Subscribe
Share
Share
Embed
You don’t know where test data will lead, but you can follow it if you plan ahead.
A few weeks ago Andy and Jake Cowperthwaite spoke with Steve Schaefer, CEO of CoolTech, about designing your device as a platform. Steve’s company was able to use their temperature management platform to develop two devices, for two distinct applications.
They were able to achieve this by designing enough flexibility into the initial design, which is what Andy and Jake discuss in depth this time around.
Speed-to-data determines go-to-market success for medical devices. You need to inform critical decisions with user data, technical demonstration data, and clinical data. We interview med tech leaders about the critical data-driven decisions they make during their product development projects.
00:00:17:08 - 00:00:19:17
Welcome back to MedTech Speed to Data
00:00:19:17 - 00:00:21:00
a Key Tech Podcast.
00:00:21:00 - 00:00:22:03
I'm your host to Andy Rogers
00:00:22:03 - 00:00:23:17
along with Jacob Cowperthwaite,
00:00:23:17 - 00:00:25:18
Senior EE and Partner here at Key Tech.
00:00:25:18 - 00:00:26:23
Jake, welcome back.
00:00:26:23 - 00:00:27:12
Hey, Andy, it's
00:00:27:12 - 00:00:29:04
good to be with you again.
00:00:29:04 - 00:00:30:03
- Yeah.
00:00:30:03 - 00:00:31:03
Man, it seems like yesterday
00:00:31:03 - 00:00:34:14
we did this really interesting interview
00:00:34:14 - 00:00:37:02
with Steve Shaefer, CEO of CoolTech.
00:00:38:03 - 00:00:39:06
Very impressive guy.
00:00:39:06 - 00:00:40:23
As I mentioned, you know,
00:00:40:23 - 00:00:43:09
fundraised billions of dollars with a B
00:00:44:21 - 00:00:47:20
or transactions with a B.
00:00:47:20 - 00:00:49:09
So a very experienced gentlemen
00:00:49:09 - 00:00:52:09
in medtech specifically.
00:00:53:01 - 00:00:56:14
One main takeaway I took from that interview was,
00:00:56:14 - 00:01:00:18
he has his "why" pretty well defined.
00:01:00:18 - 00:01:01:17
Med device development
00:01:01:17 - 00:01:04:14
in my mind's a pretty noble cause to begin with.
00:01:04:14 - 00:01:07:00
But the fact that when I asked him,
00:01:08:00 - 00:01:09:09
why is he in this business?
00:01:09:09 - 00:01:11:03
he had a pretty open-shut,
00:01:11:03 - 00:01:13:06
explanation with his colleague
00:01:14:05 - 00:01:16:18
who died of, I believe it was cancer
00:01:16:18 - 00:01:20:22
and just got the wrong therapy, caught it too late
00:01:20:22 - 00:01:22:23
and could have saved his life
00:01:22:23 - 00:01:25:10
if he had gotten caught it earlier obviously
00:01:25:10 - 00:01:27:02
and used the right therapy.
00:01:27:02 - 00:01:29:22
So if that's not a why, I don't know what is.
00:01:29:22 - 00:01:32:12
And then after that with his folks.
00:01:33:19 - 00:01:35:04
Always good to interview people
00:01:35:04 - 00:01:38:12
and understand why they're in this business.
00:01:38:12 - 00:01:40:13
And also great to see and hear
00:01:40:13 - 00:01:42:18
that the CoolStat platform
00:01:42:18 - 00:01:44:22
and the other one, the MiHelper,
00:01:44:22 - 00:01:48:07
they're on their way to market, so pretty exciting.
00:01:48:07 - 00:01:51:01
- Yeah, and obviously Steve has a ton of experience
00:01:51:01 - 00:01:53:23
and I think we learned a few things from him
00:01:53:23 - 00:01:54:19
that's for sure.
00:01:54:19 - 00:01:55:22
So it was great to have him on.
00:01:55:22 - 00:01:57:17
- Yeah, and I think,
00:01:57:17 - 00:02:00:02
we'll talk a lot about this here in this episode.
00:02:00:02 - 00:02:05:02
I think the main lesson I learned,
00:02:05:07 - 00:02:07:19
certainly we're doing this in our line of work,
00:02:07:19 - 00:02:09:12
but just hearing it from the
00:02:09:12 - 00:02:11:06
leader of this startup company
00:02:11:06 - 00:02:14:01
that's raised millions of dollars is,
00:02:14:01 - 00:02:17:06
you have to design these platforms, these products,
00:02:17:06 - 00:02:19:08
that are going into trials ultimately right
00:02:19:08 - 00:02:20:23
before going to market.
