TRANSCRIPT 🔗
1
00:00:07,680 --> 00:00:10,590
- And I'd like to begin
our conversation today
2
00:00:10,590 --> 00:00:13,380
with discussion of AI.
3
00:00:13,380 --> 00:00:15,750
So it's been a veritable
buzzword this year,
4
00:00:15,750 --> 00:00:19,050
but rather than addressing it too broadly,
5
00:00:19,050 --> 00:00:20,880
I'd like to focus in the first instance
6
00:00:20,880 --> 00:00:24,120
on the importance of its accurate usage.
7
00:00:24,120 --> 00:00:27,000
- Yeah, it's almost impossible
8
00:00:27,000 --> 00:00:30,453
not to talk about AI, right,
now in our industry, isn't it?
9
00:00:31,800 --> 00:00:34,920
And, actually, it's a great question
10
00:00:34,920 --> 00:00:38,610
because I think the use
of AI in our industry,
11
00:00:38,610 --> 00:00:42,120
there's a lot of hope, expectation,
12
00:00:42,120 --> 00:00:44,700
but also a tinge of hype as well.
13
00:00:44,700 --> 00:00:47,640
And so I think from our perspective,
14
00:00:47,640 --> 00:00:50,097
the way that we are thinking about this
15
00:00:50,097 --> 00:00:55,097
is there are elements of
where AI is useful already
16
00:00:56,400 --> 00:01:00,149
and there's areas where
AI is not yet ready
17
00:01:00,149 --> 00:01:03,150
for the mainstream.
18
00:01:03,150 --> 00:01:07,710
And you quite rightfully sort
of called out generative AI
19
00:01:07,710 --> 00:01:10,906
as being the exciting area right now.
20
00:01:10,906 --> 00:01:15,906
AI is being used to capture a broad array
21
00:01:15,960 --> 00:01:20,960
of robotics, analytics, all
the way up to generative AI,
22
00:01:21,180 --> 00:01:23,340
and it's worth just
picking that apart slightly
23
00:01:23,340 --> 00:01:26,310
'cause I think there's
value in some things
24
00:01:26,310 --> 00:01:28,860
that fall under this big catch-all.
25
00:01:28,860 --> 00:01:30,600
So, starting with the areas
26
00:01:30,600 --> 00:01:35,043
where we believe there
is great opportunity,
27
00:01:36,000 --> 00:01:38,793
we certainly see around base automation.
28
00:01:39,630 --> 00:01:44,630
So, taking a repetitive
task that humans do today,
29
00:01:45,450 --> 00:01:50,450
but not necessarily their
favorite activity in any day,
30
00:01:50,550 --> 00:01:53,223
but are still costly and time consuming,
31
00:01:54,180 --> 00:01:55,650
and those are some of the areas
32
00:01:55,650 --> 00:01:57,990
which really, really suit automation.
33
00:01:57,990 --> 00:02:01,200
So, in our world, in clinical,
34
00:02:01,200 --> 00:02:05,717
back to our previous
conversation, we have a TMF Bot
35
00:02:07,140 --> 00:02:11,610
which captures the documentation
around clinical studies,
36
00:02:11,610 --> 00:02:16,290
automatically categorizes it,
and puts it into the system.
37
00:02:16,290 --> 00:02:18,480
It's an important compliant job,
38
00:02:18,480 --> 00:02:22,822
but actually a human
neither enjoys doing it,
39
00:02:22,822 --> 00:02:24,360
but can't do it as well as a machine now.
40
00:02:24,360 --> 00:02:27,123
So, actually, it's perfect for automation.
41
00:02:29,010 --> 00:02:31,410
Another area where true generative AI
42
00:02:31,410 --> 00:02:35,460
is becoming particularly
useful is as an assistant.
43
00:02:35,460 --> 00:02:39,330
And so there are technologies
out there right now.
44
00:02:39,330 --> 00:02:41,400
I mean, just calling out ChatGPT
45
00:02:41,400 --> 00:02:46,344
has been one example of
that from OpenAI Microsoft
46
00:02:46,344 --> 00:02:50,910
and the way that can be now
integrated in with Teams
47
00:02:50,910 --> 00:02:53,490
to record the notes from a meeting,
48
00:02:53,490 --> 00:02:55,860
such as our conversation right now
49
00:02:55,860 --> 00:02:58,533
where it would capture
what you said, what I said,
50
00:02:59,370 --> 00:03:02,310
and could assign actions to each of us
51
00:03:02,310 --> 00:03:04,950
based on our conversation.
52
00:03:04,950 --> 00:03:06,120
That is a great thing.
53
00:03:06,120 --> 00:03:08,670
No one likes being the
note taker in the meeting,
54
00:03:08,670 --> 00:03:12,090
and it's something that enables people
55
00:03:12,090 --> 00:03:14,250
to focus on what they're saying.
56
00:03:14,250 --> 00:03:15,750
However, even with that,
57
00:03:15,750 --> 00:03:18,813
there's a little risk in a
compliant industry such as ours.
58
00:03:20,413 --> 00:03:25,320
And that's the catch with generative AI.
59
00:03:25,320 --> 00:03:29,913
Unlike machine learning
or traditional analytics,
60
00:03:30,750 --> 00:03:33,120
you quite often can get into a situation
61
00:03:33,120 --> 00:03:35,700
where you don't know the
genesis of the decision
62
00:03:35,700 --> 00:03:40,700
and you can get a different
answer from the same algorithm
63
00:03:41,623 --> 00:03:44,490
if you ask the same question twice.
