Popular in Professional

Leading Figures

Induction of Social Contagion

Yale University Nicholas A. Christakis

Business

The Allegory of the Favela: The Multifaceted Effects of Socioeconomic Mobility

University of Pennsylvania Leandro S. Pongeluppe

Architecture

Connecting Cities and Nature

University of Virginia Timothy Beatley

Human Rights Law

Human Rights and Corruption: Problems and Potential of Individualising a Systemic Problem

Max Planck Institute Anne Peters

Finance

Chinese Exchange Rate Policy: Lessons for Global Investors

University of California San Diego Michael Melvin

Law

A statistical approach to law school citation rankings

University of Virginia Joshua Fischman

Education

Measuring the influence of expectations, beliefs, and medication side effects

View more insights in Professional - Pharmacy

View more figures from University of Saskatchewan

View the page of David Blackburn

Read the Study

Next in Pharmacy

The many faces of the AMPA-type ionotropic glutamate receptor

Derek Bowie | McGill University

Cytotoxic Activity and Anti-Migratory Effect of Berberine-Phytantriol Liquid Crystalline on Non-Small-Cell Lung Cancer

Kamal Dua | University of Technology Sydney

David Blackburn discusses the impact of beliefs, expectations, side effects, and their combined effects on the risk for medication non‐ persistence.

Image courtesy of interviewee. March 6, 2024

Log-in or Sign-up to Faculti
Currently viewing this subject insight as a guest. You have insight(s) remaining for this month. Login to view 8000+ figures on the platform.

Generating Workspace...

Export as PDF

This video video describes a study aimed at improving adherence prediction models in healthcare, particularly concerning patients prescribed cholesterol-lowering, blood pressure-lowering, or glucose-lowering medications. The research was initiated before the COVID-19 pandemic, which disrupted some of the data collection processes.

The project involved merging patient survey responses with administrative databases to better understand non-adherence drivers. The surveys yielded a significant response, indicating public engagement, but the pandemic's impact delayed further data integration.

Key findings reveal that expectations of side effects play a crucial role in medication adherence. Specifically, only 25% of participants expected side effects, contrary to the belief that patients are overly informed about potential adverse effects. The expectation of side effects correlated differently with adherence rates; those who anticipated side effects but experienced them showed lower non-adherence rates than those who did not expect them and then suffered from side effects.

The study highlights the need to reevaluate how healthcare providers communicate about side effects, suggesting that a more balanced approach—acknowledging the inevitability of some side effects while fostering a realistic understanding of their commonality—might improve patient adherence to prescribed treatments. The author emphasizes the importance of preparing patients for the likelihood of side effects and engaging in more meaningful conversations regarding medication expectations and management.

Language:

Export as PDF

X-TIMESTAMP-MAP=LOCAL:00:00:00.000,MPEGTS:90000

1
00:00:00.075 --> 00:00:03.645
This was a study, it was sort of part of a broader project

2
00:00:03.755 --> 00:00:06.845
that we initiate years ago, right before Covid actually.

3
00:00:07.505 --> 00:00:11.485
And, we, we had done a ton of work,

4
00:00:11.535 --> 00:00:15.285
using administrative databases to try to, to try

5
00:00:15.285 --> 00:00:17.805
to create adherence prediction models.

6
00:00:18.665 --> 00:00:21.165
Um, so we were looking at adherence at the, at the,

7
00:00:21.185 --> 00:00:24.325
at the population level, most frequently.

8
00:00:24.385 --> 00:00:26.885
And we were, you know, we were really good at finding

9
00:00:27.365 --> 00:00:31.445
discreet factors that were associated with non-adherence,

10
00:00:32.065 --> 00:00:33.605
but when we put our models together,

11
00:00:34.145 --> 00:00:37.405
the overall prediction performance was always terrible.

12
00:00:37.625 --> 00:00:41.405
So, so, you know, it was, it was not a problem to identify,

13
00:00:41.705 --> 00:00:44.485
you know, five or seven variables

14
00:00:44.915 --> 00:00:47.045
that were significantly associated.

15
00:00:47.145 --> 00:00:48.285
But, but, but,

16
00:00:48.585 --> 00:00:51.205
but to actually predict, you know, to put those five

17
00:00:51.205 --> 00:00:52.405
or seven variables or, or

18
00:00:52.405 --> 00:00:56.205
however many variables we had, which was usually a lot,

19
00:00:56.705 --> 00:00:59.205
you know, to, to actually put those together to see,

20
00:00:59.385 --> 00:01:02.365
you know, what percentage of the non-adherence

21
00:01:03.205 --> 00:01:04.605
variants we could explain.

22
00:01:04.865 --> 00:01:05.965
We could never explain,

23
00:01:06.265 --> 00:01:08.045
we always could explain just a fraction

24
00:01:08.345 --> 00:01:10.205
of the non-adherence we observed in these

25
00:01:10.205 --> 00:01:11.485
population-based databases.

26
00:01:11.665 --> 00:01:13.445
So that's what led to this.

27
00:01:13.465 --> 00:01:18.005
We said, man, what if we could, you know,

28
00:01:18.265 --> 00:01:20.285
give these people a survey

29
00:01:20.545 --> 00:01:23.965
and have patient responses to merge

30
00:01:23.965 --> 00:01:25.725
with our administrative databases?

31
00:01:26.305 --> 00:01:28.205
And, and, and that's where it started.

32
00:01:28.425 --> 00:01:30.485
So we, so, so that was the,

33
00:01:30.625 --> 00:01:33.445
the overall grand vision of this project.

34
00:01:33.665 --> 00:01:34.885
And we got funding for it,

35
00:01:34.905 --> 00:01:37.245
and we got partnership with our, with our,

36
00:01:37.735 --> 00:01:38.805
provincial government.

37
00:01:39.265 --> 00:01:43.085
Um, and then, we just sent out the, we were halfway

38
00:01:43.085 --> 00:01:45.765
through sending out the surveys and covid hit, and,

39
00:01:45.825 --> 00:01:48.085
and our drug plan is still recovering from covid.

40
00:01:48.345 --> 00:01:51.245
And so, so the administrative data side,

41
00:01:51.425 --> 00:01:54.645
we still haven't, there's still more to go

42
00:01:54.645 --> 00:01:56.685
because we still haven't linked our administrative data

43
00:01:56.705 --> 00:02:00.005
to this, but we thought, man, we had, you know,

44
00:02:00.005 --> 00:02:01.205
an amazing response.

