 |
 |
Well done for having ago, Rohini
Just further on the numbers part, I really don’t like NNTs and NNHs for two main reasons
1.
There’s lots of evidence that people find it hard to compare rates with different denominators; say, an NNT of 7 and an NNH
of 9. NICE CG138 (what Neal would like to call the ‘kindness guideline’) recommends using a consistent denominator
2.
It’s not just about the absolute difference. Consider an NNT of 34 (ARR 3%, NNT rounded up) for avoiding death due to some
unpleasant condition over the next 5 years. How one interprets that is very different if that is actually a reduction from
a.
100% mortality to 97% mortality (not a huge absolute difference, but now hope where there was none before)
b.
52% mortality to 49% (still roughly half of all people die and half live)
c.
3.001% to 0.001% (a small but important risk almost abolished)
To quote the Montgomery judgement ‘The assessment is therefore fact-sensitive, and sensitive also to the characteristics of
the patient’
IMHO It’s much better to use natural frequencies and say (in the example given)
On days 14–42, and on average:
·
With hormonal therapy alone, 43 infants in 100 had spasms and 57 did not
·
With hormonal therapy plus vigabatrin, 28 infants in 100 had spasms and 72 did not
Also, of course, one would need to make clear that it’s not possible to know if any individual will be one of the 15 in 100 who benefit
or one of the 85 in 100 for whom adding vigabatrin makes no difference to what would have happened anyway
Chris Cates’s NNT online (http://www.nntonline.net/visualrx/) will do the calculations
and give you a nice graphic
Best wishes
Andy
Andy Hutchinson
Medicines Education Technical Adviser
Medicines and Technologies Programme
National Institute for Health and Care Excellence
Level 1A | City Tower | Piccadilly Plaza | Manchester M1 4BT
Tel: 07824 604962
Web:
www.nice.org.uk/medicines
email:
[log in to unmask]">[log in to unmask]
From: Evidence based health (EBH) [mailto:[log in to unmask]]
On Behalf Of Neal Maskrey
Sent: 14 June 2018 12:13
To: [log in to unmask]
Subject: Re: Number needed to treat
Yes, but conventionally the NNT would be rounded up to 7. Of course this is a short term result, maximum follow up period is 6 weeks. Whether this translates into similar, better or worse longer term benefits seems to be unknown, and it
seems to me that content-area experts would need to make a judgement about whether these findings would justify a change in routine being discussed with individual patients / parents before any longer term follow up results were available.
Best wishes
Neal
Neal Maskrey
Visiting Professor of Evidence-informed decision making, Keele University
ADVOCATE Field Studies, University of Amsterdam; a Horizon 2020 European Oral Health Project www.advocateoralhealth.com
Mobile: 07976276919