64
00:03:44,490 --> 00:03:46,950
And that of course doesn't work
with the compliant industry
65
00:03:46,950 --> 00:03:49,530
where you need to show
consistency in your decisions
66
00:03:49,530 --> 00:03:50,760
to be able to make sure
67
00:03:50,760 --> 00:03:52,830
that you're making the right decisions.
68
00:03:52,830 --> 00:03:55,950
So, let me give you an example of that.
69
00:03:55,950 --> 00:03:58,860
If a rep is meeting with a doctor,
70
00:03:58,860 --> 00:04:03,423
and I'm using generative AI to capture it,
71
00:04:04,500 --> 00:04:06,840
and I don't check my notes,
72
00:04:06,840 --> 00:04:10,923
it could have caught the
conversation incorrectly.
73
00:04:11,760 --> 00:04:15,356
And in doing so, it could have,
74
00:04:15,356 --> 00:04:19,470
instead of just talking about the drug
75
00:04:19,470 --> 00:04:21,960
within the bounds of what
you're allowed to talk about it,
76
00:04:21,960 --> 00:04:25,500
it could capture a comment
which could be taken incorrectly
77
00:04:25,500 --> 00:04:30,500
and be used as an example of
bad behavior in an interaction.
78
00:04:30,570 --> 00:04:33,150
But once it's recorded,
if it's not checked,
79
00:04:33,150 --> 00:04:35,190
it's then a discoverable item.
80
00:04:35,190 -->
Load full transcript
00:04:37,110
So again, just being very careful,
81
00:04:37,110 --> 00:04:39,750
even on an innocuous
thing like taking notes,
82
00:04:39,750 --> 00:04:41,670
there's a risk there.
83
00:04:41,670 --> 00:04:44,790
Now, if you take that forward
into some of the other areas,
84
00:04:44,790 --> 00:04:49,790
if it's into true generative
AI usage, there's opportunities
85
00:04:50,400 --> 00:04:52,440
potentially looking for drug candidates
86
00:04:52,440 --> 00:04:56,170
where, you know, you
could potentially identify
87
00:04:57,660 --> 00:05:00,772
a particular pathway and a target protein,
88
00:05:00,772 --> 00:05:03,210
and that's one potential use.
89
00:05:03,210 --> 00:05:05,640
The advantage there is that
can be checked by humans,
90
00:05:05,640 --> 00:05:08,090
see that you haven't gone
catastrophically wrong.
91
00:05:09,180 --> 00:05:11,820
Another area where generative AI
92
00:05:11,820 --> 00:05:14,550
is not yet at the same
level as machine learning
93
00:05:14,550 --> 00:05:17,713
is that we've used machine learning
94
00:05:17,713 --> 00:05:21,919
to be able to diagnose
diseases effectively.
95
00:05:21,919 --> 00:05:26,610
And there are examples in
place today such as HIV.
96
00:05:26,610 --> 00:05:29,400
There was some interesting work done
97
00:05:29,400 --> 00:05:31,830
all around epilepsy as well,
98
00:05:31,830 --> 00:05:36,274
where you could codify
what made a good diagnosis
99
00:05:36,274 --> 00:05:39,030
and you could use it,
and that was something
100
00:05:39,030 --> 00:05:41,550
that you can actually
categorize as a medical device
101
00:05:41,550 --> 00:05:42,700
'cause it's repeatable.
102
00:05:43,710 --> 00:05:45,863
If you ask it the question and
then I ask it the question,
103
00:05:47,430 --> 00:05:50,250
it'll give you the same
answer so you can check it.
104
00:05:50,250 --> 00:05:51,300
So that works really well.
105
00:05:51,300 --> 00:05:54,210
But generative AI, if you're
getting a different answer,
106
00:05:54,210 --> 00:05:58,170
that's risky, because when
generative AI gets it wrong,
107
00:05:58,170 --> 00:06:00,240
it can be catastrophically wrong.
108
00:06:00,240 --> 00:06:02,613
And unlike human beings,
109
00:06:03,510 --> 00:06:06,360
it doesn't have the
notion of common sense.
110
00:06:06,360 --> 00:06:09,840
So from the spectrum of AI
111
00:06:09,840 --> 00:06:13,560
encapsulating automation
through to true generative,
112
00:06:13,560 --> 00:06:16,050
I think there are examples
of where it can be used,
113
00:06:16,050 --> 00:06:18,303
but there's cautions associated with it.
114
00:06:19,189 --> 00:06:21,630
And that then brings you on neatly
115
00:06:21,630 --> 00:06:25,122
into the next area of how
can you make it better.
116
00:06:25,122 --> 00:06:29,850
The analytics, whether
it's machine learning
117
00:06:29,850 --> 00:06:32,910
or generative AI, are totally reliant
118
00:06:32,910 --> 00:06:35,490
on the quality of the
data that you give it.
119
00:06:35,490 --> 00:06:39,720
So if you look at open domain sources,
120
00:06:39,720 --> 00:06:41,070
you've got no control over that.
121
00:06:41,070 --> 00:06:44,670
And there were some
great examples early on
122
00:06:44,670 --> 00:06:49,670
which the training data
taught the AI tools
123
00:06:49,920 --> 00:06:54,330
very bad behaviors, be
those misogynist or racist,
124
00:06:54,330 --> 00:06:57,570
because that was the
training data that it had.
125
00:06:57,570 --> 00:07:00,000
Now, what we see in our industry
126
00:07:00,000 --> 00:07:01,470
is you don't have that luxury.
