1
00:00:00,300 --> 00:00:10,020
[Auto-generated transcript. Edits may have been applied for clarity.]
Welcome. Um, this is a guest lecture for a module on the Master's in Translational Health Sciences,

2
00:00:10,020 --> 00:00:15,750
which is run jointly through continuing education and the Department of Primary Care Health Sciences.

3
00:00:16,110 --> 00:00:17,970
My name is Trish Greenhouse.

4
00:00:18,120 --> 00:00:26,550
I'm program director, and I want to extend a really warm welcome to not just to our own master students, most of whom are in the room here,

5
00:00:26,730 --> 00:00:33,480
but also lots of people who've joined online and, uh, also some who are in the room who know that master student.

6
00:00:33,630 --> 00:00:38,010
So I think we've got a hugely diverse group here.

7
00:00:38,310 --> 00:00:42,660
And now I'm going to introduce Carolyn, and I'm so pleased you're here.

8
00:00:43,170 --> 00:00:51,329
Um, you are on visiting fellowship at the this institute in Cambridge, which is a it's run by Mary Dixon Woods.

9
00:00:51,330 --> 00:00:57,510
And they do wacky things with implementation science, pushing back on some of the assumptions.

10
00:00:57,900 --> 00:01:02,700
And I hear from Karen, you have the most amazing interdisciplinary background,

11
00:01:02,700 --> 00:01:10,829
which started with an undergraduate joint harnessing chi busy ology, looking up if you don't know that it's fun and sociology.

12
00:01:10,830 --> 00:01:15,630
And then you went on to do various management t organisationally things.

13
00:01:16,080 --> 00:01:25,680
Um, which equipped you for hanging out with the Institute of Health Policy Management and Evaluation at the fabulous University of Toronto.

14
00:01:26,160 --> 00:01:32,310
Um, but the the reason why I'm so excited is, you know, this Barry is an enablers thing.

15
00:01:32,520 --> 00:01:38,250
What am I say to you guys at the beginning of the week? You give me barriers in the neighbourhoods, you might pass you.

16
00:01:38,460 --> 00:01:42,860
You challenge those barriers and enablers framing. You probably get a distinction.

17
00:01:42,870 --> 00:01:45,960
Well, you're going to tell us how it's done. Thank you.

18
00:01:46,110 --> 00:01:52,650
Thank you. And I understand it's going to be one of those back and forth sessions where there's going to be lots of interactivity.

19
00:01:53,100 --> 00:01:56,489
Uh, questions are going to pop up. I think you are okay with them.

20
00:01:56,490 --> 00:02:01,020
Put your hands in the all right stuff. I'm just going to go and sit in the corner much.

21
00:02:03,080 --> 00:02:06,739
I think. I think I'm on camera now for the online people. Okay, I'll try and stand here.

22
00:02:06,740 --> 00:02:10,370
Novel wandered. Thanks so much and so happy to be here.

23
00:02:10,370 --> 00:02:14,929
Um, really. Thank you. Uh, to Caroline and Trish for the invitation to chat with the module.

24
00:02:14,930 --> 00:02:18,409
I had a chance to read a little bit more about what you guys have been learning all week.

25
00:02:18,410 --> 00:02:25,610
And, uh, what an exciting opportunity, um, for for you guys and for, um, for me to be able to engage with you in what you've been learning this week.

26
00:02:26,240 --> 00:02:32,750
Um, so Esther says I'm from the University of Toronto, in addition to my associate professor role and senior investigating role,

27
00:02:32,750 --> 00:02:37,070
I hold what's called a Canada research chair in implementing digital health innovation.

28
00:02:37,430 --> 00:02:44,450
So in Canada, we have these kind of nationally recognised chairs where we get to present an issue we think really matters,

29
00:02:44,780 --> 00:02:48,439
and the government gets to decide if we get to go and do it or not.

30
00:02:48,440 --> 00:02:54,350
And so I'm one of those and I pitch that implementation matters a lot and digital health matters.

31
00:02:54,950 --> 00:02:56,899
And that's what I've been working on the last five years.

32
00:02:56,900 --> 00:03:01,820
And uh, we'll be renewing that for, for the next five, uh, because it's a journey that continues.

