Evidence NNT Vetsurgeon.org

Re: Evidence NNT Vetsurgeon.org

J G Wray — Thu, 11 Oct 2012 22:58:32 GMT

[quote user="Malcolm Ness"]No you can't, for NNT you really need an untreated control group. [/quote]

Weeeell, that's what I always thought, then I recalled the peptic ulcer studies in Table 2 of the first reference I gave.

Whaddyathink?

JGW

Re: Evidence NNT Vetsurgeon.org

Malcolm Ness — Thu, 11 Oct 2012 18:49:07 GMT

[quote user="J G Wray"]

You know where this is going Mr N... what about newer simpler techniques?

Your centre could use a control group of TPLO patients for example.

[/quote]

No you can't, for NNT you really need an untreated control group.

Re: Evidence NNT Vetsurgeon.org

J G Wray — Thu, 11 Oct 2012 17:57:27 GMT

[quote user="Malcolm Ness"]Aspirational but not really a goer for a great number of reasons.[/quote]

You know where this is going Mr N... what about newer simpler techniques?

Your centre could use a control group of TPLO patients for example.

[quote user="Malcolm Ness"]Finally, and particularly in the field of surgery, NNT calculations can be nothing more than a complex way of discovering that some surgeons get better results and fewer complications than others - in effect, the stats tell you more about the surgeon than the surgery.[/quote]

'Twas ever thus. It would certainly help non-referral vets make a better choice for their patients and clients and allow better surgeons to charge a premium.

JGW

Re: Evidence NNT Vetsurgeon.org

Malcolm Ness — Thu, 11 Oct 2012 17:37:21 GMT

Aspirational but not really a goer for a great number of reasons.

First, Number needed to treat in isolation is not terribly useful - you need to know number needed to harm (effectively and expression of complication/risk rate) Second, you need a control group for comparison and controlled clinical studies which, in this context must be prospective, randomised and very tightly protocolled are essential. The difficulties associated with the execution of such studies are the real reason that we don't have ready access to these statistics. Added to that we need very tight, objective inclusion criteria and tight, objective outcome measures (which generally, we don't have especially in vet surgery). Not least we need a reliable and sensible way of evaluating and classifying complications and therein lies a serious "rub". For example, an early attempt at NNT estimation for some cruciate ops lumped all complications in together giving equal statistical weight to a minor wound infection that resolved completely and a catastrophic complication that needed amputation. Finally, and particularly in the field of surgery, NNT calculations can be nothing more than a complex way of discovering that some surgeons get better results and fewer complications than others - in effect, the stats tell you more about the surgeon than the surgery.

Re: Evidence NNT Vetsurgeon.org

J G Wray — Thu, 11 Oct 2012 17:26:26 GMT

[quote user="Mark Holmes"]Can I make it quite clear that its the other one. [/quote]

Me too.

Jon Wray

Re: Evidence NNT Vetsurgeon.org

Mark Holmes — Thu, 11 Oct 2012 17:22:27 GMT

[quote user="Hannah Wynne Richards"]

What is NNT ? I REALLY HATE abreviations I don't understand

Wynne

[/quote]

Read the link

Re: Evidence NNT Vetsurgeon.org

Hannah Wynne Richards — Thu, 11 Oct 2012 17:13:56 GMT

What is NNT ? I REALLY HATE abreviations I don't understand

Wynne

Re: Evidence NNT Vetsurgeon.org

Mark Holmes — Thu, 11 Oct 2012 17:11:26 GMT

Can I make it quite clear that its the other one.