127
00:07:01,470 --> 00:07:04,500
It's funny when it's
in a newspaper article.
128
00:07:04,500 --> 00:07:06,960
It's really bad when it's in medicine.
129
00:07:06,960 --> 00:07:11,960
So in our world actually
high-quality, harmonized data
130
00:07:12,840 --> 00:07:14,160
is really valuable.
131
00:07:14,160 --> 00:07:18,474
And so on the one hand, on the
absolute extreme treatments
132
00:07:18,474 --> 00:07:20,880
and suggestions for treatments,
133
00:07:20,880 --> 00:07:24,120
but even down to the notion of coming up
134
00:07:24,120 --> 00:07:26,340
with the next best action for a rep,
135
00:07:26,340 --> 00:07:28,410
what should they do next?
136
00:07:28,410 --> 00:07:31,436
And certainly where we
see our role at Veeva
137
00:07:31,436 --> 00:07:34,620
is doing the 'grindy' work,
138
00:07:34,620 --> 00:07:38,100
the hard bit, of giving harmonized data,
139
00:07:38,100 --> 00:07:40,530
and then working with our partners
140
00:07:40,530 --> 00:07:44,366
for them to use their
algorithms on top of it
141
00:07:44,366 --> 00:07:47,790
to help pharma companies
where appropriate.
142
00:07:47,790 --> 00:07:52,050
- Okay, so if we talk about
this 'grindy' work and Veeva,
143
00:07:52,050 --> 00:07:55,440
where does this all apply,
and what's the impact of it
144
00:07:55,440 --> 00:07:58,500
within specialized or
personalized medicine?
145
00:07:58,500 --> 00:08:03,022
- If you think about, and we
talked about this in the past,
146
00:08:03,022 --> 00:08:08,022
this sort of notion, and in
fact we heard it here at Summit
147
00:08:08,370 --> 00:08:13,370
with the panel with
physicians, they actually want,
148
00:08:13,950 --> 00:08:18,950
physicians, KOLs want
input from the industry.
149
00:08:19,290 --> 00:08:21,630
They want to understand
150
00:08:21,630 --> 00:08:23,910
what new treatments are, how to use them,
151
00:08:23,910 --> 00:08:28,410
they want medical education,
but it's got to be useful,
152
00:08:28,410 --> 00:08:32,520
and each interaction has to be
useful, and each interaction
153
00:08:32,520 --> 00:08:34,920
should build off the previous interaction,
154
00:08:34,920 --> 00:08:36,990
whether it was you talking to them first
155
00:08:36,990 --> 00:08:39,180
or me talking to them.
156
00:08:39,180 --> 00:08:43,200
Whoever goes next should be a step beyond
157
00:08:43,200 --> 00:08:46,380
where we each covered previously.
158
00:08:46,380 --> 00:08:47,943
And that's quite
difficult.
159
00:08:49,170 --> 00:08:52,860
It's quite difficult unless
you bring all the information
160
00:08:52,860 --> 00:08:54,840
about that interaction together,
161
00:08:54,840 --> 00:08:57,360
unless you understand that individual
162
00:08:57,360 --> 00:09:00,510
and what their journey is,
what their level of interest,
163
00:09:00,510 --> 00:09:03,960
and you understand the information
that you have available.
164
00:09:03,960 --> 00:09:08,790
So, what we do is we make sure
that the interaction data,
165
00:09:08,790 --> 00:09:13,740
whether it's with medical,
clinical, or the field teams,
166
00:09:13,740 --> 00:09:16,263
is stored and managed in a consistent way.
167
00:09:17,670 --> 00:09:22,670
We take all of that body of
information about physicians
168
00:09:23,340 --> 00:09:26,940
and make it available in
one of our tools, Link,
169
00:09:26,940 --> 00:09:30,270
so that you can truly understand
about that individual,
170
00:09:30,270 --> 00:09:31,500
what they're interested in,
171
00:09:31,500 --> 00:09:33,540
the sort of congresses they've been to,
172
00:09:33,540 --> 00:09:35,523
so you can form a picture of them.
173
00:09:36,720 --> 00:09:40,560
And then, ultimately, that
will be a great source
174
00:09:40,560 --> 00:09:44,460
of information to pull together
with when you last saw them,
175
00:09:44,460 --> 00:09:47,850
what meetings they're going to,
and come up with suggestions
176
00:09:47,850 --> 00:09:50,430
as to what the next
interaction might look like
177
00:09:50,430 --> 00:09:52,350
or how you might do it.
178
00:09:52,350 --> 00:09:55,770
But all of that requires
that you've got accurate,
179
00:09:55,770 --> 00:09:58,350
harmonized, and consistent data
180
00:09:58,350 --> 00:10:01,560
so the algorithm can do
its thing on top of it.
181
00:10:01,560 --> 00:10:02,393
- Absolutely.
182
00:10:02,393 --> 00:10:03,960
So we're talking not just clean data,
183
00:10:03,960 --> 00:10:06,180
but high-quality data.
- Yeah.
184
00:10:06,180 --> 00:10:11,130
- And you also went on,
I believe, to this notion
185
00:10:11,130 --> 00:10:12,990
that's coming in at the moment of inbound
186
00:10:12,990 --> 00:10:16,560
as well as outbound when it
comes to engagement with HCPs.
187
00:10:16,560 --> 00:10:19,080
Perhaps you could expand on that slightly.
