You are here

Pil B's points on the ACD- reposted from string further down page.

I thought it worth reposting Phils summation of his reading of the ACD from the discussion lower down the page.

I agree totally with Phil's take on the model NICE have used as one that is not based on the real life experience... from clinicians/researchers in the field and patients experiences and responses to TKI's.

I know they have recognised this because the ACD concludes with a statement that they themselves (using their own model) would 'very likely' come to a different (more positive) conclusion should researchers conduct more robust RCT's.

In other words (and I am sorry to repeat myself) preferably:
4 arm double blind randomised trials with strict controls on patient cohorts and no moving between arms, no dose escalation/reductions.
Such trials are not possible and are unethical.
If such a trial were ethically acceptable,
how can you blind the drug indentities as they have differing side effects and in the case of nilotinib a completely different administration regime.
If the patient didn't work it out immediately then the clinicians certainly would.

Also how would they recruit to a trial that would randomise some to 'best supportive care' i.e no treatment? It is a nonsense but dressed up as Science.
The best scientists/researchers/clinicians (like Brian Druker) follow hunches after all ;o) and how do we possibly measure a hunch?

Sandy

Phil's post:

'I am having real problems with this...' Phil B

Having worked through the bulk of this, I have to say I'm finding their conclusions very strange indeed. Having rejected the manufacturers' own cost effectiveness analysis they appear to have gone out on a limb with their own analysis and ignored some very unlikely predictions it makes.
They seem to have:
1. predicted their overall survival curve based on MCyR response rates
2. calculated expected times on treatment from the study data
3. taken standard assumptions on length of AP and BP
4. assumed the balancing number is time in chronic phase without treatment after the treatment in question failed.

The results of the exercise are showing time to progression from chronic phase WITHOUT TREATMENT for the HD Imatinib and Nilotinib arms of 7 to 8 years on average. I can't see how this can possibly make sense given that TOTAL life expectency pre TKIs was only 3 to 5 years from diagnosis.

The conclusion that Dasatinib is not cost effective arises ONLY because of the effect of this 'calculated' length of treatment-free chronic phase being so much longer in their HD Imatinib / Nilotinib models, but the fact the the figure produced makes no sense at all in comparison with empirical data (and thus their model clearly doesn't accurately model reality) hasn't been mentioned.

On top of ignoring all the best scientific data, there are a few more sneaky things they've done. In particular they have rigged the model so that a very large proportion of patients die of non-cml causes before or not long after they would have succumbed to CML and therefore getting minimal benefit to set against the cost.

They have done this by:
1.assuming all the patients are 60 years old to begin with;
2 exaggerating the survival times without treatment; and
3.modelling survival times on treatment as a smooth distribution tailing off rapidly when what you actually have is distinct populations of non-responders, poor responders and good responders with the bulk of this last group effectively CML-free with normal life expectancy.

If you instead model say a 40 year old with resistance due to a single kinase mutation which we know dasatinb / nilotinib works against then your best estimate of their survival time would be in line with Dr Druker's 30 year estimate for Glivec responders. The bulk of this would be AFTER the patent expires and therefore on a cheap generic. The cost per year gained is then very favourable indeed.

You could say they have ignored patent expiry and exaggerated patient expiry!

The other big thing missing, as already mentioned below. is the whole issue of mutation testing, drug level testing, etc. which in a high proportion of cases lets the doctors target the drugs to where they will do most good and greatly improve value for money.

Best wishes all

Phil

Phil, it is great to have your input and it is very much appreciated.
Hope your vicious cold is improving. I am almost back to normal ;o)
I am finishing a letter to DH Health ministers... which must be very carefully worded and targeted so that they actually take us seriously. I plan to add an appendix to the letter and will be in touch with you about this.

best.......Sandy