33
00:03:02,450 --> 00:03:05,899
And one of the reasons I really like implementation science, um,

34
00:03:05,900 --> 00:03:10,280
is because it kind of helps us with this problem that you guys have probably talked a lot about,

35
00:03:10,400 --> 00:03:15,620
like that notorious 17 years gap between discovery and actual adoption.

36
00:03:15,920 --> 00:03:20,330
And it's a really sticky one is this paper was out in 2023 from Jama.

37
00:03:20,570 --> 00:03:28,010
But yeah, and I seen this title in a lot of papers over a lot of years, and it just doesn't seem to be budging very much.

38
00:03:28,640 --> 00:03:34,940
Um, and I'm going to talk a little bit about how we might think about using implementation science theory in particular,

39
00:03:34,940 --> 00:03:39,379
and theories that underpin implementation science to maybe help shrink this a

40
00:03:39,380 --> 00:03:43,250
little bit more than they've been able to do over the last couple of decades.

41
00:03:43,880 --> 00:03:47,630
So I'm already going to put you guys to work. I'm gonna put you to work a couple of times.

42
00:03:47,810 --> 00:03:50,960
Maybe you have a piece of paper in front of you. Maybe you have your phone, your laptop.

43
00:03:51,500 --> 00:03:56,360
Um, I'd like us to think a little bit about what we've actually been measuring when we're doing change.

44
00:03:57,080 --> 00:04:01,340
So on your piece of paper or your note on the phone, even better, find a new piece of paper.

45
00:04:01,340 --> 00:04:04,430
I'll show you. Why. Because drawing pictures is a fun part of this work.

46
00:04:04,880 --> 00:04:10,400
I want to think of you guys to think about maybe a reform or an innovation that you've been a part of.

47
00:04:10,850 --> 00:04:16,970
Maybe you've studied it as a researcher, or maybe you've implemented it as an organisational leader.

48
00:04:17,240 --> 00:04:24,680
Maybe you designed it as an innovator. Um, and think about what, um, your experience of doing that was.

49
00:04:24,680 --> 00:04:29,540
So what you're going to do was on the left side of your piece of paper or on one part of your phone,

50
00:04:29,960 --> 00:04:34,580
you're going to write down what the innovation is or what current in 2020 likes to call the thing.

51
00:04:34,700 --> 00:04:38,870
What is your thing, your innovation, the implementation that they were going to put in place?

52
00:04:38,870 --> 00:04:43,880
What's your thing? And on the right side, what did you hope the thing was going to do for you?

53
00:04:43,940 --> 00:04:48,319
What was your outcome that you were trying to get to? So have a thought on that.

54
00:04:48,320 --> 00:04:52,830
And then in the middle, write down all the things you kind of cared about or measure.

55
00:04:52,850 --> 00:04:58,970
So if you're researching it or if you were trying to drive it, what were your KPIs, your key performance indicators?

56
00:04:59,210 --> 00:05:03,890
What were the things you kind of measured between the thing and the outcome?

57
00:05:04,190 --> 00:05:08,210
I'll give you just like a couple of minutes to think about that. But I take a sense of this lovely coffee that was.

58
00:05:09,430 --> 00:05:18,300
Like you to take a look at your list. Um, did anybody write down explicitly something about how you went from the thing to the outcome?

59
00:05:18,780 --> 00:05:23,339
Yeah. And we often do that. Right? What's the thing that we're doing? What were the outcomes that we measured.

60
00:05:23,340 --> 00:05:31,829
And implementation folks will do that as well. But there's all this stuff that's happening in the middle, right, that we know what the thing is.

61
00:05:31,830 --> 00:05:35,700
And then we measure this outcome and then we think, okay, now we we're going to make it work.

62
00:05:35,700 --> 00:05:39,540
And then we try to hand it to people. And then it doesn't really take the way we think it should.

63
00:05:40,170 --> 00:05:43,680
And that's where what you guys have been learning about should be able to help with that.

64
00:05:43,680 --> 00:05:54,509
And in fact, when we go back to that, um, paper from 2023, it suggests to us that the burgeoning field of implementation science written in 2023,

65
00:05:54,510 --> 00:05:58,980
I feel like it was a couple, at least a couple of decades old at that point, but that should be helping us out.

66
00:05:59,900 --> 00:06:05,460
But not all. Implementation science is kind of the same and may or may not be helping us on that journey.

67
00:06:05,480 --> 00:06:08,870
So we'll talk about that a bit today. So hold on to your piece of paper.