188
00:10:19,080 --> 00:10:22,263
- Yeah, this is a really interesting area.
189
00:10:23,166 --> 00:10:26,730
If you look at the
generational muscle mass
190
00:10:26,730 --> 00:10:29,103
of the life sciences industry,
191
00:10:30,382 --> 00:10:32,998
it's created an industrial-scale
192
00:10:34,665 --> 00:10:36,298
push mechanism for information.
193
00:10:37,320 --> 00:10:40,500
So, if you are a physician,
I go and see you,
194
00:10:40,500 --> 00:10:43,923
I talk about what I want
to talk about to you.
195
00:10:46,560 --> 00:10:49,200
What we actually hear,
time and time again,
196
00:10:49,200 --> 00:10:52,320
and we heard it on stage here at Summit,
197
00:10:52,320 --> 00:10:57,240
was that physicians want
to speak to those companies
198
00:10:57,240 --> 00:11:01,500
who have information that is
incremental and useful to them.
199
00:11:01,500 --> 00:11:03,420
Now, the catch about the current model
200
00:11:03,420 --> 00:11:08,420
is that assumes that the visiting MSL
201
00:11:08,700 --> 00:11:11,760
or rep turns up at the right time,
202
00:11:11,760 --> 00:11:15,600
but actually for that
information to be useful for me,
203
00:11:15,600 --> 00:11:17,460
it needs to also be timely.
204
00:11:17,460 --> 00:11:19,800
And that may not be
when a visit is planned.
205
00:11:19,800 --> 00:11:22,530
So, for example, if I've
just seen a patient,
206
00:11:22,530 --> 00:11:25,110
I want to learn about
new oncology treatments
207
00:11:25,110 --> 00:11:27,963
in a specific area, specific tumor type,
208
00:11:29,190 --> 00:11:32,241
I need to be able to reach out proactively
209
00:11:32,241 --> 00:11:35,157
and find out if that information is there.
210
00:11:35,157 --> 00:11:39,150
And so this notion of pull
becomes really important.
211
00:11:39,150 --> 00:11:42,390
And in fact, we've
recently conducted a study
212
00:11:42,390 --> 00:11:46,470
where we found that where you
make pull channels available
213
00:11:46,470 --> 00:11:50,370
so the physician can
actually ask for help,
214
00:11:50,370 --> 00:11:53,910
that makes up 30% of all interactions.
215
00:11:53,910 --> 00:11:57,390
And we've also seen that
those interactions on average
216
00:11:57,390 --> 00:11:58,680
answered in about five minutes.
217
00:11:58,680 --> 00:12:01,800
So all of a sudden, you've
shifted the whole dynamic
218
00:12:01,800 --> 00:12:03,330
and you've become useful
219
00:12:03,330 --> 00:12:06,180
because you're actually
addressing a real need.
220
00:12:06,180 --> 00:12:08,310
And then where we see that usage,
221
00:12:08,310 --> 00:12:11,700
we don't just see the pull channels go up,
222
00:12:11,700 --> 00:12:13,800
we also see all other channels go up
223
00:12:13,800 --> 00:12:16,980
because you've built a real
dialogue with the individual.
224
00:12:16,980 --> 00:12:19,140
- So now I want to talk about value,
225
00:12:19,140 --> 00:12:22,020
because at the end of the day,
it all comes down to money
226
00:12:22,020 --> 00:12:23,970
or cost at least.
- Yeah.
227
00:12:23,970 --> 00:12:28,410
- So can you tell us how
medical will prove pivotal
228
00:12:28,410 --> 00:12:29,550
as scrutiny of value
229
00:12:29,550 --> 00:12:32,640
for money intensifies
in Europe, for example?
230
00:12:32,640 --> 00:12:34,110
- Yeah, yeah.
231
00:12:34,110 --> 00:12:36,900
It's interesting because there
was a McKinsey study showing
232
00:12:36,900 --> 00:12:39,390
that medical budgets haven't increased
233
00:12:39,390 --> 00:12:41,013
in the last five years.
234
00:12:42,690 --> 00:12:44,610
Counter to that, though, is the fact
235
00:12:44,610 --> 00:12:46,320
that when you speak to HCPs,
236
00:12:46,320 --> 00:12:51,240
they say they really value
their interactions with medical
237
00:12:51,240 --> 00:12:54,803
'cause it provides educational information
238
00:12:54,803 --> 00:12:58,650
and new data on products that
they may not otherwise have.
239
00:12:58,650 --> 00:13:01,830
So there's a bit of a gap there.
240
00:13:01,830 --> 00:13:04,800
So everyone's kind of known
that medical was really useful
241
00:13:04,800 --> 00:13:07,863
and effective, but no one's
been able to track it.
242
00:13:08,820 --> 00:13:10,830
So we've been using our data,
243
00:13:10,830 --> 00:13:14,247
our interaction data, to
look at what the impact is.
244
00:13:14,247 --> 00:13:15,510
And what we've seen
245
00:13:15,510 --> 00:13:20,160
is that where MSLs are included pre-launch
246
00:13:20,160 --> 00:13:25,160
and used to heavily
educate around a treatment
247
00:13:26,580 --> 00:13:30,390
three months or more in
advance of the launch itself,
248
00:13:30,390 --> 00:13:33,900
we see an uplift of 1.5
times the prescription
249
00:13:33,900 --> 00:13:36,063
of that product, which is just huge.