00:02:20:23 - 00:02:23:10
You really have to design them with flexibility
00:02:23:10 - 00:02:26:17
and thinking about it as a platform,
00:02:26:17 - 00:02:28:04
because inevitably,
00:02:28:04 - 00:02:30:20
you're going to work with an outside partner
00:02:30:20 - 00:02:31:23
or build a product yourself,
00:02:31:23 - 00:02:33:21
design and build your product yourself
00:02:33:21 - 00:02:36:18
and put it out in the real world.
00:02:36:18 - 00:02:40:04
And not surprisingly, there's going to be challenges.
00:02:40:04 - 00:02:42:01
There's gonna be changes that need to be made.
00:02:42:01 - 00:02:47:16
And he was talking about their nasal cannula,
00:02:47:16 - 00:02:49:21
the disposable, having to change the way
00:02:49:21 - 00:02:51:03
in which they apply
00:02:51:03 - 00:02:54:08
the high flow low humidity air therapy
00:02:54:08 - 00:02:57:04
from, in both nostrils and out through the mouth,
00:02:57:04 - 00:03:00:06
to, in one nostril and out the other.
00:03:00:06 - 00:03:01:08
Now, that's a pretty significant
00:03:01:08 - 00:03:03:09
change in your therapy.
00:03:04:05 - 00:03:08:21
And he made that pivot successfully.
00:03:08:21 - 00:03:10:14
They are making that pivot successfully.
00:03:10:14 - 00:03:14:06
But definitely in talking with you after that interview,
00:03:14:06 - 00:03:16:15
brought up a couple different areas
00:03:16:15 - 00:03:18:10
or I guess ways in which
00:03:18:10 - 00:03:23:05
you can design these products as a platform.
00:03:23:05 - 00:03:25:14
So, I mean, you really, at the end of the day,
00:03:25:14 - 00:03:28:10
you wanna design these products where
00:03:28:10 - 00:03:30:07
if you need to make a change,
00:03:30:07 - 00:03:32:02
you don't have to go back to the drawing board.
00:03:32:02 - 00:03:34:10
It's sort of like Planning 101
00:03:34:10 - 00:03:36:04
and Implementation 101.
00:03:36:04 - 00:03:37:16
But I guess, Jake,
00:03:37:16 - 00:03:38:21
since you've been at Key Tech
00:03:38:21 - 00:03:40:18
for longer than and I have,
00:03:40:18 - 00:03:43:22
let's see, I'm approaching 16 years.
00:03:43:22 - 00:03:45:14
You're a year or two ahead of me,
00:03:45:14 - 00:03:46:21
I've lost count at this point.
00:03:46:21 - 00:03:48:18
But in your experience,
00:03:48:18 - 00:03:51:10
how have you seen products
00:03:51:10 - 00:03:55:19
be developed with flexibility in mind,
00:03:55:19 - 00:03:57:03
broadly and more specifically,
00:03:57:03 - 00:03:58:13
maybe just as a platform
00:03:58:13 - 00:04:01:01
so that you can accommodate these change
00:04:01:01 - 00:04:02:09
as they're required?
00:04:02:09 - 00:04:05:17
- First of all, you're doing a clinical trial for a reason
00:04:05:17 - 00:04:07:01
and that's because you might not know
00:04:07:01 - 00:04:08:23
exactly what you're gonna get for data.
00:04:08:23 - 00:04:12:03
So you've gotta be able to change things on the fly
00:04:13:05 - 00:04:17:00
and you can't spend years or even months doing that.
00:04:17:00 - 00:04:18:21
When you think about configurability,
00:04:19:20 - 00:04:22:03
the obvious place to start is software.
00:04:22:03 - 00:04:24:05
When you design a platform, you gotta make sure
00:04:24:05 - 00:04:25:08
that you have a process
00:04:25:08 - 00:04:27:13
for making software upgrades.
00:04:27:13 - 00:04:29:12
Can you do it on site?
00:04:29:12 - 00:04:30:11
Can you do it remotely?
00:04:30:11 - 00:04:31:16
What sort of testing is needed?
00:04:31:16 - 00:04:33:18
How long is it going to take?
00:04:33:18 - 00:04:35:13
You need to keep all that in mind.
00:04:35:13 - 00:04:37:07
The other thing with software is,
00:04:37:07 - 00:04:39:07
you need to plan ahead for what might change.