45
00:02:01.585 --> 00:02:04.525
We sent out 11,000 surveys. We got 4,000 back.

46
00:02:05.145 --> 00:02:07.285
Uh, just, this is population wide,

47
00:02:07.385 --> 00:02:09.005
so people just, just responding.

48
00:02:09.005 --> 00:02:12.165
We, we, we gave them no, incentive.

49
00:02:12.465 --> 00:02:15.085
Uh, these are just good citizens responding to our survey.

50
00:02:15.145 --> 00:02:16.685
And, I can tell you more about that later.

51
00:02:16.945 --> 00:02:21.805
But the, you know, that's where we, that's

52
00:02:21.805 --> 00:02:22.845
where we started and there,

53
00:02:22.865 --> 00:02:24.965
and we're, I guess we're sort of halfway through and,

54
00:02:25.145 --> 00:02:26.965
and these couple papers we've done are, are

55
00:02:26.965 --> 00:02:31.325
where we're at now, this current project, I think to me,

56
00:02:31.585 --> 00:02:35.525
really, it really makes us look at this idea

57
00:02:35.745 --> 00:02:37.125
of expectations.

58
00:02:37.505 --> 00:02:40.085
The, the expectations side of our findings was,

59
00:02:40.225 --> 00:02:41.885
was most interesting to me.

60
00:02:42.345 --> 00:02:46.885
So, so where this sits, I think, is we connect this idea

61
00:02:47.105 --> 00:02:49.445
of expectations for side effects, which,

62
00:02:49.445 --> 00:02:53.005
if you look at a lot of literature is, is,

63
00:02:53.005 --> 00:02:55.165
setting people up for no SIBO effects.

64
00:02:55.265 --> 00:02:58.205
You know, the idea, there's a lot of literature that says,

65
00:02:58.235 --> 00:03:01.405
geez, we should be careful about telling people

66
00:03:01.475 --> 00:03:04.245
what side effects to expect because they're gonna get them.

67
00:03:05.065 --> 00:03:07.765
Um, and, and so we call that an no SIBO effects, and,

68
00:03:07.825 --> 00:03:09.445
and we assume that that's a, that

69
00:03:09.445 --> 00:03:11.725
that creates a pessimistic view of drugs.

70
00:03:12.185 --> 00:03:13.525
If you're expecting a side effect,

71
00:03:13.525 --> 00:03:16.085
and then you'll get it, you'll, you'll,

72
00:03:16.085 --> 00:03:19.285
you'll physically sense a, a, a side effect

73
00:03:19.285 --> 00:03:21.245
or the side effect that maybe you wouldn't

74
00:03:21.245 --> 00:03:22.325
have if you didn't expect it.

75
00:03:22.545 --> 00:03:23.605
That's an no sibo.

76
00:03:24.265 --> 00:03:27.445
Um, and, and so we associate these expectations

77
00:03:27.465 --> 00:03:31.005
for side effects as a, you know, as,

78
00:03:31.065 --> 00:03:34.365
as pessimistic thinking, as, as sort of a, a harbinger of,

79
00:03:34.585 --> 00:03:36.485
of, of a poor outcome of somebody

80
00:03:36.485 --> 00:03:39.525
that's gonna end up being non likely to be non-adherent

81
00:03:39.525 --> 00:03:41.565
because they're, they already have negative

82
00:03:41.565 --> 00:03:42.845
expectations for these drugs.

83
00:03:43.285 --> 00:03:47.965
I think our findings really provide a little bit

84
00:03:47.965 --> 00:03:51.665
of a different perspective on what we, you know,

85
00:03:51.665 --> 00:03:53.385
currently think about expectations.

86
00:03:53.775 --> 00:03:58.225
What we did was, we partnered with our,

87
00:03:58.565 --> 00:04:03.145
our provincial insurance drug insurance, who is,

88
00:04:03.525 --> 00:04:04.705
you know, I, I live in Canada.

89
00:04:04.925 --> 00:04:09.145
We have sort of what we call universal, healthcare,

90
00:04:09.235 --> 00:04:11.625
which maybe it isn't technically, but it's close.

91
00:04:12.725 --> 00:04:17.505
Um, and so our, our drug provider here,

92
00:04:17.645 --> 00:04:21.345
has about 90% of our province are considered beneficiaries

93
00:04:21.365 --> 00:04:22.425
of our drug plan.

94
00:04:23.365 --> 00:04:26.585
And so we partnered with our provincial drug plan.

95
00:04:26.645 --> 00:04:30.585
We said, you guys, we,

96
00:04:30.605 --> 00:04:31.745
we had a list.

97
00:04:32.125 --> 00:04:34.265
Uh, they, they created this all on their own,

98
00:04:34.415 --> 00:04:35.865
obviously independent of us.

99
00:04:36.645 --> 00:04:41.425
And, they, they received claims in real time

100
00:04:42.325 --> 00:04:45.945
for all prescription drugs dispensed from pharmacies in

101
00:04:45.945 --> 00:04:47.385
our, in our province.

102
00:04:48.085 --> 00:04:51.505
And so what they did is they identified all

103
00:04:51.505 --> 00:04:56.025
of the individuals who received a first prescription

104
00:04:56.525 --> 00:04:59.025
for a cholesterol lowering, blood pressure lowering

105
00:04:59.085 --> 00:05:01.065
or glucose lowering drug.

106
00:05:01.605 --> 00:05:04.465
So in, in, not exactly real time,

107
00:05:04.565 --> 00:05:06.185
but about every six weeks

108
00:05:06.365 --> 00:05:11.025
or so, they would run the report of all people that got,

109
00:05:11.085 --> 00:05:12.985
you know, one of those three classes of drugs.

110
00:05:13.405 --> 00:05:16.905
And we, we, we had a sort of a cohort people of,

111
00:05:17.325 --> 00:05:18.905
we investigated their history

112
00:05:19.325 --> 00:05:22.145
and they had to have at least two years of,

113
00:05:22.645 --> 00:05:24.025
of beneficiary status.

114
00:05:24.125 --> 00:05:25.505
So we had, we had to have knowledge of 'em

115
00:05:25.505 --> 00:05:28.305
for the last two years, and they couldn't have had any

116
00:05:28.915 --> 00:05:31.945
prior fill for whatever, you know, for one

117
00:05:31.945 --> 00:05:33.265
of those three drug families.