68
00:06:08,870 --> 00:06:12,680
Think about your thing and your outcome, and let's start working on filling in the middle of it.

69
00:06:13,610 --> 00:06:21,020
Um, but what often happens for us as we really notice that there's determinants in the literature, we really like determinants.

70
00:06:21,230 --> 00:06:29,660
We did a really interesting, um, one study led by, uh, Genevieve Rouleau and uh, in published just in 2024,

71
00:06:29,660 --> 00:06:33,980
where we looked at just studies, uh, digitally implementation of digital health.

72
00:06:34,280 --> 00:06:39,590
And we wanted to look at the theories, models and frameworks that were being used in those studies over a 20 year period.

73
00:06:40,100 --> 00:06:43,820
Uh, we looked found 156 studies that were published in that time.

74
00:06:44,150 --> 00:06:48,260
68 different theories, models and frameworks were used.

75
00:06:48,830 --> 00:06:54,469
The vast majority of determinants, frameworks, and really surprisingly,

76
00:06:54,470 --> 00:07:02,330
about 20% of those determinants frameworks were just cipher a lots of consolidated framework for implementation research that's getting used.

77
00:07:02,690 --> 00:07:11,510
We can talk a lot about why that's the case. But overall the field in digital health and others really love determinant studies.

78
00:07:11,960 --> 00:07:16,580
We love it. Last year uh for a talk I want you know,

79
00:07:16,580 --> 00:07:24,680
for for fun I wanted to see in just 2025 how many studies of determinants in digital health were out there and there were over 100.

80
00:07:25,130 --> 00:07:28,430
This is not a systematic review. This is just me playing around with Google Scholar.

81
00:07:28,700 --> 00:07:34,210
But that's an awful lot of papers on just determinants. So as a field we just really love digital.

82
00:07:35,330 --> 00:07:41,840
Um, and I really encourage to hear that you guys were talk to you about this at the beginning, that maybe we need to move beyond,

83
00:07:42,260 --> 00:07:49,070
um, because there's a concern growing amongst implementation scientists and other people that as the science has matured,

84
00:07:49,160 --> 00:07:58,309
uh, it's arguably becoming structurally comfortable that we're really good at cataloguing determinants and strategies,

85
00:07:58,310 --> 00:08:06,680
but we're really under theorising mechanisms, which is really the key thing that happens between the thing and the outcome.

86
00:08:07,280 --> 00:08:17,239
Um, and as I wrote in 2023 with, uh, their colleagues that theoretical products at describing what have generally outpaced explanations of how in

87
00:08:17,240 --> 00:08:24,470
implementation science and as evidenced by those numerous determinant frameworks but fewer explanatory theories.

88
00:08:25,010 --> 00:08:28,399
So we saw that in our scoping review and others who are also working in the field

89
00:08:28,400 --> 00:08:31,700
doing similar scoping reviews in other like outside of just digital health,

90
00:08:32,120 --> 00:08:38,029
are overwhelmingly seeing this to be the case. So it's really important to remember that determinants matter.

91
00:08:38,030 --> 00:08:43,400
They're important for for some key things, but they're not really explanations all the time.

92
00:08:43,940 --> 00:08:47,780
Strategies that we also publish aren't necessarily causal processes.

93
00:08:48,110 --> 00:08:51,440
And outcomes don't necessarily mean we've understood what happened.

94
00:08:51,710 --> 00:08:55,190
Right. We see this in trials all the time. We we saw an outcome and that's awesome.

95
00:08:55,190 --> 00:09:01,850
And then we try to pick the thing up and put it to another place. And we can't for the life of us figure out why it didn't happen in the other place.

96
00:09:02,450 --> 00:09:06,230
And that's all the middle part between our thing and our outcome.

97
00:09:06,890 --> 00:09:12,080
So mechanistic thinking or mechanisms are a way forward in this.

98
00:09:12,080 --> 00:09:19,100
And that's and this is not a brand new space. A lot of implementation scientists and other theorists have been grappling with mechanisms.

99
00:09:19,100 --> 00:09:22,580
Realist thinking has been grappling with mechanisms for a long time.

100
00:09:23,010 --> 00:09:27,589
Um, so I'm going to spend today talking about how we can think differently about mechanisms, um,

101
00:09:27,590 --> 00:09:32,450
the different theoretical foundations that help us uncover where they are and some of the challenges

102
00:09:32,450 --> 00:09:37,550
we might face if we're really going in and trying to understand this component of implementation.

