking into consideration their views and positions and by making our methods clear, the content of the final product may become more acceptable to the target audience. Moreover, presenting an overview of the existing recommendations also sheds light on how various European institutions use and apply existing evidence to their settings – common and different approaches and positions become visible.
Finally, we complemented the European-focused summary of existing recommendations with an independent body of evidence from relevant systematic reviews of the Cochrane library. Separating the evidence from the recommendations allows the editorial group to gauge how different institutions considered factors such as acceptability, feasibility, potential harms, cost and others.
This review has limitations. There was a limited timeline that was defined by the editorial group and editorial process of the WHO pocket book. To address this, a single author was identified for conducting the literature search, data extraction and data synthesis. When we found no systematic review addressing one of our key questions, or when a systematic review was found but was outdated, it was not possible to conduct a needed systematic review ourselves. In addition, it was not possible to assess the methodological quality of included systematic reviews and guidelines using the AMSTAR-2 tool and the AGREE tool respectively.
The strengths of this review are the alignment to the principles of good evidence synthesis for policy – inclusive, rigorous, transparent and accessible [13, 14]. This review is inclusive, comprising of a range of relevant sources for both available evidence and recommendations. We believe that findings are relevant and useful to policymakers, especially for guiding those involved in the health promotion in primary health care for the paediatric pocket book for the WHO European Region. Although these evidence summaries do not assess external validity, we focused on gathering relevant sources from the European or similarly applicable settings when it comes to recommendations. The editorial group is invited to take this into consideration when moving from evidence to recommendations. We used a systematic and rigorous approach for all the topics. The study was conducted by an independent author and peer-reviewed for ensuring good quality. We developed a protocol prior to the evidence synthesis, reported the methods we followed, and acknowledged the limitations of our work with transparency. Finally, timeliness is key, and despite the time constraint, this compilation of evidence synthesis was made available in due time for its use by the WHO paediatric pocket book panel.