118
00:05:33.445 --> 00:05:34.785
So we call them new users.

119
00:05:35.065 --> 00:05:36.545
I think for the most part, they probably were.

120
00:05:37.245 --> 00:05:41.145
Um, and so if they met that criteria at the insurer,

121
00:05:41.645 --> 00:05:46.025
the insurer mailed out our SER survey that we gave to them,

122
00:05:46.735 --> 00:05:47.945
they mailed out the survey,

123
00:05:48.485 --> 00:05:51.945
the envelope had our return address on it.

124
00:05:52.165 --> 00:05:53.665
So, you know, typical survey.

125
00:05:54.085 --> 00:05:56.745
Um, so this is a typical cross-sectional survey of drug

126
00:05:56.745 --> 00:05:57.745
Users. But the thing that

127
00:05:57.745 --> 00:05:59.085
really distinguished our

128
00:05:59.085 --> 00:06:00.525
project, I think is the timing.

129
00:06:01.185 --> 00:06:02.765
So we really captured them.

130
00:06:03.225 --> 00:06:06.325
We probably, you know, captured them somewhere between,

131
00:06:07.345 --> 00:06:09.605
you know, two to three weeks

132
00:06:09.745 --> 00:06:13.245
and probably four months of when they first

133
00:06:13.985 --> 00:06:17.725
got a dispensation for their first drug at a pharmacy.

134
00:06:18.145 --> 00:06:21.645
And then the, and then this big tool, we had a, we,

135
00:06:21.705 --> 00:06:26.245
we created a, a survey, that was

136
00:06:27.125 --> 00:06:29.325
designed for people who were new users

137
00:06:29.345 --> 00:06:30.365
of one of those three drugs.

138
00:06:30.985 --> 00:06:35.005
Um, we designed it, really, we,

139
00:06:35.005 --> 00:06:36.445
we had 52 questions.

140
00:06:36.785 --> 00:06:38.605
We, we wanted, you know, we real,

141
00:06:38.625 --> 00:06:41.285
as you can imagine in this adherence business, we wanted

142
00:06:41.285 --> 00:06:43.645
to ask 150 questions,

143
00:06:44.145 --> 00:06:46.525
but it was just even at 52, we were quite nervous.

144
00:06:47.185 --> 00:06:50.645
So we really condensed a lot of the, you know, the,

145
00:06:50.645 --> 00:06:54.285
the traditional batteries of belief scales

146
00:06:54.425 --> 00:06:56.205
and, knowledge scales.

147
00:06:56.205 --> 00:06:59.125
And we, we really had to make hard decisions and, and, and,

148
00:06:59.145 --> 00:07:01.645
and condense those questions so we could get a breadth

149
00:07:01.645 --> 00:07:02.885
of information on these people.

150
00:07:03.785 --> 00:07:06.445
Um, and then ultimately, you know, we said, you know,

151
00:07:06.445 --> 00:07:08.645
what are one of the key questions for our outcome?

152
00:07:08.665 --> 00:07:11.245
And this study was, you know, what, what about today?

153
00:07:11.545 --> 00:07:13.685
You know, have you, are you still taking your drug today?

154
00:07:13.685 --> 00:07:16.445
Or have you stopped it? And that was sort of ended up our,

155
00:07:16.505 --> 00:07:19.005
our outcome of non persistence were the people

156
00:07:19.105 --> 00:07:21.325
who had stopped that or said, you know, as

157
00:07:21.325 --> 00:07:23.965
of today filling this out, I, I'm not taking it anymore.

158
00:07:24.265 --> 00:07:25.605
We sent out 12,000.

159
00:07:26.065 --> 00:07:27.365
We didn't send out our, the, the,

160
00:07:27.385 --> 00:07:29.885
the provider sent out 12,000 surveys.

161
00:07:30.005 --> 00:07:32.845
We got about 4,000 back, which we're very happy with.

162
00:07:33.185 --> 00:07:36.485
We, we would've kept going, but we, but covid hit.

163
00:07:37.025 --> 00:07:38.685
And so priorities on the,

164
00:07:38.715 --> 00:07:41.645
obviously is on the insurance provider, was, you know,

165
00:07:41.645 --> 00:07:45.085
we gotta, guys, we, we can't be sending out letters for,

166
00:07:45.465 --> 00:07:48.925
for, for this discrete project at, at the time like this.

167
00:07:48.945 --> 00:07:50.845
And they had many other things they had to deal with.

168
00:07:50.905 --> 00:07:53.845
So, so anyways, we ended up with, yeah, about 4,000 returns.

169
00:07:54.345 --> 00:07:56.965
The, a a a couple of just descriptive pieces

170
00:07:56.995 --> 00:07:58.085
that really interested me.

171
00:07:58.825 --> 00:08:03.245
One was, so we had three questions that we, we tried

172
00:08:03.245 --> 00:08:06.605
to capture beliefs, again, beliefs, you know,

173
00:08:06.625 --> 00:08:09.965
and when, when you look at the, the, the, the validated way

174
00:08:09.965 --> 00:08:11.885
to capture beliefs, you, you probably have to have

175
00:08:12.485 --> 00:08:15.165
y you know, a numerous battery of questions.

176
00:08:15.165 --> 00:08:20.005
We had three, we, when we looked at that, a very,

177
00:08:20.195 --> 00:08:23.645
very small percentage of people were, we were categorized

178
00:08:23.645 --> 00:08:26.685
as having negative beliefs before starting the drug.

179
00:08:26.705 --> 00:08:29.485
And that's how we asked it said, you know, you know, what,

180
00:08:29.705 --> 00:08:30.925
how do you feel about this drug?

181
00:08:30.945 --> 00:08:35.885
And, very, very lower than expected

182
00:08:36.685 --> 00:08:39.355
percentage of people who had negative beliefs going in.

183
00:08:39.355 --> 00:08:40.595
So they were all seemed,

184
00:08:40.595 --> 00:08:44.195
most people on average were very positive about the drugs,

185
00:08:44.695 --> 00:08:46.355
you know, using the questions we used.