250
00:13:37,200 --> 00:13:39,210
Now, on top of that, what we actually see
251
00:13:39,210 --> 00:13:40,890
is the stickiness of the product
252
00:13:40,890 --> 00:13:44,640
goes forward into the next
year, year and a half,
253
00:13:44,640 --> 00:13:46,950
because you've actually
set the market up properly,
254
00:13:46,950 --> 00:13:48,330
it's being used in the right way,
255
00:13:48,330 --> 00:13:50,040
so you're getting the right outcome,
256
00:13:50,040 --> 00:13:52,110
so it tends to be stickier.
257
00:13:52,110 --> 00:13:54,210
So it is very much the shift,
258
00:13:54,210 --> 00:13:58,170
especially as we move more
towards 70% of treatments
259
00:13:58,170 --> 00:14:00,993
being in specialty or targeted treatments,
260
00:14:02,040 --> 00:14:03,960
making sure that the physician
261
00:14:03,960 --> 00:14:06,390
can match the right treatments
to the right patients,
262
00:14:06,390 --> 00:14:08,313
and then therefore get the outcomes.
263
00:14:09,300 --> 00:14:13,020
- Okay, so in short, it's
about the sort of pressures
264
00:14:13,020 --> 00:14:15,480
that are abounding when
it comes to biopharma
265
00:14:15,480 --> 00:14:17,370
demonstrating value for money
266
00:14:17,370 --> 00:14:19,380
alongside that clinical efficacy,
267
00:14:19,380 --> 00:14:21,960
and as you said, the adherence factor.
268
00:14:21,960 --> 00:14:25,350
Perhaps we can address
the Veeva Pulse findings
269
00:14:25,350 --> 00:14:29,340
that less than half of
HCPs are now accessible.
270
00:14:29,340 --> 00:14:30,930
So if I give a little bit of history
271
00:14:30,930 --> 00:14:32,490
to the conversation
we're going to be having,
272
00:14:32,490 --> 00:14:34,980
the Veeva Pulse Field Trends Report,
273
00:14:34,980 --> 00:14:39,390
which analyzes over 600
million HP interactions
274
00:14:39,390 --> 00:14:43,350
and activities annually from more than 80%
275
00:14:43,350 --> 00:14:46,590
of commercial biopharma
field teams worldwide,
276
00:14:46,590 --> 00:14:50,670
its analysis compiles
real-time transactional data
277
00:14:50,670 --> 00:14:54,450
recorded in Veeva CRM
and Veeva Data products,
278
00:14:54,450 --> 00:14:56,910
to deliver a view of engagement activity
279
00:14:56,910 --> 00:14:59,280
across the life sciences.
280
00:14:59,280 --> 00:15:00,690
Indexed by Veeva quarterly,
281
00:15:00,690 --> 00:15:03,630
the data aims to help
companies effectively
282
00:15:03,630 --> 00:15:05,670
and accurately benchmark performance
283
00:15:05,670 --> 00:15:07,980
and to set the right actionable goals
284
00:15:07,980 --> 00:15:10,560
for continued growth and impact.
285
00:15:10,560 --> 00:15:15,090
But before we dive deeper into
the recent report's findings,
286
00:15:15,090 --> 00:15:17,110
I thought it was interesting that PMLiVE
287
00:15:17,970 --> 00:15:21,000
earlier in the year reported
that the new reality,
288
00:15:21,000 --> 00:15:23,070
so to speak, of HCP engagement,
289
00:15:23,070 --> 00:15:27,150
is that HCPs have more
options than ever before.
290
00:15:27,150 --> 00:15:28,770
And for many in the industry,
291
00:15:28,770 --> 00:15:31,830
this has made cutting through the noise
292
00:15:31,830 --> 00:15:34,410
in the online space a
significant challenge.
293
00:15:34,410 --> 00:15:38,670
Turning now to the Veeva
Pulse Field Trends Report,
294
00:15:38,670 --> 00:15:43,670
this has recently found that
HCP access has returned,
295
00:15:43,890 --> 00:15:47,820
one can perhaps say unfortunately,
to its pre-pandemic state
296
00:15:47,820 --> 00:15:52,820
with only 45% of HCPs
accessible to biopharms now.
297
00:15:53,400 --> 00:15:58,320
That's compared to 60%
a mere 18 months ago.
298
00:15:58,320 --> 00:16:02,220
And half of HCPs are more
selective across specialties,
299
00:16:02,220 --> 00:16:04,860
which exacerbates the problem,
restricting engagement
300
00:16:04,860 --> 00:16:07,410
to three or fewer companies.
301
00:16:07,410 --> 00:16:11,940
Indeed, according to the report,
immunology and dermatology
302
00:16:11,940 --> 00:16:14,160
seem to be the most
accessible at the moment,
303
00:16:14,160 --> 00:16:17,613
while internal medicine and
psychiatry are some way behind.
304
00:16:18,510 --> 00:16:21,090
So despite access challenges,
305
00:16:21,090 --> 00:16:25,440
some field teams though are
extending HCP relationships
306
00:16:25,440 --> 00:16:29,790
by becoming one importantly
connected resource
307
00:16:29,790 --> 00:16:32,760
across sales, marketing, and medical
308
00:16:32,760 --> 00:16:35,790
in order to build more relevant
and trusted relationships,
309
00:16:35,790 --> 00:16:38,100
an important word that trust,
310
00:16:38,100 --> 00:16:41,550
in other words, coordinated
engagement modeling.
311
00:16:41,550 --> 00:16:44,040
So Chris, after all that preamble,
312
00:16:44,040 --> 00:16:45,810
can you tell us more about this trend?
