[HN Gopher] The case for clinical trial abundance
___________________________________________________________________
The case for clinical trial abundance
Author : apsec112
Score : 24 points
Date : 2024-12-19 16:16 UTC (3 days ago)
(HTM) web link (ifp.org)
(TXT) w3m dump (ifp.org)
| CharlieDigital wrote:
| I worked in this space for ~10 years and learned a lot.
|
| In 2011, the bootstrapped startup I was at built the first COTS
| trial site electronic document exchange portal. Up until then,
| trial teams would use a mish mash of shipping paper, fax, FTP,
| E-mail, etc.
|
| Our product was a finalist for the Transcelerate Shared
| Investigator Portal (SIP) project to unify this process on one
| platform.
|
| I saw the internal workings of IT teams in many large pharmas and
| fought to modernize the internal workings of our own team.
|
| I say all this as preface. The industry as a whole is highly
| political, incredibly conservative when it comes to tech, and has
| high barriers to innovation.
|
| I use this term "political" broadly to cover an array of issues I
| observed. One of which is that there a lot of jostling for
| influence because that's how consultants and vendors make money.
| A lot of folks with influence are long-time industry veterans
| with domain expertise, but no technical prowess paired with it.
|
| In the end of the Transcelerate SIP selection process, Veeva
| submitted a non-existent product (at the time) and ultimately
| Cognizant (the consultancy running the selection) built their own
| because of course! Millions of dollars in recurring revenue were
| on the line!
|
| Lots of good ideas have come and gone because the entrenched
| interests in this space extract rent from the friction. One
| working group I was a part of was iterating a spec as a
| spreadsheet instead of a technical format like a standard API...
|
| One point I particularly disagree with the OP is CT.gov. It is
| fine, IMO, because it is meant for industry and not laypersons.
| It has an excellent API[0] and there are plenty of third parties
| that build consumer-facing apps for trial matching. For example,
| I built a really basic one using LLMs to match natural language
| criteria to clinical trials and emails you when a matching trial
| pops up in the change feed[1]. The main problem with CT.gov right
| now is that it lacks a bit in data quality standards. There's a
| lot of variation in completeness of data dependent on country of
| origin and even between companies.
|
| The lack of technical standards in the industry really hurts
| innovation. For example, clinical trial protocols should arguably
| be a standard technical template. The FDA has done some work in
| this area[2], but industry has largely not adopted it. You can
| find public protocols attached to some trials on CT.gov, but
| every company has a different structure. Yet it is this document
| that underpins the design and execution of the trial (on top of
| being hard for trial sites to adapt between different sponsors).
|
| [0] https://clinicaltrials.gov/data-api/api#extapi
|
| [1] https://zeeq.ai
|
| [2] https://osp.od.nih.gov/policies/clinical-research#tab1/
___________________________________________________________________
(page generated 2024-12-22 23:00 UTC)