186
00:08:47.015 --> 00:08:49.715
Um, the other thing that we found very interesting

187
00:08:50.325 --> 00:08:53.235
after the fact, after this idea of, we had one question

188
00:08:53.235 --> 00:08:56.855
that said,

189
00:08:57.275 --> 00:09:02.215
before you started this medication, did you expect to,

190
00:09:03.155 --> 00:09:04.935
did you, did you expect side effects?

191
00:09:05.035 --> 00:09:07.135
Or, you know, this, I can't remember the wording exactly,

192
00:09:07.195 --> 00:09:09.255
but before you started, did you expect side effects?

193
00:09:09.515 --> 00:09:13.175
And only 25% of people expected

194
00:09:13.195 --> 00:09:15.175
to get side effects from this drug.

195
00:09:15.475 --> 00:09:17.895
You know, these 4,000 people only, only a quarter

196
00:09:17.895 --> 00:09:20.895
of 'em actually expected to get drug side effects

197
00:09:20.895 --> 00:09:21.895
before starting the drug.

198
00:09:22.835 --> 00:09:25.695
So that, that was very, very surprising to me.

199
00:09:25.875 --> 00:09:28.535
Uh, I, I'm a, I'm a a professor of pharmacy,

200
00:09:28.535 --> 00:09:30.655
and sometimes I think that's all we do is tell our students

201
00:09:30.795 --> 00:09:34.255
to tell everybody what side effects they're gonna get.

202
00:09:34.315 --> 00:09:36.895
You know, that's your, your role in life is to tell people

203
00:09:36.895 --> 00:09:37.975
what side effects they're gonna get.

204
00:09:38.075 --> 00:09:42.055
So, so we, we have our, our people here tuned up,

205
00:09:42.075 --> 00:09:44.415
and we, we, I thought,

206
00:09:44.555 --> 00:09:46.415
or I think that, you know, that, that,

207
00:09:46.415 --> 00:09:48.735
especially on the pharmacy side, on the dispensation side,

208
00:09:48.735 --> 00:09:53.015
that we, we prime people who say, geez, you know, this,

209
00:09:53.045 --> 00:09:54.335
this, there's gonna be side effects.

210
00:09:54.395 --> 00:09:56.415
So, so quarter people expect it.

211
00:09:57.235 --> 00:09:58.655
Um, and,

212
00:09:58.755 --> 00:10:03.175
and just as expected, of those quarter of people

213
00:10:03.175 --> 00:10:06.615
that expected to get side effects, they were more likely

214
00:10:06.615 --> 00:10:08.295
to get side effects than people who didn't expect.

215
00:10:08.295 --> 00:10:12.335
So there's our no SIBO effect replicated in our, in our,

216
00:10:12.595 --> 00:10:13.775
you know, real world survey.

217
00:10:14.715 --> 00:10:19.625
Um, but then when we started to model all of this,

218
00:10:19.965 --> 00:10:22.505
all of these factors, and we had, we had numerous factors

219
00:10:22.535 --> 00:10:26.425
that we modeled, a couple interesting things.

220
00:10:27.085 --> 00:10:31.145
You know, first off, the drivers of non persistence

221
00:10:31.145 --> 00:10:33.025
and that early non persistence, the people

222
00:10:33.025 --> 00:10:35.865
that quit already, the main drivers side effects

223
00:10:36.045 --> 00:10:38.445
and negative beliefs, bar none.

224
00:10:38.985 --> 00:10:40.605
So huge odds ratios.

225
00:10:41.145 --> 00:10:43.885
And, and, and, when you throw those in the model,

226
00:10:43.955 --> 00:10:47.365
they make all of these other sort of, you know, discreet

227
00:10:47.995 --> 00:10:51.005
factors that we throw in there, forgetfulness and,

228
00:10:51.665 --> 00:10:54.605
and, cost related, non-adherence, all, all this sort

229
00:10:54.605 --> 00:10:57.005
of things, all those get blown out.

230
00:10:57.005 --> 00:10:58.765
It get diluted by the, the, the,

231
00:10:58.785 --> 00:11:00.725
the overpowering effect of side.

232
00:11:00.745 --> 00:11:02.005
Did you get side effects? Yes.

233
00:11:02.545 --> 00:11:03.685
Or do you have negative beliefs?

234
00:11:03.905 --> 00:11:06.165
Yes, those two overpowered everything.

235
00:11:07.025 --> 00:11:10.125
Um, and then this expectations

236
00:11:10.185 --> 00:11:14.205
for side effects was totally unexpected to us,

237
00:11:14.315 --> 00:11:17.565
performed differently in the univariate analysis,

238
00:11:17.585 --> 00:11:20.125
in the univariate analysis, an expectation for side effects

239
00:11:20.715 --> 00:11:23.685
increased your risk of non-adherence, non persistence.

240
00:11:24.705 --> 00:11:28.445
In the multi-variate, there was an interaction we found,

241
00:11:28.715 --> 00:11:32.125
when we ran it, it, it, the odds ratio flipped from a 1 3 7

242
00:11:32.145 --> 00:11:33.165
to 0.57.

243
00:11:33.905 --> 00:11:36.605
So we, we searched for interaction with everything,

244
00:11:36.665 --> 00:11:38.645
and we found this interaction with side effects.

245
00:11:39.185 --> 00:11:43.705
So as it turns out, if you,

246
00:11:44.205 --> 00:11:47.505
if you experience side effects from your medication,

247
00:11:47.505 --> 00:11:50.425
obviously your, your risk of non persistence is much higher.

248
00:11:51.125 --> 00:11:54.085
But if you expected to get side effects

249
00:11:54.705 --> 00:11:59.045
before having a real side effect, your risk

250
00:11:59.105 --> 00:12:01.885
of non persistence was, was half.

251
00:12:02.945 --> 00:12:04.005
If you, for for,

252
00:12:04.105 --> 00:12:07.045
and I'll just back up in case it wasn't clear, sorry, if,

253
00:12:07.065 --> 00:12:09.885
if, if you had side effects, but you expected them

254
00:12:09.885 --> 00:12:14.405
before, your risk of non-serious persistence was half

255
00:12:15.445 --> 00:12:17.645
compared to if you, if you experienced side effects

256
00:12:17.705 --> 00:12:19.645
but hadn't expected them prior.