313
00:16:45,810 --> 00:16:47,460
- Well, first of
all, great summary.
314
00:16:47,460 --> 00:16:48,810
- Thank you.
315
00:16:48,810 --> 00:16:52,233
- I think what we're really
seeing is exactly as you said.
316
00:16:53,340 --> 00:16:55,500
Going into the pandemic,
317
00:16:55,500 --> 00:16:57,540
we were seeing these kind of numbers:
318
00:16:57,540 --> 00:17:00,180
half of all doctors not taking any visits
319
00:17:00,180 --> 00:17:02,130
from pharmaceutical companies,
320
00:17:02,130 --> 00:17:05,283
and of those, the vast majority
speaking to three or fewer.
321
00:17:06,870 --> 00:17:09,060
That obviously dropped off a cliff
322
00:17:09,060 --> 00:17:10,620
when the pandemic happened
323
00:17:10,620 --> 00:17:14,490
and we saw a pickup in digital engagement,
324
00:17:14,490 --> 00:17:16,950
and then when people were
allowed to reconnect again,
325
00:17:16,950 --> 00:17:18,774
there was a spike
326
00:17:18,774 --> 00:17:23,774
whilst doctors and pharmaceutical
companies reconnected.
327
00:17:24,540 --> 00:17:27,090
And we've seen it return, pretty much,
328
00:17:27,090 --> 00:17:29,160
to exactly where it was before.
329
00:17:29,160 --> 00:17:32,550
If you remember back from
when we were here in Madrid
330
00:17:32,550 --> 00:17:36,570
in November and we saw
the panel of doctors,
331
00:17:36,570 --> 00:17:39,150
I thought they summarized it beautifully.
332
00:17:39,150 --> 00:17:43,410
So, in their words, what they look at
333
00:17:43,410 --> 00:17:48,410
is they need information on
how to treat their patients,
334
00:17:50,067 --> 00:17:52,860
they need to know which
treatments are available
335
00:17:52,860 --> 00:17:54,330
and whether they're applicable,
336
00:17:54,330 --> 00:17:57,630
but they need to sometimes
know that almost in real-time,
337
00:17:57,630 --> 00:18:00,243
and they're looking for true education.
338
00:18:01,080 --> 00:18:04,260
So I think when you
pick apart those numbers
339
00:18:04,260 --> 00:18:06,330
and see what sits behind it,
340
00:18:06,330 --> 00:18:09,750
what you're really seeing
is those that add value
341
00:18:09,750 --> 00:18:12,603
in that relationship are
actually sought after.
342
00:18:13,440 --> 00:18:17,310
But those that have a superficial message
343
00:18:17,310 --> 00:18:21,120
or the key thing that
we heard from that panel
344
00:18:21,120 --> 00:18:24,510
was that the most frustrating thing
345
00:18:24,510 --> 00:18:29,340
is if they get interactions with companies
346
00:18:29,340 --> 00:18:31,740
that don't build off the previous one,
347
00:18:31,740 --> 00:18:35,550
and it's just hammering the
same point over and over again,
348
00:18:35,550 --> 00:18:37,410
so that then starts to point
349
00:18:37,410 --> 00:18:40,230
to if you can add value
in the relationship,
350
00:18:40,230 --> 00:18:42,720
and if there's true education,
351
00:18:42,720 --> 00:18:46,740
and if each interaction
builds off the last one,
352
00:18:46,740 --> 00:18:49,890
then you are seeing access levels.
353
00:18:49,890 --> 00:18:53,580
But
pharmaceutical
companies have to be aware
354
00:18:53,580 --> 00:18:56,490
that those doctors are
being very selective
355
00:18:56,490 --> 00:19:00,180
and they are going to get limited chances
356
00:19:00,180 --> 00:19:01,560
of building that trust,
357
00:19:01,560 --> 00:19:04,160
and when they do, they're
going to have to maintain it.
358
00:19:05,070 --> 00:19:05,903
- Exactly.
359
00:19:05,903 --> 00:19:08,550
And knowing this, knowing this right way
360
00:19:08,550 --> 00:19:13,550
to approach the HCPs and
deliver a personalized approach,
361
00:19:13,680 --> 00:19:15,944
this all depends on data, doesn't it?
362
00:19:15,944 --> 00:19:16,777
- Absolutely.
363
00:19:16,777 --> 00:19:21,330
And you touched on some really
great examples in your intro.
364
00:19:21,330 --> 00:19:26,330
So we've heard from our
customers, for example,
365
00:19:26,550 --> 00:19:28,860
that they've had instances
366
00:19:28,860 --> 00:19:31,620
where they've been
conducting clinical trials
367
00:19:31,620 --> 00:19:34,590
with real leaders in the field,
368
00:19:34,590 --> 00:19:36,393
in a particular therapeutic area,
369
00:19:37,320 --> 00:19:42,320
and then to have their
medical service leads come in
370
00:19:43,470 --> 00:19:46,200
and talk to those same doctors
371
00:19:46,200 --> 00:19:49,680
and to not know that they were
on their clinical studies.
372
00:19:49,680 --> 00:19:52,260
And that's why that
linkage between clinical
373
00:19:52,260 --> 00:19:54,180
and medical is so important,
374
00:19:54,180 --> 00:19:56,550
because there's only so many experts,
375
00:19:56,550 --> 00:20:00,240
especially as we move into specialty
376
00:20:00,240 --> 00:20:02,610
and rare and ultra-rare diseases.