103
00:09:39,140 --> 00:09:47,360
So there are lots of different definitions and understandings, and generally not an one agreed upon, um, definition that is being used.

104
00:09:47,750 --> 00:09:51,350
Our realist friends will suggest mechanisms are kind of this, uh,

105
00:09:51,350 --> 00:09:57,259
resources and reasoning components that trigger that are triggered within a particular context to lead to an outcome.

106
00:09:57,260 --> 00:10:01,250
Those are the context mechanism, outcomes groupings that you tend to see.

107
00:10:01,760 --> 00:10:07,999
Um, del can explain expanded that to suggest mechanisms exist in resources and reasoning, um, as well.

108
00:10:08,000 --> 00:10:15,110
So we can kind of think about them in that way. Uh, Lewis Carroll, Lewis and a group of colleagues out of the states have been working in,

109
00:10:15,110 --> 00:10:20,840
in mechanisms for quite some time, and they've been trying to catalogue them to the best of their ability for a long time.

110
00:10:21,080 --> 00:10:26,750
And they define them as processes through which implementation strategies exert effects.

111
00:10:27,290 --> 00:10:34,460
Um, so very much focussed on the implementation component, how you connect your strategy for doing the thing to lead to the outcome you're expecting.

112
00:10:34,940 --> 00:10:40,099
And then, uh, more on the theoretic kind of theoretical, sociological, middle range theory spaces.

113
00:10:40,100 --> 00:10:45,950
We have folks like Romain and Tracy Finch doing normalisation process theory thinking, where they suggest mechanisms,

114
00:10:45,950 --> 00:10:53,569
or these generative social processes, uh, that explain how a practice becomes normalised, at least in the the example of that particular theory.

115
00:10:53,570 --> 00:10:57,260
But there are other middle range theories that effectively say that same thing.

116
00:10:57,260 --> 00:11:03,680
But there's a causal thing that's happening under the surface that's leading from the thing to the outcome or the action.

117
00:11:04,280 --> 00:11:12,200
But generally, they're all agreeing that mechanisms are about explaining how and why a change happens.

118
00:11:12,890 --> 00:11:18,770
Um, the how and why being critical determines sort of telling you what, not necessarily how or why they're functioning.

119
00:11:20,110 --> 00:11:25,350
So another way I like to share this idea with people is using a light switch analogy.

120
00:11:25,360 --> 00:11:31,420
I used to use a trail running analogy, which only worked for a very small pocket of people who knew what trail running.

121
00:11:32,380 --> 00:11:35,590
So I was strongly encouraged to not use that analogy.

122
00:11:35,840 --> 00:11:39,150
So now I'm on a light switch and a light bulb, which makes a lot more sense.

123
00:11:39,160 --> 00:11:43,840
So if you standing in this lovely room and we would like to light up the room a little bit.

124
00:11:43,840 --> 00:11:48,760
So I see some lights here and I go over to the wall and I turn the light switch on and off the light went on.

125
00:11:49,030 --> 00:11:57,040
Magic. However, for that light to turn on because that light switch was switched, there's all this stuff happening in the walls of the building.

126
00:11:57,400 --> 00:12:02,590
There's all the stuff happening in the environment around us to create power that goes to that light switch,

127
00:12:02,890 --> 00:12:08,770
to make sure that that light went on, because that light switch was turned on and not a fan was turned on or not.

128
00:12:08,770 --> 00:12:16,870
A light in another room was turned on. And the key with mechanisms is they tend to exist in places we don't always obviously see them.

129
00:12:17,440 --> 00:12:20,620
We're kind of in the walls. They're kind of in our minds.

130
00:12:20,620 --> 00:12:24,549
They're kind of in our social norms that exist in our environments around us.

131
00:12:24,550 --> 00:12:31,240
So they're not readily available for measurement, and they're not readily available for us to see or understand them,

132
00:12:31,240 --> 00:12:37,930
which is part of why we don't go in that direction when we go to evaluate or measure or understand what we see,

133
00:12:38,050 --> 00:12:45,790
because it's not necessarily something we can't see. So they matter because we might have asked them this time.