257
00:12:20.335 --> 00:12:22.805
Those people, the people that had side effects

258
00:12:22.865 --> 00:12:24.005
but hadn't expected them,

259
00:12:24.375 --> 00:12:28.845
their non persistence rate was double those who, who,

260
00:12:29.785 --> 00:12:32.085
had expected them when they, you know, so, so,

261
00:12:32.385 --> 00:12:36.005
so side effects were, were associated with non persistence.

262
00:12:36.005 --> 00:12:37.485
Either way, they increased your risk

263
00:12:37.585 --> 00:12:40.405
of non persistence no matter what subgroup we looked at.

264
00:12:40.745 --> 00:12:43.605
But, but it seemed like the expectations sort

265
00:12:43.605 --> 00:12:45.445
of attenuated the risk.

266
00:12:46.505 --> 00:12:48.845
Um, and there's lots of ways to look at this and,

267
00:12:48.845 --> 00:12:50.765
and so I don't wanna oversimplify this, but,

268
00:12:50.905 --> 00:12:54.645
but that was sort of, to me, the most interesting part of,

269
00:12:55.345 --> 00:12:58.525
of that, that result of, of side effects.

270
00:12:58.805 --> 00:13:00.565
O other than all the other, you know,

271
00:13:00.565 --> 00:13:02.165
the other interesting things I just told,

272
00:13:02.275 --> 00:13:04.125
this is my intuition, I'll tell you, but it,

273
00:13:04.125 --> 00:13:07.845
but it's not founded in psychological psychologic theory.

274
00:13:07.865 --> 00:13:10.605
And so I'm such an amateur once we start getting into the

275
00:13:10.605 --> 00:13:13.245
theory pieces, I'm, I'm, I'm such an amateur and, and,

276
00:13:13.265 --> 00:13:15.005
and right now we're planning next steps.

277
00:13:15.125 --> 00:13:16.525
I think there's lots to learn here.

278
00:13:16.525 --> 00:13:20.125
Here's what I, here's what I think might be happening is,

279
00:13:20.625 --> 00:13:24.045
is I, I told you before, you know that healthcare is,

280
00:13:24.045 --> 00:13:27.045
especially in the, in the professional pharmacy, I think we,

281
00:13:27.345 --> 00:13:29.885
you know, we're there, there's a duty to warn.

282
00:13:30.085 --> 00:13:32.445
I think legally in a lot of, in a lot of locations,

283
00:13:32.445 --> 00:13:34.085
there's a duty to warn of what,

284
00:13:34.115 --> 00:13:35.125
what are the negative effects

285
00:13:35.125 --> 00:13:36.965
of drugs when you're dispensing them?

286
00:13:36.965 --> 00:13:39.085
And I think we're, and I think, you know,

287
00:13:39.085 --> 00:13:40.565
we're pretty good in the healthcare system of,

288
00:13:40.905 --> 00:13:42.165
of warning people about drugs.

289
00:13:42.385 --> 00:13:46.045
But, but what I also think could be going on here,

290
00:13:46.225 --> 00:13:47.965
and that might explain this,

291
00:13:48.475 --> 00:13:52.445
only 25% expect side effects is we say we tend

292
00:13:52.445 --> 00:13:57.175
to really soften the information about drug

293
00:13:57.195 --> 00:13:58.255
safety to people.

294
00:13:58.455 --> 00:14:02.135
I, I think healthcare providers typically are so motivated

295
00:14:03.275 --> 00:14:05.175
in, in, or, or,

296
00:14:05.175 --> 00:14:08.175
or they're, they're so focused on motivating patients

297
00:14:08.515 --> 00:14:10.775
to buy in to a, to a therapy.

298
00:14:10.875 --> 00:14:13.655
And I think they do this with all good intentions.

299
00:14:13.715 --> 00:14:17.255
You know, I, this isn't a pharmaceutical industry conspiracy

300
00:14:17.395 --> 00:14:18.415
or trying to sell more drug.

301
00:14:18.495 --> 00:14:20.215
I think, I think healthcare providers are,

302
00:14:20.395 --> 00:14:21.935
are really focused on trying

303
00:14:21.935 --> 00:14:23.455
to have a positive spin on drugs.

304
00:14:23.475 --> 00:14:26.415
And some of my colleagues, you know, and,

305
00:14:26.415 --> 00:14:28.415
and some of the, the, the, the, the sort

306
00:14:28.415 --> 00:14:29.615
of small scale studies

307
00:14:29.675 --> 00:14:32.815
before look at say, well, you know, what, if you frame

308
00:14:33.725 --> 00:14:35.135
side effect conversations

309
00:14:35.195 --> 00:14:39.535
and you say, well, rather than 5%, you know, could get

310
00:14:40.815 --> 00:14:41.975
diarrhea something,

311
00:14:42.075 --> 00:14:45.705
why don't we say 95% don't get diarrhea?

312
00:14:46.085 --> 00:14:48.865
You know, what if what if we frame our side

313
00:14:48.865 --> 00:14:49.905
effect expectations?

314
00:14:50.285 --> 00:14:52.405
And so this is where I think, this is my hunch.

315
00:14:52.405 --> 00:14:55.245
It's totally you, you, you know, totally unverified.

316
00:14:55.345 --> 00:14:59.565
But what, what if we say, what if,

317
00:14:59.635 --> 00:15:01.525
what if we're too positive to people?

318
00:15:01.585 --> 00:15:03.485
So we say, you know, side effects happen.

319
00:15:03.655 --> 00:15:04.725
These side effects could happen,

320
00:15:04.785 --> 00:15:06.645
but don't worry, this won't happen to you.

321
00:15:07.825 --> 00:15:10.205
Uh, so I, I wonder if everybody's walking out

322
00:15:10.205 --> 00:15:11.965
of there knowing that side effects can happen,

323
00:15:12.025 --> 00:15:13.725
but they all think that it's not gonna happen to them

324
00:15:13.725 --> 00:15:15.045
because we, you know, we're trying

325
00:15:15.045 --> 00:15:16.325
to paint this flowery picture.

326
00:15:16.625 --> 00:15:20.125
And, and to some respect, I wonder if we're failing

327
00:15:20.345 --> 00:15:23.845
to prepare people for the inevitable

328
00:15:24.735 --> 00:15:27.005
occurrence of, of side effects, most

329
00:15:27.005 --> 00:15:29.565
of our side effects are indeed nuisance side effects,

330
00:15:29.565 --> 00:15:31.965
where most of the things that happen from these run

331
00:15:31.965 --> 00:15:33.605
of the mill drugs are, are not severe.

