In March we discussed this paper which examined whether there was an association between second hand smoking and meningococcal disease risk. We critiqued the paper together on 24th March, and below is a summary of the discussion.
Q.1 Were the aims of this study clear?
Overall we agreed the aim was clear. The authors had identified a gap that other reviews had not addressed; whether second hand smoking is associated with meningococcal disease, or not.
Q.2 Was the systematic review comprehensive?
It was noted that the search method employed by the authors meant that no grey literature was included, although this was probably because grey literature is unlikely to include comparators (which was a defined criteria for this review).
It was agreed that experiemental studies (such as randomised control trials) in this area were both unlikely to have occured due to it being unethical!
We had some discussion about publication bias assessment; some people felt this was good practice to try and quantify possible missing studies (e.g. by funnel plot), whereas others questioned the validity of the ‘trim and fill’ technique (which essentially balances the funnel plot by creating extra studies).
Q.3 Was the meta-analysis appropraite?
@Duncautumnstore queried whether combining different study types was appropraite for a meta-analysis, and suggested that a study group sub-group analysis may have been preferable if there were enough studies.
Q.4 Do you believe the results? Could anything else explain these findings?
@KateT_Health suggested that the authors’ conclusion of their results represented the ‘true effect of SHS on meningo’ was over-stated, particularly in light of the likely publication bias.
@AbbeyDiaz_says noted that the original studies did not account for counfounders such as “ethnicity, indigenous status, immunisation status, how much smoked/where smoked/other smking behaviours, other enviro factors.” This limits the definitiveness of the results.
However overall #PHTwitJC participants found the results convincing, and agreed that the evidence was compelling enough to suggest an association between smoking and meningococcal disease.
Q.5 What implications do the findings have for public health practice & policy?
It is clear that this study adds to the weight of evidence regarding the harms of smoking!
We discussed that pregnancy is prime time for bombarding expectant mothers with information, and were unsure whether this would be a sensible addition as the information and support available regarding smoking is already well established.
We explored whether this risk information would be useful during outbreak investigations, however @carotomes pointed out one would have to be careful to not suggest parents were responsible for the disease risk.
For further information on publication bias and other types of bias in population data, check out the Health Knowledge page here.
For further information on the Bradford-Hill criteria for causation, the Wikipedia page does a good overview here.