377
00:20:02,610 --> 00:20:05,010
And so having strong
relationships with them
378
00:20:05,010 --> 00:20:08,250
and building off that
foundation is important.
379
00:20:08,250 --> 00:20:09,780
The other thing that we are seeing
380
00:20:09,780 --> 00:20:12,420
is that when you have those interactions,
381
00:20:12,420 --> 00:20:14,340
they do need to be
tailored to the individual,
382
00:20:14,340 --> 00:20:16,290
and that's quite difficult today.
383
00:20:16,290 --> 00:20:17,850
So what we are looking at
384
00:20:17,850 --> 00:20:22,850
is how can we combine sales,
marketing, and medical
385
00:20:23,910 --> 00:20:27,510
so that with input from clinical
386
00:20:27,510 --> 00:20:32,070
to understand who we've
spoken to about what
387
00:20:32,070 --> 00:20:33,210
and what did they value
388
00:20:33,210 --> 00:20:36,030
and how do they want to
be communicated with,
389
00:20:36,030 --> 00:20:38,850
i.e., in-person, digital, or mixture.
390
00:20:38,850 --> 00:20:42,870
And then the final element is
there was always this notion
391
00:20:42,870 --> 00:20:44,820
that it was a push message.
392
00:20:44,820 --> 00:20:49,170
So pharmaceutical companies
would reach out to doctors,
393
00:20:49,170 --> 00:20:51,270
and it would all be about, right,
394
00:20:51,270 --> 00:20:54,240
okay, I want to tell you about.
395
00:20:54,240 --> 00:20:57,420
Actually, what we hear from physicians
396
00:20:57,420 --> 00:20:59,790
is, no, no, sometimes there's time
397
00:20:59,790 --> 00:21:02,130
when I need to know the answer
398
00:21:02,130 --> 00:21:05,280
to is this drug right for my patient?
399
00:21:05,280 --> 00:21:07,770
Under what circumstances?
What can you tell me?
400
00:21:07,770 --> 00:21:09,840
Can you help me with this?
401
00:21:09,840 --> 00:21:11,970
And there, what we've been doing
402
00:21:11,970 --> 00:21:14,310
is experimenting with two-way chat,
403
00:21:14,310 --> 00:21:18,630
and with compliant chat we are seeing 30%
404
00:21:18,630 --> 00:21:22,260
of all those interactions
actually triggered by physicians
405
00:21:22,260 --> 00:21:25,579
who have a question, need it answered,
406
00:21:25,579 --> 00:21:27,900
and want it answered pretty quickly.
407
00:21:27,900 --> 00:21:31,530
And again, all of that
builds this circle of trust,
408
00:21:31,530 --> 00:21:34,230
especially if you
understand every interaction
409
00:21:34,230 --> 00:21:35,910
and can build off it.
410
00:21:35,910 --> 00:21:38,190
- Circle of trust, two-way chat,
411
00:21:38,190 --> 00:21:40,470
basically the very nature
of a human relationship.
412
00:21:40,470 --> 00:21:41,850
- There's so many companies
413
00:21:41,850 --> 00:21:46,850
and they're targeting some
fairly specialist areas,
414
00:21:47,550 --> 00:21:49,860
and doctors are human beings,
415
00:21:49,860 --> 00:21:51,990
and they will focus their efforts
416
00:21:51,990 --> 00:21:55,260
on those where they feel
that they get the most value.
417
00:21:55,260 --> 00:22:00,120
And that means that every
interaction is precious
418
00:22:00,120 --> 00:22:02,379
and it has to add value.
419
00:22:02,379 --> 00:22:03,212
- Absolutely.
420
00:22:03,212 --> 00:22:06,210
So if we think about this notion
421
00:22:06,210 --> 00:22:10,030
and perhaps what might exist
geographically differently
422
00:22:11,220 --> 00:22:13,140
when it comes to HCP access,
423
00:22:13,140 --> 00:22:14,970
considering the global trends now
424
00:22:14,970 --> 00:22:19,350
and taking into consideration
the U.S., as well as Europe,
425
00:22:19,350 --> 00:22:22,260
Asia, but also Latin
America, for instance,
426
00:22:22,260 --> 00:22:24,060
Latin America was found
427
00:22:24,060 --> 00:22:28,890
to have the lowest average
meeting duration of all regions.
428
00:22:28,890 --> 00:22:32,280
Can you perhaps tell us why
there is this discrepancy
429
00:22:32,280 --> 00:22:35,070
between the connectivity, as it were?
430
00:22:35,070 --> 00:22:39,240
- Yeah, I mean, when we
look at this across markets,
431
00:22:39,240 --> 00:22:42,030
in some places, it's easier
to understand than others.
432
00:22:42,030 --> 00:22:47,030
So, for example, as we look
into markets like the UK
433
00:22:47,160 --> 00:22:50,790
where you've got a lot
of collective agreements
434
00:22:50,790 --> 00:22:53,640
and you go through healthcare trusts,
435
00:22:53,640 --> 00:22:58,170
the direct interaction with
physicians is even more limited
436
00:22:58,170 --> 00:23:01,503
and they're even more selective
about who they talk to.