332
00:15:34.105 --> 00:15:35.645
Um, but, but they do happen.

333
00:15:35.645 --> 00:15:39.445
They happen in, in, in our, in our survey, you know, 25%

334
00:15:39.445 --> 00:15:41.965
of people reported experiencing some sort

335
00:15:41.965 --> 00:15:43.285
of fi side effect in this window

336
00:15:43.345 --> 00:15:45.125
of when this survey came out.

337
00:15:45.785 --> 00:15:49.965
So, so that's my hunch. I I think maybe we're too positive.

338
00:15:50.045 --> 00:15:53.925
I think maybe, you know, we need to be more, more clear

339
00:15:54.695 --> 00:15:56.685
about the inevitability of side effects

340
00:15:56.745 --> 00:15:59.245
and we need to prepare people, you know, you know, where's,

341
00:15:59.695 --> 00:16:01.005
where is your commitment level?

342
00:16:01.225 --> 00:16:03.085
You know, where are you at if that, if is

343
00:16:03.085 --> 00:16:07.645
that side effect gonna, gonna crush your whole framework

344
00:16:07.645 --> 00:16:08.965
of why you're taking that drug?

345
00:16:08.965 --> 00:16:10.805
And all of a sudden the balance is tipped the

346
00:16:10.805 --> 00:16:11.845
other direction and I'm done.

347
00:16:12.345 --> 00:16:14.325
Um, so, so I, I don't know, I think that's

348
00:16:14.325 --> 00:16:15.525
where we're going anyways.

349
00:16:15.585 --> 00:16:19.245
And looking at just sort of, you know, just,

350
00:16:19.245 --> 00:16:22.205
just this hunch about where your beliefs start,

351
00:16:22.515 --> 00:16:24.445
when you get side effects, what they do,

352
00:16:24.555 --> 00:16:26.885
what those side effects, how they can change beliefs and,

353
00:16:27.385 --> 00:16:31.525
and are people indeed ready to start these lifelong drugs

354
00:16:32.115 --> 00:16:34.005
without maybe proper preparation.

355
00:16:34.055 --> 00:16:36.405
Since this has come out, I, I think differently about this.

356
00:16:36.605 --> 00:16:37.685
I used to be that guy.

357
00:16:37.845 --> 00:16:41.085
I used to be the guy that said, we don't need

358
00:16:41.085 --> 00:16:44.445
to be scaring people about starting a blood pressure drug.

359
00:16:44.455 --> 00:16:46.485
Blood pressure drugs are insanely safe,

360
00:16:47.065 --> 00:16:48.365
for, for most people.

361
00:16:48.425 --> 00:16:49.845
You know, we, we know the risks.

362
00:16:50.045 --> 00:16:51.085
I think we can avoid the risks,

363
00:16:51.505 --> 00:16:55.765
but I think that, that, you know, in our good,

364
00:16:55.785 --> 00:17:00.205
in our well-intentioned approach to, to try

365
00:17:00.205 --> 00:17:03.685
to calm people's nerves, it's scary

366
00:17:03.685 --> 00:17:04.725
to start new drugs, right?

367
00:17:04.775 --> 00:17:08.085
Scary, you know, it, it, it's, it's difficult

368
00:17:08.085 --> 00:17:09.085
for people to get their head

369
00:17:09.085 --> 00:17:11.765
around taking a chemical drug every day.

370
00:17:12.465 --> 00:17:16.085
So, so I used to be on that camp of we need to

371
00:17:16.765 --> 00:17:20.725
reassure people that this is not some major, you know,

372
00:17:21.165 --> 00:17:23.365
chemical entity that's gonna change, you know,

373
00:17:23.365 --> 00:17:24.725
the way your brain works or something.

374
00:17:25.305 --> 00:17:29.085
Um, so, so where I'm at now is I, I I think

375
00:17:29.755 --> 00:17:32.285
that we need to be more direct

376
00:17:32.465 --> 00:17:34.125
and we, I think we're gonna be, need

377
00:17:34.125 --> 00:17:36.765
to be more than just checking a box.

378
00:17:36.875 --> 00:17:39.165
Yeah. And it could happen, or this could happen,

379
00:17:39.265 --> 00:17:40.525
but don't worry, it won't be you.

380
00:17:41.325 --> 00:17:42.925
I think we need to be more direct and,

381
00:17:42.985 --> 00:17:44.325
and actually preparing people

382
00:17:44.385 --> 00:17:47.805
and saying it, this is probably gonna happen to you to,

383
00:17:47.985 --> 00:17:50.205
so one of the, maybe headache is uncommon,

384
00:17:50.385 --> 00:17:53.205
but of all of these things going to happen, one

385
00:17:53.205 --> 00:17:54.605
of them is probably gonna happen to you.

386
00:17:55.265 --> 00:17:58.565
Um, let's talk about, you know, let's, let's start preparing

387
00:17:58.565 --> 00:18:00.725
for that and thinking of what that means and,

388
00:18:00.825 --> 00:18:03.125
and how that makes you feel about this drug and,

389
00:18:03.125 --> 00:18:06.405
and whether you're, you're ready to commit through that.

390
00:18:06.705 --> 00:18:10.485
You know, like for what we mostly would experience as,

391
00:18:10.665 --> 00:18:13.485
as a short term challenge or,

392
00:18:13.485 --> 00:18:14.885
or side effect that we usually get over.

393
00:18:14.905 --> 00:18:17.965
So, you know, I, I'm, I'm now thinking that yes,

394
00:18:18.835 --> 00:18:22.205
this is more than just a, a, a pep rally,

395
00:18:22.325 --> 00:18:23.805
a cheerleading session saying,

396
00:18:23.905 --> 00:18:25.045
everything's gonna be great.

397
00:18:25.535 --> 00:18:27.885
Don't worry about it. I think, I think we have

398
00:18:27.885 --> 00:18:29.165
to have more serious conversations.

399
00:18:29.745 --> 00:18:31.285
At least that's what I'm starting to think now.