00:04:39:07 - 00:04:43:07
So either you use configuration files or variables.
00:04:43:07 - 00:04:44:07
Make sure that it allows you
00:04:44:07 - 00:04:46:15
to adjust those as needed.
00:04:46:15 - 00:04:48:05
And when you make those adjustments,
00:04:48:05 - 00:04:50:15
you don't wanna be introducing new issues.
00:04:50:15 - 00:04:51:20
For example, let's say that
00:04:51:20 - 00:04:52:19
you increase the flow rate
00:04:52:19 - 00:04:55:02
of a device like CoolStat.
00:04:55:02 - 00:04:56:10
You don't wanna be all of a sudden
00:04:56:10 - 00:04:58:01
getting a lot of pressure sensor alarms
00:04:58:01 - 00:04:58:22
because of that change.
00:04:58:22 - 00:05:00:15
So you need to think ahead,
00:05:00:15 - 00:05:02:06
when you change something, what's the impact?
00:05:02:06 - 00:05:03:08
Have you addressed it?
00:05:03:08 - 00:05:04:22
Is it going to be okay?
00:05:04:22 - 00:05:08:12
And then finally with software is data logging.
00:05:08:12 - 00:05:11:00
You wanna get as much data as you can.
00:05:11:00 - 00:05:13:05
Even things that you think you might not need,
00:05:13:05 - 00:05:15:19
you may decide later it's useful data.
00:05:15:19 - 00:05:16:22
So always err on the side
00:05:16:22 - 00:05:18:11
of logging more data
00:05:18:11 - 00:05:19:12
rather than less.
00:05:20:09 - 00:05:22:22
Second area would be electrical.
00:05:22:22 - 00:05:24:09
So with electrical design,
00:05:25:08 - 00:05:28:13
you wanna leave room for changes and upgrades.
00:05:28:13 - 00:05:30:08
Things like the power supply.
00:05:30:08 - 00:05:31:19
Make sure that you can not only provide
00:05:31:19 - 00:05:34:07
enough power for the system at the time,
00:05:34:07 - 00:05:36:22
but if you add features to that platform later,
00:05:36:22 - 00:05:39:07
will you be able to power those.
00:05:39:07 - 00:05:41:01
Things like maybe thermal control,
00:05:41:01 - 00:05:42:23
is your device running hot already
00:05:42:23 - 00:05:45:05
and will it run hotter if you add something new?
00:05:45:05 - 00:05:45:23
You wanna make sure that you
00:05:45:23 - 00:05:47:22
have some overhead there.
00:05:47:22 - 00:05:49:16
Another thing you can do is include hooks
00:05:49:16 - 00:05:52:17
on your circuit boards for sensors and actuators.
00:05:52:17 - 00:05:54:18
If you decide you need to add more sensors,
00:05:54:18 - 00:05:56:17
you've already got it built into your circuit board,
00:05:56:17 - 00:05:58:22
and then it's really just a matter of plugging them in.
00:05:58:22 - 00:06:00:17
The last area would be mechanical.
00:06:00:17 - 00:06:02:11
I think to build a platform,
00:06:02:11 - 00:06:04:06
you wanna think about modularity.
00:06:04:06 - 00:06:08:11
So can you change parts of the hardware
00:06:08:11 - 00:06:10:11
without impacting the entire device?
00:06:10:11 - 00:06:13:07
Can you take a module swap in a new one
00:06:13:07 - 00:06:15:09
and get the functionality that you need
00:06:15:09 - 00:06:16:19
from that new module?
00:06:16:19 - 00:06:19:02
And I think Steve had a good example of that
00:06:19:02 - 00:06:22:19
where he was able to change only the tube set
00:06:22:19 - 00:06:25:11
to adapt the way that the air was reaching
00:06:25:11 - 00:06:27:04
the users nostrils as opposed
00:06:27:04 - 00:06:28:21
to changing the entire system
00:06:28:21 - 00:06:30:11
including the hardware.
00:06:30:11 - 00:06:32:08
And then lastly, you wanna be
00:06:32:08 - 00:06:34:03
thinking about serviceability.
00:06:34:03 - 00:06:37:18
So with a platform, if you're gonna need to change it,
00:06:37:18 - 00:06:38:23
you wanna be able to open it up,
00:06:38:23 - 00:06:41:01
get in there, change components,
00:06:41:01 - 00:06:44:02
swap things out without damaging anything,
00:06:44:02 - 00:06:46:02
or without having a hard time putting it back together.