437
00:23:02,640 --> 00:23:05,820
And we are seeing those
discrepancies across all markets,
438
00:23:05,820 --> 00:23:09,180
almost with the U.S. being the benchmark
439
00:23:09,180 --> 00:23:11,313
and places like Latin America even lower,
440
00:23:12,510 --> 00:23:15,450
where you've got high prevalence
of specialty medicine,
441
00:23:15,450 --> 00:23:17,490
that's where you tend
to see where there's new
442
00:23:17,490 --> 00:23:18,660
and innovative treatments,
443
00:23:18,660 --> 00:23:21,900
that's where you tend
to see more engagement
444
00:23:21,900 --> 00:23:25,860
because the doctors need to
understand the treatments
445
00:23:25,860 --> 00:23:27,600
and how to use them.
446
00:23:27,600 --> 00:23:31,290
So I think what's interesting is, yes,
447
00:23:31,290 --> 00:23:34,440
we see discrepancies by
markets that could be cultural,
448
00:23:34,440 --> 00:23:38,880
it could be product mix,
but equally we are seeing
449
00:23:38,880 --> 00:23:43,880
that the core trend almost
goes up and down in unison.
450
00:23:44,100 --> 00:23:48,580
And it's this combination of
value in every interaction,
451
00:23:49,500 --> 00:23:52,590
building off the previous
one, and mix of channels
452
00:23:52,590 --> 00:23:55,293
that suits the individual,
including two-way.
453
00:23:56,970 --> 00:24:00,690
- So you've spoken a lot about education
454
00:24:00,690 --> 00:24:05,070
during this conversation,
so if we think pre-launch,
455
00:24:05,070 --> 00:24:07,980
what role can education of KOLs
456
00:24:07,980 --> 00:24:11,190
or field medical key opinion leaders,
457
00:24:11,190 --> 00:24:14,916
play in facilitating matters?
458
00:24:14,916 --> 00:24:19,800
- We've seen that if you target
459
00:24:19,800 --> 00:24:23,190
and provide support to those KOLs
460
00:24:23,190 --> 00:24:27,210
six months or more before
launch, differential outcomes,
461
00:24:27,210 --> 00:24:30,510
both in terms of initial prescriptions,
462
00:24:30,510 --> 00:24:34,980
but really interestingly the maintenance
463
00:24:34,980 --> 00:24:35,970
of those prescriptions,
464
00:24:35,970 --> 00:24:38,910
so the adherence to it going forward.
465
00:24:38,910 --> 00:24:40,980
And if you take a step back
466
00:24:40,980 --> 00:24:43,110
and just look at it from
a logical perspective,
467
00:24:43,110 --> 00:24:44,970
it kind of makes sense.
468
00:24:44,970 --> 00:24:49,410
If you are educating the market
to how it should be used,
469
00:24:49,410 --> 00:24:51,870
when it should be used, what
are the right patient sets,
470
00:24:51,870 --> 00:24:54,360
you're going to get better outcomes.
471
00:24:54,360 --> 00:24:57,322
And those KOLs are so, so important
472
00:24:57,322 --> 00:25:02,322
to educating the medical
community as a whole.
473
00:25:03,030 --> 00:25:05,880
So understanding who they are,
474
00:25:05,880 --> 00:25:09,030
understanding how they
like to communicate,
475
00:25:09,030 --> 00:25:12,037
and understanding any
track record or history
476
00:25:12,037 --> 00:25:14,310
does what we touched on previously,
477
00:25:14,310 --> 00:25:17,220
which is building that
trust-based relationship.
478
00:25:17,220 --> 00:25:19,410
And so what we are
seeing is those companies
479
00:25:19,410 --> 00:25:21,210
that understand and target that
480
00:25:21,210 --> 00:25:25,680
get considerably better
differential outcomes.
481
00:25:25,680 --> 00:25:29,340
- Now, is there a takeaway
from our conversation,
482
00:25:29,340 --> 00:25:32,820
something that can be summarized
in maybe not one sentence,
483
00:25:32,820 --> 00:25:34,290
but a couple of sentences?
484
00:25:34,290 --> 00:25:38,190
- Yeah, I think it's this core theme
485
00:25:38,190 --> 00:25:41,430
of every interaction has to matter
486
00:25:41,430 --> 00:25:44,250
and it should build off the previous one.
487
00:25:44,250 --> 00:25:47,460
So our advice to all of our customers
488
00:25:47,460 --> 00:25:50,850
is to look very carefully at the needs
489
00:25:50,850 --> 00:25:54,240
of your key opinion leaders and physicians
490
00:25:54,240 --> 00:25:57,870
and make sure you are
directly targeting their needs
491
00:25:57,870 --> 00:26:00,060
in their context, in the way
492
00:26:00,060 --> 00:26:02,970
that they want to be communicated with.
493
00:26:02,970 --> 00:26:06,540
And in doing that, when
you take the perspective
494
00:26:06,540 --> 00:26:11,010
of that doctor, they lead very busy lives,
495
00:26:11,010 --> 00:26:14,220
that they're doing important
work treating patients,
496
00:26:14,220 --> 00:26:17,640
and you are giving them the data in a form
497
00:26:17,640 --> 00:26:19,560
that's consumable and useful
498
00:26:19,560 --> 00:26:21,810
in the context of what they're doing,
499
00:26:21,810 --> 00:26:26,070
not in some kind of mass
email or mass splurge.
500
00:26:26,070 --> 00:26:29,010
So that targeting is everything.
501
00:26:29,010 --> 00:26:30,960
- Targeting is everything.
- Yeah.
502
00:26:30,960 --> 00:26:32,310
- I'll take that as your summation.
503
00:26:32,310 --> 00:26:34,590
Thank you very much, Chris.
It's been a pleasure.
504
00:26:34,590 --> 00:26:37,090
- Thank you very much, Nicole.