Patient Adherence Prediction Models and Expectations

00:00:00 - This was a study, it was sort of part of a broader project that we initiate years ago, right before Covid actually. And, we, we had done a ton of work, using administrative databases to try to, to try to create adherence prediction models. so we were looking at adherence at the, at the, at the population level, most frequently. And we were, you know, we were really good at finding discreet factors that were associated with non-adherence,
00:00:34 - the overall prediction performance was always terrible. So, so, you know, it was, it was not a problem to identify, you know, five or seven variables that were significantly associated. But, but, but, but to actually predict, you know, to put those five or seven variables or, or however many variables we had, which was usually a lot, you know, to, to actually put those together to see, you know, what percentage of the non-adherence
00:01:04 - We could never explain, we always could explain just a fraction of the non-adherence we observed in these population-based databases. So that's what led to this. We said, man, what if we could, you know, give these people a survey and have patient responses to merge with our administrative databases? And, and, and that's where it started. So we, so, so that was the, the overall grand vision of this project.
00:01:34 - and we got partnership with our, with our, provincial government. and then, we just sent out the, we were halfway through sending out the surveys and covid hit, and, and our drug plan is still recovering from covid. And so, so the administrative data side, we still haven't, there's still more to go because we still haven't linked our administrative data to this, but we thought, man, we had, you know, an amazing response. We sent out 11,000 surveys. We got 4,000 back.
00:02:07 - so people just, just responding. We, we, we gave them no, incentive. these are just good citizens responding to our survey. And, I can tell you more about that later. But the, you know, that's where we, that's where we started and there, and we're, I guess we're sort of halfway through and, and these couple papers we've done are, are where we're at now, this current project, I think to me, really, it really makes us look at this idea
00:02:37 - The, the expectations side of our findings was, was most interesting to me. So, so where this sits, I think, is we connect this idea of expectations for side effects, which, if you look at a lot of literature is, is, setting people up for no SIBO effects. You know, the idea, there's a lot of literature that says, geez, we should be careful about telling people what side effects to expect because they're gonna get them. and, and so we call that an no SIBO effects, and,

Impact of Expectations on Drug Adherence

00:03:09 - that creates a pessimistic view of drugs. If you're expecting a side effect, and then you'll get it, you'll, you'll, you'll physically sense a, a, a side effect or the side effect that maybe you wouldn't have if you didn't expect it. That's an no sibo. and, and so we associate these expectations for side effects as a, you know, as, as pessimistic thinking, as, as sort of a, a harbinger of, of, of a poor outcome of somebody that's gonna end up being non likely to be non-adherent
00:03:41 - expectations for these drugs. I think our findings really provide a little bit of a different perspective on what we, you know, currently think about expectations. What we did was, we partnered with our, our provincial insurance drug insurance, who is, you know, I, I live in Canada. We have sort of what we call universal, healthcare, which maybe it isn't technically, but it's close.
00:04:17 - has about 90% of our province are considered beneficiaries of our drug plan. And so we partnered with our provincial drug plan. We said, you guys, we, we had a list. they, they created this all on their own, obviously independent of us. And, they, they received claims in real time for all prescription drugs dispensed from pharmacies in
00:04:48 - And so what they did is they identified all of the individuals who received a first prescription for a cholesterol lowering, blood pressure lowering or glucose lowering drug. So in, in, not exactly real time, but about every six weeks or so, they would run the report of all people that got, you know, one of those three classes of drugs. And we, we, we had a sort of a cohort people of,
00:05:19 - and they had to have at least two years of, of beneficiary status. So we had, we had to have knowledge of 'em for the last two years, and they couldn't have had any prior fill for whatever, you know, for one of those three drug families. So we call them new users. I think for the most part, they probably were. and so if they met that criteria at the insurer, the insurer mailed out our SER survey that we gave to them, they mailed out the survey,
00:05:52 - So, you know, typical survey. so this is a typical cross-sectional survey of drug Users. But the thing that really distinguished our project, I think is the timing. So we really captured them. We probably, you know, captured them somewhere between, you know, two to three weeks and probably four months of when they first got a dispensation for their first drug at a pharmacy. And then the, and then this big tool, we had a, we,

Survey Design and Patient Expectations Study

00:06:27 - designed for people who were new users of one of those three drugs. we designed it, really, we, we had 52 questions. We, we wanted, you know, we real, as you can imagine in this adherence business, we wanted to ask 150 questions, but it was just even at 52, we were quite nervous. So we really condensed a lot of the, you know, the, the traditional batteries of belief scales
00:06:56 - And we, we really had to make hard decisions and, and, and, and condense those questions so we could get a breadth of information on these people. and then ultimately, you know, we said, you know, what are one of the key questions for our outcome? And this study was, you know, what, what about today? You know, have you, are you still taking your drug today? Or have you stopped it? And that was sort of ended up our, our outcome of non persistence were the people who had stopped that or said, you know, as of today filling this out, I, I'm not taking it anymore. We sent out 12,000.
00:07:27 - the provider sent out 12,000 surveys. We got about 4,000 back, which we're very happy with. We, we would've kept going, but we, but covid hit. And so priorities on the, obviously is on the insurance provider, was, you know, we gotta, guys, we, we can't be sending out letters for, for, for this discrete project at, at the time like this. And they had many other things they had to deal with. So, so anyways, we ended up with, yeah, about 4,000 returns. The, a a a couple of just descriptive pieces
00:07:58 - One was, so we had three questions that we, we tried to capture beliefs, again, beliefs, you know, and when, when you look at the, the, the, the validated way to capture beliefs, you, you probably have to have y you know, a numerous battery of questions. We had three, we, when we looked at that, a very, very small percentage of people were, we were categorized as having negative beliefs before starting the drug. And that's how we asked it said, you know, you know, what,
00:08:30 - And, very, very lower than expected percentage of people who had negative beliefs going in. So they were all seemed, most people on average were very positive about the drugs, you know, using the questions we used. the other thing that we found very interesting after the fact, after this idea of, we had one question that said, before you started this medication, did you expect to,
00:09:05 - Or, you know, this, I can't remember the wording exactly, but before you started, did you expect side effects? And only 25% of people expected to get side effects from this drug. You know, these 4,000 people only, only a quarter of 'em actually expected to get drug side effects before starting the drug. So that, that was very, very surprising to me. I, I'm a, I'm a a professor of pharmacy, and sometimes I think that's all we do is tell our students to tell everybody what side effects they're gonna get.