00:06:46:02 - 00:06:49:09
So you wanna think about avoiding things like glues
00:06:49:09 - 00:06:51:23
and other mechanisms that would prevent
00:06:51:23 - 00:06:53:13
serviceability and access.
00:06:53:13 - 00:06:54:23
- Yeah, those are great points, Jake.
00:06:54:23 - 00:06:57:07
I would add, it's a little bit harder
00:06:57:07 - 00:06:59:01
to sort of give concrete examples,
00:06:59:01 - 00:07:00:18
but from a user perspective,
00:07:01:14 - 00:07:02:23
before you get into a trial,
00:07:02:23 - 00:07:05:14
you're going to do pre-formative evaluations
00:07:05:14 - 00:07:08:15
and you're gonna put different product anatomies
00:07:08:15 - 00:07:13:09
in front of potential end users and let them tell you
00:07:13:09 - 00:07:15:08
what the ideal product architecture
00:07:15:08 - 00:07:16:11
and use case should be.
00:07:16:11 - 00:07:18:14
So that when you get in the trial,
00:07:18:14 - 00:07:21:16
your changes are very technical
00:07:21:16 - 00:07:24:09
or related to the actual therapy you're applying
00:07:24:09 - 00:07:26:22
and not how a user puts on a mask
00:07:26:22 - 00:07:31:05
or loads up a GUI or engages with a GUI.
00:07:31:05 - 00:07:34:10
You wanna lock that in early while the costs are low
00:07:34:10 - 00:07:36:16
by doing your pre-performative evaluations,
00:07:36:16 - 00:07:38:10
doing them thoroughly.
00:07:39:19 - 00:07:41:10
Any other areas, Jake?
00:07:41:10 - 00:07:44:06
- I mean, I guess another thing to think about is,
00:07:44:06 - 00:07:46:10
whenever you're designing a platform
00:07:46:10 - 00:07:48:20
and you're trying to include flexibility,
00:07:48:20 - 00:07:51:17
you've gotta think, what's the extra effort?
00:07:52:15 - 00:07:53:23
There's pros and cons of everything.
00:07:53:23 - 00:07:55:04
You can't have it all.
00:07:55:04 - 00:07:57:08
So if you're designing in these extra hooks
00:07:58:12 - 00:08:00:16
on your circuit boards, if you're making sure
00:08:00:16 - 00:08:04:00
that your software is easily reconfigured,
00:08:04:00 - 00:08:05:23
it's gonna take some extra effort.
00:08:05:23 - 00:08:07:23
So you're gonna spend that time upfront
00:08:07:23 - 00:08:09:10
to get that stuff right,
00:08:09:10 - 00:08:11:03
but it's gonna pay dividends down the road.
00:08:11:03 - 00:08:12:17
It's way more difficult
00:08:12:17 - 00:08:14:16
to make those changes later on
00:08:14:16 - 00:08:16:10
if you haven't built those hooks in originally.
00:08:16:10 - 00:08:19:11
So I think at Key Tech we would always recommend
00:08:19:11 - 00:08:21:14
making sure that you've planned ahead
00:08:21:14 - 00:08:23:11
when you're doing the design early.
00:08:23:11 - 00:08:25:00
- The CoolStat is a great example
00:08:25:00 - 00:08:27:12
of developing a product as a platform
00:08:27:12 - 00:08:30:10
so that you can make pivots along the way.
00:08:30:10 - 00:08:32:07
The therapy was changed,
00:08:32:07 - 00:08:34:16
the consumable changed slightly,
00:08:34:16 - 00:08:36:06
and they were able to make those changes
00:08:36:06 - 00:08:38:20
and kind of keep going with the trial.
00:08:38:20 - 00:08:40:22
Another kind of more extreme example
00:08:40:22 - 00:08:42:20
we'll talk about in the next episode
00:08:43:20 - 00:08:47:02
is how the CoolStat platform was used
00:08:47:02 - 00:08:49:00
as a means to conduct a trial
00:08:49:00 - 00:08:50:21
for completely different therapy.
00:08:50:21 - 00:08:53:21
This is for migraines for their MiHelper platform.
00:08:53:21 - 00:08:56:06
So we'll get into that in another episode.
00:08:56:06 - 00:09:00:11
So for now, Jake, thanks for joining the interview
00:09:00:11 - 00:09:03:01
and this post-show chat.
00:09:03:01 - 00:09:05:02
And we'll talk to everybody soon.
00:09:05:02 - 00:09:05:22
- Thanks a lot, Andy.
00:09:05:22 - 00:09:06:22
We'll talk again in a bit.