134
00:12:45,790 --> 00:12:51,910
It's a large ocean liner, and here's our thing chugging along and we're like, this is definitely going to work.

135
00:12:52,060 --> 00:13:00,310
But then what happens is we face a resistance iceberg. And often at the top of that iceberg are all those determinants that we can see and measure.

136
00:13:00,880 --> 00:13:06,040
But what the mechanisms that are important are the ones that actually underpin the things that we see and measure.

137
00:13:06,040 --> 00:13:11,409
They don't exist differently from the mechanisms. They're actually kind of part of why or differently from the determinants.

138
00:13:11,410 --> 00:13:14,860
They actually are part of why the determinants are the determinants in the first place.

139
00:13:15,310 --> 00:13:23,590
And they rest on they're influenced by these mechanisms that are things like affect and beliefs and norms and values that lead to things like,

140
00:13:23,920 --> 00:13:30,010
um, general resistance or workflow problems or resource allocation issues.

141
00:13:30,310 --> 00:13:35,770
There are underlying pieces about why the things we see at the surface are actually happening at all.

142
00:13:36,100 --> 00:13:43,060
So we need a way to kind of dive deeper into what the determinants are in order to understand why they matter to us in the first place.

143
00:13:43,360 --> 00:13:46,240
And that's where our action probably needs to happen.

144
00:13:46,870 --> 00:13:52,509
Um, that's why when you go and put something in practice and you fix all your work flow in one place, you're in a great fixed work flow.

145
00:13:52,510 --> 00:13:56,110
And then I go do it over here. And that work flow fix just isn't working.

146
00:13:56,410 --> 00:13:59,950
It's because something else is happening underneath the surface that you need to uncover.

147
00:14:01,430 --> 00:14:08,090
Um, now, there's also some new exciting studies that have shot that has shown, particularly this one from Williams and company,

148
00:14:08,360 --> 00:14:13,160
that mechanisms can have actual like causal explanations for implementation success.

149
00:14:13,430 --> 00:14:18,290
This one came out just a month ago, which I'm super psyched about is in all my talks over the next few months.

150
00:14:18,710 --> 00:14:24,560
Um, and certainly Carl May and Tracy Vance. You design normalisation process theory are also super pumped about this.

151
00:14:24,890 --> 00:14:32,530
Uh, because they ran, uh, they did a secondary analysis of data out of a hybrid type three trial of it implementing a digital, um, tool.

152
00:14:32,540 --> 00:14:39,800
This was a on clinician decision support tool or. Sorry, this was a, uh, tool used in mental health and outpatient clinics for youth.

153
00:14:40,250 --> 00:14:48,469
Uh, they were able to use the Nomad's tool, which measures mechanisms of coherence and collective action and cognitive participation,

154
00:14:48,470 --> 00:14:51,170
all the mechanisms underlying normalisation process theory.

155
00:14:51,560 --> 00:14:56,719
And they were able to see, like causal reasons for implementation working when those mechanisms were turned on,

156
00:14:56,720 --> 00:15:01,320
when that like dependent outcome occurred. Um, which was very exciting.

157
00:15:01,340 --> 00:15:07,400
Um, and then they saw that there was a statement of the implementation over a six month period when those mechanisms were turned on.

158
00:15:07,700 --> 00:15:14,600
So not only important when you're putting anything in place, but maybe also important to sustain the new thing over time.

159
00:15:14,610 --> 00:15:20,329
So really encouraging data coming out. So that's one way of doing finding out what your mechanisms are.

160
00:15:20,330 --> 00:15:27,950
You can go out and actually study them and figure out causal pathways. But uh, has done kind of suggests there are lots of other ways we can find, um,

161
00:15:27,950 --> 00:15:34,280
mechanisms in the world that we can look for things like strong association, which happened with the Williams paper.

162
00:15:34,490 --> 00:15:38,059
You can dive in and do some specificity work or consistency work,

163
00:15:38,060 --> 00:15:44,330
which is actually getting into being able to quantify what a mechanism is and tested empirically over time.

164
00:15:44,870 --> 00:15:53,060
Um, you can do experimental implementations of experimental manipulation or look at things occurring over time and look for patterns over time.

165
00:15:53,630 --> 00:16:00,020
Um, all of these things are quite complex and require a good amount of measurement so that you can one understand the mechanism is there,

166
00:16:00,020 --> 00:16:02,630
and to be able to measure it consistently over time.

