144 – Big Data & Real-World Evidence (S10E9)

Welcome you. You've given us a really fascinating
set of materials this time, really digging into
how big data, massive amounts of data, and real
-world evidence are shaking things up in healthcare
decision -making, especially around new drugs.
Absolutely. Yeah, our mission for this deep dive
is, well, to pull out the key insights from all
that material you sent over. We're seeing this
fundamental shift, really, in understanding drug
effectiveness and safety moving beyond those
very controlled clinical trials. Right. Looking
at what happens sort of on the ground. in everyday
practice. Exactly. So we'll look at where this
real -world data comes from, the powerful analytic
tools needed, and yeah, some concrete examples
of the impact. What struck me looking through
this was just the sheer scale of it all. The
volume and the variety of data we have now. It
paints a much richer picture, doesn't it? It
really does. It's a whole new level of detail
compared to what we had before. Okay, so let's
kick things off with the basics then. When we
say real -world evidence, RWE, what are we actually
talking about? OK, so simply put, RWE is the
clinical evidence about the usage and potential
benefits or risks of a medical product derived
from analysis of real -world data, RWD. And that
RWD is collected from actual patient experiences
out in the wild, so to speak. Precisely, not
in the artificial setting of a randomized controlled
trial or RCT. That's the key difference. Right,
because RCTs, while they're the gold standard
for proving efficacy initially, They have really
strict rules, don't they? Who can be in the trial?
Who can't? Exactly. Strict inclusion and exclusion
criteria, which means the people in the trial
might not perfectly represent the broader patient
population you see in clinics every day. Those
patients often have multiple conditions or are
taking other meds. Yeah, real life is messy.
Right. Not like a neat study protocol. Uh -huh.
And that's where big data analytics really becomes
crucial. We finally have the computational power,
the tools to actually process and make sense
of these huge complex data sets from the real
world. So it's not just having the data mountain,
it's having the tools to mine it for insights.
That's it, exactly. Imagine like a million -piece
jigsaw puzzle. Before, we could maybe only look
at small sections. Now, big data analytics, often
using things like AI, let's just see the bigger
picture, find subtle patterns, connections we'd
completely miss otherwise. OK, that makes sense.
Finding the signal and the noise, basically.
Yeah. it gives us a much, much deeper understanding
of drug performance, safety profiles, across
really diverse groups of people. All right, so
where does all this real world data, the RWD,
actually... Come from? Sounds like it's pulled
from many places. It is, yeah. Think about all
the digital breadcrumbs from healthcare interactions,
electronic health records, EHRs are a massive
source. You know, your doctor's digital notes,
diagnoses, lab results, prescriptions. Okay,
the records kept by hospitals and clinics. Got
it. Then there's claims and billing data. Right.
From insurance companies, payers. That tells
you a lot about which treatments people are actually
getting, how often the costs involved. It paints
a picture of healthcare utilization. Right, patterns
of care. What else? Increasingly, patient -generated
health data. This is huge. Stuff from wearables,
fitness trackers, sleep monitors. Also, data
patients enter themselves into health apps or
patient surveys. Huh. So individuals are directly
contributing their own data points? That adds
another layer. A really important one. And we
also have disease registries. These are focused
databases collecting specific information on
patients with a particular condition, maybe cystic
fibrosis or a type of cancer. Great for long
-term tracking. So really targeted datasets for
specific diseases. Yeah. And finally, don't forget
pharmacovigilance data. These are the adverse
event report side effects that doctors, patients,
or companies report after a drug is on the market.
Wow. Okay, so from your smartwatch to insurance
claims to a side effect report, it's quite a
mix. And all this feeds into generating that
real world evidence. Exactly. It all comes together,
gets analyzed, and hopefully gives us that more
complete picture. And this RWE is now being actively
used to inform some really critical decisions.
OK, let's get into that. How is it being applied?
Where are we seeing the impact? Well, one major
area is regulation. You know, bodies like the
FDA, they're increasingly looking at and using
RWE. especially for post -market surveillance.
Monitoring drugs after they're approved. Precisely.
Understanding long -term safety, seeing how effective
a drug really is once millions are using it,
not just the few thousand in a trial. It can
help spot rare side effects that didn't show
up earlier. That seems incredibly important for
public health. It is, and RWE can also help understand
how well a drug works in specific subgroups,
maybe older patients or those with kidney problems
who might have been excluded from trials. Sometimes
it can even support expanding a drug's approved
use, its label, based on compelling real -world
data. So potentially speeding up access to treatments
for different patient groups. What about actual
health care delivery? How doctors treat patients
day to day? Yeah, RWE is starting to inform clinical
practice guidelines, helping guide doctors on
the best treatment pathways based on broader
evidence. And it's a big driver for personalized
medicine. How so? Well, imagine using insights
from RWE. to tailor a treatment choice for your
specific patient based on observing outcomes
in thousands of similar real -world patients,
similar age, similar other conditions, you know?
Right. Moving beyond just the average trial result
to what works for this type of person. Exactly.
The ultimate aim is always improving patient
outcomes, making care more effective and potentially
more efficient too. Okay. And you mentioned drug
development itself is a key area being changed
by this. How is RWE fitting into that pipeline?