The Impact of Expectations on Drug Adherence

00:09:36 - what side effects they're gonna get. So, so we, we have our, our people here tuned up, and we, we, I thought, or I think that, you know, that, that, especially on the pharmacy side, on the dispensation side, that we, we prime people who say, geez, you know, this, this, there's gonna be side effects. So, so quarter people expect it. and, and just as expected, of those quarter of people that expected to get side effects, they were more likely
00:10:08 - So there's our no SIBO effect replicated in our, in our, you know, real world survey. but then when we started to model all of this, all of these factors, and we had, we had numerous factors that we modeled, a couple interesting things. You know, first off, the drivers of non persistence and that early non persistence, the people that quit already, the main drivers side effects and negative beliefs, bar none.
00:10:41 - And, and, and, when you throw those in the model, they make all of these other sort of, you know, discreet factors that we throw in there, forgetfulness and, and, cost related, non-adherence, all, all this sort of things, all those get blown out. It get diluted by the, the, the, the overpowering effect of side. Did you get side effects? Yes. Or do you have negative beliefs? Yes, those two overpowered everything. and then this expectations
00:11:14 - performed differently in the univariate analysis, in the univariate analysis, an expectation for side effects increased your risk of non-adherence, non persistence. In the multi-variate, there was an interaction we found, when we ran it, it, it, the odds ratio flipped from a 1 3 7 to 0.57. So we, we searched for interaction with everything, and we found this interaction with side effects. So as it turns out, if you,
00:11:47 - obviously your, your risk of non persistence is much higher. But if you expected to get side effects before having a real side effect, your risk of non persistence was, was half. If you, for for, and I'll just back up in case it wasn't clear, sorry, if, if, if you had side effects, but you expected them before, your risk of non-serious persistence was half
00:12:17 - but hadn't expected them prior. Those people, the people that had side effects but hadn't expected them, their non persistence rate was double those who, who, had expected them when they, you know, so, so, so side effects were, were associated with non persistence. Either way, they increased your risk of non persistence no matter what subgroup we looked at. But, but it seemed like the expectations sort of attenuated the risk.

Rethinking Communication About Drug Side Effects

00:12:48 - and so I don't wanna oversimplify this, but, but that was sort of, to me, the most interesting part of, of that, that result of, of side effects. O other than all the other, you know, the other interesting things I just told, this is my intuition, I'll tell you, but it, but it's not founded in psychological psychologic theory. And so I'm such an amateur once we start getting into the theory pieces, I'm, I'm, I'm such an amateur and, and, and right now we're planning next steps. I think there's lots to learn here.
00:13:20 - is I, I told you before, you know that healthcare is, especially in the, in the professional pharmacy, I think we, you know, we're there, there's a duty to warn. I think legally in a lot of, in a lot of locations, there's a duty to warn of what, what are the negative effects of drugs when you're dispensing them? And I think we're, and I think, you know, we're pretty good in the healthcare system of, of warning people about drugs. But, but what I also think could be going on here, and that might explain this,
00:13:52 - to really soften the information about drug safety to people. I, I think healthcare providers typically are so motivated in, in, or, or, or they're, they're so focused on motivating patients to buy in to a, to a therapy. And I think they do this with all good intentions. You know, I, this isn't a pharmaceutical industry conspiracy or trying to sell more drug.
00:14:20 - are really focused on trying to have a positive spin on drugs. And some of my colleagues, you know, and, and some of the, the, the, the, the sort of small scale studies before look at say, well, you know, what, if you frame side effect conversations and you say, well, rather than 5%, you know, could get diarrhea something, why don't we say 95% don't get diarrhea? You know, what if what if we frame our side
00:14:50 - And so this is where I think, this is my hunch. It's totally you, you, you know, totally unverified. But what, what if we say, what if, what if we're too positive to people? So we say, you know, side effects happen. These side effects could happen, but don't worry, this won't happen to you. so I, I wonder if everybody's walking out of there knowing that side effects can happen, but they all think that it's not gonna happen to them because we, you know, we're trying to paint this flowery picture. And, and to some respect, I wonder if we're failing
00:15:24 - occurrence of, of side effects, most of our side effects are indeed nuisance side effects, where most of the things that happen from these run of the mill drugs are, are not severe. but, but they do happen. They happen in, in, in our, in our survey, you know, 25% of people reported experiencing some sort of fi side effect in this window of when this survey came out. So, so that's my hunch. I I think maybe we're too positive. I think maybe, you know, we need to be more, more clear

Preparing Patients for Realistic Drug Expectations

00:15:56 - and we need to prepare people, you know, you know, where's, where is your commitment level? You know, where are you at if that, if is that side effect gonna, gonna crush your whole framework of why you're taking that drug? And all of a sudden the balance is tipped the other direction and I'm done. so, so I, I don't know, I think that's where we're going anyways. And looking at just sort of, you know, just, just this hunch about where your beliefs start, when you get side effects, what they do, what those side effects, how they can change beliefs and,
00:16:32 - without maybe proper preparation. Since this has come out, I, I think differently about this. I used to be that guy. I used to be the guy that said, we don't need to be scaring people about starting a blood pressure drug. Blood pressure drugs are insanely safe, for, for most people. You know, we, we know the risks. I think we can avoid the risks, but I think that, that, you know, in our good, in our well-intentioned approach to, to try
00:17:03 - to start new drugs, right? Scary, you know, it, it, it's, it's difficult for people to get their head around taking a chemical drug every day. So, so I used to be on that camp of we need to reassure people that this is not some major, you know, chemical entity that's gonna change, you know, the way your brain works or something. so, so where I'm at now is I, I I think that we need to be more direct
00:17:34 - to be more than just checking a box. Yeah. And it could happen, or this could happen, but don't worry, it won't be you. I think we need to be more direct and, and actually preparing people and saying it, this is probably gonna happen to you to, so one of the, maybe headache is uncommon, but of all of these things going to happen, one of them is probably gonna happen to you. let's talk about, you know, let's, let's start preparing for that and thinking of what that means and, and how that makes you feel about this drug and,
00:18:29 - You know, like for what we mostly would experience as, as a short term challenge or, or side effect that we usually get over. So, you know, I, I'm, I'm now thinking that yes, this is more than just a, a, a pep rally, a cheerleading session saying, everything's gonna be great. Don't worry about it. I think, I think we have to have more serious conversations. At least that's what I'm starting to think now.

Faculti AI can make mistakes

Research

Policy

Analysis