167
00:16:03,200 --> 00:16:12,799
And because those things are hard to do when you're trying to measure something like social norms often land on the plausibility or of or coherence,

168
00:16:12,800 --> 00:16:15,530
um, requirement of identifying mechanisms.

169
00:16:15,920 --> 00:16:24,440
So Carol Lewis in that group, um, back in 2020, had done a review of the literature to see how mechanisms were being unveiled in the literature.

170
00:16:24,620 --> 00:16:32,690
And overwhelmingly plausibility. Coherence was the main way that we were uncovering it, because we were still really in a qualitative space,

171
00:16:32,990 --> 00:16:37,400
working in smaller environments and not being able to do these bigger replication pieces of work.

172
00:16:37,760 --> 00:16:42,200
So the Williams paper is very encouraging and shows us ways that we can embed mechanism,

173
00:16:42,200 --> 00:16:45,620
mechanistic measurement into bigger trials with bigger numbers,

174
00:16:45,620 --> 00:16:51,139
and start getting into those higher order measurement opportunities that cast and suggests for

175
00:16:51,140 --> 00:16:57,470
us what I will say is what I'm going to present next as we go through some other mechanisms.

176
00:16:57,830 --> 00:17:05,300
Um, tends to rest on plausibility or coherence, um, as a way to understand that these are mechanisms available to us,

177
00:17:05,570 --> 00:17:11,960
because that's kind of a lot of the data that we have available to us. So I think it's still conversational and still things that need to be tested.

178
00:17:12,380 --> 00:17:17,630
Um, I'm going to talk about why theory helps us when in the plausibility or coherence space.

179
00:17:18,920 --> 00:17:21,829
Another way. We've uncovered this in some of the work that we've done,

180
00:17:21,830 --> 00:17:28,309
and that's one of these really fun piece of work I'm doing with colleague, uh, Jay Shaw in, uh, University of Toronto.

181
00:17:28,310 --> 00:17:31,490
And I think a number of people here all know Jay was like, hey, do you know Jay?

182
00:17:31,550 --> 00:17:40,370
Yeah. Uh, so we're trying to uncover, um, how to change practice for community health workers, um, and understand what their needs are.

183
00:17:40,370 --> 00:17:48,229
And we're doing a bunch of co-design work in terms of understanding their roles, uh, in community health work in different environments across Canada.

184
00:17:48,230 --> 00:17:55,070
So we have sites in Ontario where I'm from. We have sites in B.C., um, as well, where Nancy Clark is leading that part of the work.

185
00:17:55,190 --> 00:18:01,280
And so we were trying to understand what they were facing in terms of the challenges of doing their work as community health workers.

186
00:18:01,760 --> 00:18:07,370
And in order for us to dive deeper into understanding root causes or mechanisms leading

187
00:18:07,370 --> 00:18:11,570
to those barriers and enablers that they were experiencing as community health workers,

188
00:18:11,900 --> 00:18:18,710
we just kept asking why? So this is an image from one of our mural boards that we used in our co-production work.

189
00:18:19,130 --> 00:18:22,470
So we'd say, okay, what's the challenge? You're experiencing a mental health.

190
00:18:22,490 --> 00:18:26,120
You say, oh, why do you think that is? And explains this? Okay, well why do you think that is?

191
00:18:26,120 --> 00:18:27,200
And then they explain to us again.

192
00:18:27,500 --> 00:18:34,250
And so they only see two layers of this, but it just keeps going and going and going until you get down to I can't say why anymore.

193
00:18:34,670 --> 00:18:40,850
And that's where you kind of get to like values and norms and beliefs when there isn't really another why.

194
00:18:40,850 --> 00:18:45,410
It's just what's deeply rooted in people, in systems, in our environments.

195
00:18:45,890 --> 00:18:48,590
Um, I hear you heard lurkers love a complexity theory.

196
00:18:48,590 --> 00:18:55,340
So think about all those factors and complexity theory, then into those root reasons why things are happening or not happening.

197
00:18:55,610 --> 00:18:59,240
And that's when some of that, like magical mechanistic stuff starts happening.

198
00:18:59,660 --> 00:19:06,020
Um, and that's where you can start uncovering it and understanding it, at least from a coherence and plausibility, um, perspective.

199
00:19:07,760 --> 00:19:11,980
I'm going to make you go to work again. Go back to your pieces of paper or your phones.

