4 ms·

And in large part because FDA multi-stage trials are very expensive to get through and the risk of failure is high. I know it is popular to hate on the pharmacy

And in large part because FDA multi-stage trials are very expensive to get through and the risk of failure is high. I know it is popular to hate on the pharmacy companies right now - and deservedly so - but the research costs are not totally made up like you are implying.

Zigurd 3y ago
Maximized is not "totally made up." You will find software companies maximizing R&D tax credits, too, for example. A JHU study found median costs for clinical trials of $19 million. That's expensive relative to software. If I piled HIPAA, payments processing, and whatever telcos might require for hooking up to SS7 in one app (call it "X") I doubt I could get halfway to $19 million. But $19 million is peanuts compared to pharma revenue. I stand by the assertion that pharma gets undeserved sympathy for long patent terms.
- kbenson 3y ago
  > I stand by the assertion that pharma gets undeserved sympathy for long patent terms. Sure. That's an assertion that's easy to stand by because it's so easy for it to be true because it's so open ended and unbounded. I think most people believe the pharma companies are playing it up some and getting some undue sympathy, but 10x the sympathy and 10% more sympathy are vastly different situations and your assertion is "true" for both. It's less a question of whether that assertion is true, and more about to what degree, because that's what really matters in a discussion like this. A statement about the median cost of a single study across an industry compared to pharma revenue (across an industry, across all studies done, whether resulting in a successful product that can make it to market and earn revenue), is not something we can make useful assertions from, whether or not I believe your assertion to be true (I do). It's just throwing disjointed numbers around and using implications instead of data to link them, which isn't a useful way to convince people who want to understand what's going on.
  - Zigurd 3y ago
    How is an industry-wide median cost "disjoint?" What number would better represent typical clinical trials costs?
    - kbenson 3y ago
      Is it the median cost for a trial or for all trials that a company does? A company might have many trials going on at once, of greater or lesser costs, but revenue is an aggregation of everything over a year. A company with $100 million in revenue and 10 trials ongoing and a company with a $1 billion in revenue and 10 trials ongoing would possibly lead one to specific conclusions, but even when it's broken down that much it doesn't really mean anything for the examples presented, because the mean cost of the trials that the company in question has ongoing could be entirely different. The 10 trials the $100 million revenue company is administering could average $1 million, and the 10 trials the $1 billion company is administering could cost 80 million each, by nature of the types of drugs they and markets they are aiming for. It could also be that there's lots of low cost trials around $20 million, but slightly less than half of the trials that are done are very expensive, and 4-5 times the cost. This would be more obvious if there were values for bot mean and median, instead of just median. The looser the data, the less you can actually infer. Any time someone tells me the median cost for something across an entire industry, my immediate reaction is to wonder what that means, not assume I actually know what it means, because it's far too little data to really know. How many trials are done in a year on average? How many a year do companies average based on revenue? Or simple, what's the actual amount spent on trials compared to revenue, and over how many years are those trials run? Is yearly cost of a trial really a good indicator, if trials take 5-10 years to run, so a failed one results in not just loss of money spent but opportunity cost over many years? Note: If you provided a reference for that median number I would have actually looked it up to see if it made sense to me, or if I had additional questions about how they went about their measurements. Instead, all I have a questions about what the numbers mean because I have no context.
      - Zigurd 3y ago
        Here is a link to the JHU announcement: https://publichealth.jhu.edu/2018/cost-of-clinical-trials-for-new-drug-FDA-approval-are-fraction-of-total-tab https://publichealth.jhu.edu/2018/cost-of-clinical-trials-fo... You will like some of the numbers on that site even less than the $19 million median cost for a trial. The study was published in JAMA, link on the JHU page. I hope this does not sound like an appeal to authority, but good luck discrediting the numbers.

Zigurd 3y ago

Maximized is not "totally made up." You will find software companies maximizing R&D tax credits, too, for example. A JHU study found median costs for clinical trials of $19 million. That's expensive relative to software. If I piled HIPAA, payments processing, and whatever telcos might require for hooking up to SS7 in one app (call it "X") I doubt I could get halfway to $19 million. But $19 million is peanuts compared to pharma revenue. I stand by the assertion that pharma gets undeserved sympathy for long patent terms.

kbenson 3y ago
> I stand by the assertion that pharma gets undeserved sympathy for long patent terms. Sure. That's an assertion that's easy to stand by because it's so easy for it to be true because it's so open ended and unbounded. I think most people believe the pharma companies are playing it up some and getting some undue sympathy, but 10x the sympathy and 10% more sympathy are vastly different situations and your assertion is "true" for both. It's less a question of whether that assertion is true, and more about to what degree, because that's what really matters in a discussion like this. A statement about the median cost of a single study across an industry compared to pharma revenue (across an industry, across all studies done, whether resulting in a successful product that can make it to market and earn revenue), is not something we can make useful assertions from, whether or not I believe your assertion to be true (I do). It's just throwing disjointed numbers around and using implications instead of data to link them, which isn't a useful way to convince people who want to understand what's going on.
- Zigurd 3y ago
  How is an industry-wide median cost "disjoint?" What number would better represent typical clinical trials costs?
  - kbenson 3y ago
    Is it the median cost for a trial or for all trials that a company does? A company might have many trials going on at once, of greater or lesser costs, but revenue is an aggregation of everything over a year. A company with $100 million in revenue and 10 trials ongoing and a company with a $1 billion in revenue and 10 trials ongoing would possibly lead one to specific conclusions, but even when it's broken down that much it doesn't really mean anything for the examples presented, because the mean cost of the trials that the company in question has ongoing could be entirely different. The 10 trials the $100 million revenue company is administering could average $1 million, and the 10 trials the $1 billion company is administering could cost 80 million each, by nature of the types of drugs they and markets they are aiming for. It could also be that there's lots of low cost trials around $20 million, but slightly less than half of the trials that are done are very expensive, and 4-5 times the cost. This would be more obvious if there were values for bot mean and median, instead of just median. The looser the data, the less you can actually infer. Any time someone tells me the median cost for something across an entire industry, my immediate reaction is to wonder what that means, not assume I actually know what it means, because it's far too little data to really know. How many trials are done in a year on average? How many a year do companies average based on revenue? Or simple, what's the actual amount spent on trials compared to revenue, and over how many years are those trials run? Is yearly cost of a trial really a good indicator, if trials take 5-10 years to run, so a failed one results in not just loss of money spent but opportunity cost over many years? Note: If you provided a reference for that median number I would have actually looked it up to see if it made sense to me, or if I had additional questions about how they went about their measurements. Instead, all I have a questions about what the numbers mean because I have no context.
    - Zigurd 3y ago
      Here is a link to the JHU announcement: https://publichealth.jhu.edu/2018/cost-of-clinical-trials-for-new-drug-FDA-approval-are-fraction-of-total-tab https://publichealth.jhu.edu/2018/cost-of-clinical-trials-fo... You will like some of the numbers on that site even less than the $19 million median cost for a trial. The study was published in JAMA, link on the JHU page. I hope this does not sound like an appeal to authority, but good luck discrediting the numbers.