Thread by @VPrasadMDMPH

In our new paper in @JAMANetworkOpen we take a deep look into cost-effectiveness (CEA) studies of cancer drugs

Bottom line: If a CEA study is funded by pharma, it is 40x (OMG!) more likely to find the drug is cost effective

A 🧵 explaining what we found
jamanetwork.com/journals/jaman…

For every cancer drug indication approved between 2015-20, we searched for cost-effectiveness studies

we found between 0 and 9 per drug!!

Some trials were industry sponsored & others neutral

Here are the baseline characteristics of the studies we looked it.

Only 1/2 to 2/3 of drugs have even shown they improve survival

The rest have unknown effects on survival

That is not good enough

It is FDA failure! (these days that's common)

In descriptive analysis, 96% of industry sponsored studies find a drug is cost effective, while this is only true for 30% of non industry sponsored studies

Don't worry! I bet the authors of the 4% of studies sponsored by pharma finding it not cost effective have been fired.

In a multivariate model, being funded by pharma is associated with an odds ratio of 40!!! that a study concludes a drug is cost -effective

This relationship is MASSIVE

Tobacco as a lung carcinogen is only OR 20!!!

Finally, here is a waterfall plot of the incremental cost-effectiveness of drugs BEYOND the threshold

Some cancer drugs are 1 MILLION dollars per QUALY beyond the threshold!!!

This is madness!

We spend so much on drugs that either (a) don't improve survival or quality of life (b) we don't know if they do or (c) do so only marginally

No society can exist funding these drugs

It is untenable

It will bankrupt us all

They say Rome collapsed when half of all days were holidays; perhaps the same fate will happen in the USA when half of GDP is spent on cancer drugs that don't work or work marginally

Read the full paper here:
jamanetwork.com/journals/jaman…

& Follow @vkprasadlab for research updates