It's touching multiple stages. Right at the beginning,
RWE can help with identifying and validating
new drug targets. How does that work? By analyzing
real -world data on disease progression, unmet
needs, how existing treatments are performing
or failing. Researchers can get clues about which
biological pathways or molecules are truly driving
the disease in actual patients, making them potentially
better targets for new therapies. So it grounds
the lab work and the reality of disease in populations.
Yeah, kind of. Think about cancer drug discovery
shifting towards specific molecular targets.
RWE helps confirm which targets are most relevant
across diverse patient groups. Interesting. What
about clinical trials themselves? Can RWE make
those better? Definitely. It can help optimize
trial design. For instance, using RWD to better
identify the right patients to enroll those most
likely to respond or those with the highest unmet
need. This can make trials more efficient, maybe
smaller or faster. Refining the recruitment.
And there's a lot of buzz around using RWE to
create external control arms. External controls,
what are those? So instead of having a placebo
group within your current trial, especially maybe
for rare diseases or certain cancers where placebos
are tough ethically, you might use historical
RWD from similar patients treated previously
with the standard of care as your comparison
group. Ah, using past data as the benchmark instead
of a concurrent placebo group. Exactly. It's
complex, needs careful statistical handling,
but it holds promise for accelerating development
in some areas. I can see how that would be useful.
And then, of course, there's the massive role
RWE plays in post -market surveillance, which
we touched on. Monitoring safety and effectiveness
continuously once a drug is launched, detecting
rare events, understanding real -world usage
patterns. It's vital. It really creates this
continuous learning loop, doesn't it? feeding
insights back into practice and future development.
Absolutely. Now obviously handling all this data
It's not trivial. Yeah, I was gonna ask. Analyzing
these massive messy data sets must need some
serious tech. It definitely does. You need sophisticated
analytic platforms. These systems have to integrate
data from all those different sources we mentioned,
EHRs, claims, wearables, clean it, standardize
it, which is a huge task in itself, and then
apply advanced analytical methods. Not something
you can do in Excel. Not even close. We're talking
techniques like machine learning, AI. They're
becoming essential tools to sift through the
complexity, identify those subtle patterns, predict
outcomes, things humans just couldn't do at this
scale. So AI is becoming the engine to unlock
the value in the RWD. In many ways, yes, but
it's also crucial to be aware of potential biases
in RWD. It's not perfectly curated like trial
data. So the analytical methods have to be incredibly
rigorous to ensure the evidence the RWE we generate
is actually valid and reliable. Garbage in, garbage
out still applies. Right, the quality of the
analysis is paramount. Now, our outline mentioned
connecting this to some concepts like process
analytical technology, PA, for manufacturing.
How do those worlds link up? That's a really
interesting parallel, actually. While the source
materials didn't give us direct RWE clinical
outcomes linked to PAIT, the underlying philosophy
is very similar. It's all about using data for
better control and understanding. OK, explain
that link. Well, think about PAIT in drug manufacturing.
We talked about using real -time sensors and
data analysis during production to monitor critical
quality attributes, ensuring the process stays
within spec, making adjustments on the fly. It's
continuous monitoring for quality control. right,
data -driven manufacturing. Now think about RWE.
It's essentially continuous monitoring for drug
performance after manufacturing out in the real
world. We're using real -world data streams to
track safety and effectiveness on an ongoing
basis. So just like Peatney monitors the process,
RWE monitors the product's impact in practice.
Ah, I see. So it's analogous continuous data
streams for continuous understanding just at
different stages of the drug life cycle. Exactly.
One focuses on making the drug consistently well,
the other on understanding how the well -made
drug performs in diverse people over time. Both
rely heavily on collecting and intelligently
analyzing data streams. That's a neat connection.
What about the idea of scaling up? Yeah, remember
discussing things like using lab tools to predict
how drug particles might behave when you scale
up production from a small batch to a huge one.
Predicting performance at scale. RWE is kind
of doing the same thing, but for patients. Clinical
trials are like the small lab batch, a limited
controlled population. RWE tries to understand
and predict how the drug will perform when scaled
up to the entire diverse real -world patient
population. It's about understanding performance
beyond the initial controlled setting. Okay,
predicting real -world performance based on initial
signals, analogous to predicting manufacturing
performance. That makes a lot of sense. It does.
The core idea is using data proactively, either
to ensure manufacturing quality or to understand
and optimize patient outcomes. OK, so wrapping
things up. It feels like the key takeaway is
that big data analytics is genuinely transforming
our ability to generate and use real world evidence.
Without a doubt, it's providing a much more comprehensive,
nuanced understanding of how medicines actually
work and don't work outside the confines of traditional
trials. And this, RWE, isn't just academic. It's
directly influencing major decisions, regulatory
pathways, clinical guidelines, how we even design
the next generation of drugs. Correct. We're
definitely uncovering new aha moments as we get
better at tapping into this wealth of real -world
experience data. It's pretty powerful stuff.
So maybe a final thought for you the listener.
Just consider how all the health data points
generated about you, maybe from doctor visits,
prescriptions, perhaps even your fitness app,
might be contributing anonymously, of course,
to this bigger picture. How this collective data
could shape more personalized medicine in the
future, and how new treatments get developed
and watched over time. Yeah, and you know, what
are the ethical considerations there? As more
and more personal health information gets used
for RWE, it's definitely something to think about.
Definitely food for thought.