Hi Nathalie,
Good questions and I'm sure my answers may not be accepted by all, but with regards the issue of clinical guidelines, you have to create a hierarchy of evidence and then put that evidence into perspective of the population that you are referring to. For example, ESHRE is a European organization and a good number of the countries in your demographics use a government-sponsored IVF system. This may contrast markedly with other healthcare systems (e.g. US, Middle East, etc.). As a prime example, single embryo transfers are highly advocated in Europe not because they increase the live-birth rate (which they don't and are known to have a lower pregnancy and live-birth rate than double embryo transfers), but the main driving force behind their use is overall cost-effectiveness due to the high cost of caring and dealing with the complications of preterm labor (in this population mainly due to multiple gestations). Now back to your point, narrative reviews are not necessarily evidence-based. In this case, if I found a narrative review on the single embryo transfers published in Human Reproduction (ESHRE publication) then almost certainly it will highly support the use of SETs, while a similar review in a US publication (e.g. Fertility and Sterility) or Middle East journal (e.g. Middle East Fertility Society Journal) may have differing points of view on the same topic. Therefore, if you are making guidelines, you should almost always leave narrative reviews as the last source of evidence.
As for evaluating RCTs vs. systematic reviews of RCTs, you should be using both. A systematic review may tackle a common PICO (e.g. Clinical efficacy and adverse events rates of single vs. double embryo transfers in women undergoing assisted reproduction) may not take into account all the subtle differences in patient populations, geographic distributions, etc. Therefore even thought it does provide more robust evidence on the overall topic, the individual RCTs may provide more content specific information (e.g. Clinical efficacy and adverse events rates of single vs. double embryo transfers in women over the age of 40 undergoing assisted reproduction in Sweden). Therefore you are not double-counting, but should be transparent to let the reader know how many RCTs being evaluated are also part of a systematic review that is also providing evidence.
Consensus meetings, in my opinion, are 'super' narrative reviews. They don't have to be evidence-based, they can be politically motivated, motivated by sponsorship, or other 'unknown' factors and some have almost no transparency to how they came up with their final conclusions. Again, they are a good source of direction and a way to identify key trials/ reviews on the topic, but I wouldn't put too much emphasis on every word they dictate clearly their methods, and how decisions were made (assuming that they were evidence-based).
Now I am expecting the furry of all readers who have written narrative reviews and sat on consensus committee panels.
Hope this helps.
Ahmed
-----Original Message-----
From: Evidence based health (EBH) [mailto:[log in to unmask]] On Behalf Of Nathalie Vermeulen
Sent: October 26, 2011 8:30 AM
To: [log in to unmask]
Subject: which evidence to include..
Dear all,
As ESHRE, the European Society for Human Reproduction and Embryology, and I are new in writing evidence-based clinical guidelines, some issues have raised from the start of evaluating the evidence.
Here are some issues on which I would like to ask your opinion.
1. narrative reviews. Would you exclude all narrative reviews or could you include them at an evidence level similar to an expert opinion in case of limited evidence on a particular topic?
2. In case an RCT is already included in a "high quality" systematic (Cochrane) review, do you automatically exclude this RCT from your list of papers to evaluate, or do you check for additional information in the RCT, with the risk of mentioning the same results twice in your evidence table (separately and as part of the meta-analysis).
3. How do you deal with a "report on a consensus meeting"? Exclude or include at the evidence level of expert opinion?
Thank you in advance for your assistance!
Best wishes,
Nathalie
Dr. Nathalie Vermeulen
Research specialist
E.S.H.R.E.
Meerstraat 60
1852 Grimbergen
Belgium
Tel (general): + 32 (0)2 269 09 69
Tel (direct): +32 (0)2 263 64 61
Fax: + 32 (0)2 269 56 00
E-mail [log in to unmask]