200
00:19:12,310 --> 00:19:15,580
You have your thing on the left. You have your outcome on the right.

201
00:19:16,090 --> 00:19:18,250
We've talked a little bit about what the middle could be.

202
00:19:18,850 --> 00:19:24,770
If you wrote some things down in the middle already, stuff you measured, can you think about that a little deeper?

203
00:19:24,790 --> 00:19:28,989
If you saw something on the surface that occurred that maybe got in the way or helped what

204
00:19:28,990 --> 00:19:35,080
might be underlying that thing that got in the way or helped your thing to its outcome?

205
00:19:35,320 --> 00:19:37,060
And I got to a something with you.

206
00:19:37,450 --> 00:19:44,680
Um, I'm curious if anyone wants to share with us what you were thinking about, or possibly how hard that might have been.

207
00:19:45,430 --> 00:19:51,970
Um, the risk one is a really interesting one, right? That, um, if you presented a new idea and it feels less risky,

208
00:19:52,300 --> 00:20:00,010
then you're kind of driving down into individual kind of behaviour mechanism that our tolerance for risk really shapes how we behave or not behave.

209
00:20:00,400 --> 00:20:06,850
Um, excellent example of how you can think about that. Moving into mechanistic where we're in can are great questions to ask.

210
00:20:06,850 --> 00:20:11,500
Like where is the mechanism. I'm going to present some stuff that's going to make that answer really hard to figure out.

211
00:20:11,950 --> 00:20:19,950
Um, but uh, I think the piece you, you clicked in on the midwife, introducing something gets you there, right?

212
00:20:19,960 --> 00:20:24,580
The interest. So where you would start is the midwife introduce something that's like the action we did.

213
00:20:25,030 --> 00:20:30,970
And then you got to ask, well, why does that work? Why is that helping a person actually use a technology?

214
00:20:31,240 --> 00:20:36,340
And there's lots of really interesting literature coming out around how we can transfer

215
00:20:36,340 --> 00:20:42,610
trust from an individual into trust to a technology that if someone presents one to us,

216
00:20:42,610 --> 00:20:47,769
that if we have a trusting, good relationship with them, that's transferred into the technology space.

217
00:20:47,770 --> 00:20:51,219
And it also comes from, um, trust transfer in teams.

218
00:20:51,220 --> 00:20:56,680
So that's in the team's literature for, um, like a long time where people can work within a team,

219
00:20:56,680 --> 00:21:01,870
even if it's a new person, because they trust the team and you could potentially transfer them to technology.

220
00:21:01,870 --> 00:21:05,290
Some newer papers coming out on that. Very cool.

221
00:21:05,290 --> 00:21:10,200
So as you start getting into the next level of why that will help you.

222
00:21:10,270 --> 00:21:14,560
I'm going to show you some things about, um, some theory that can help you.

223
00:21:15,400 --> 00:21:25,090
So I did a little bit of a deep dive with our example here about how we can try to use theory to inform understanding why a mechanism works or not.

224
00:21:25,870 --> 00:21:30,279
And, you know, Tristan, a nice job of talking about my very multi-disciplinary background.

225
00:21:30,280 --> 00:21:31,870
There's political science in there, too.

226
00:21:32,230 --> 00:21:38,830
So I had, you know, 4 or 5 different disciplinary backgrounds, and I think that's probably why I like implementation science,

227
00:21:39,280 --> 00:21:43,390
because implementation science is like the science for people who don't want to pick things,

228
00:21:43,990 --> 00:21:47,530
because it requires understanding cognitive and social psychology.

229
00:21:47,710 --> 00:21:52,190
We need to understand why people do things and don't do things. You need to understand sociology.

230
00:21:52,190 --> 00:21:55,569
You need to understand why societies shift or move, or in anthropology,

231
00:21:55,570 --> 00:21:59,650
you have to understand why cultures shift or move, and how culture sharing groups behave.

232
00:22:00,040 --> 00:22:07,750
And all of these disciplinary traditions have been around for a long time, have been doing the work of understanding change.

233
00:22:07,960 --> 00:22:12,580
Right. We don't need to reinvent a lot of that. And some of this is like deeply, empirically based.

234
00:22:12,940 --> 00:22:16,270
Um, they show it in labs, they show it consistently in real world.

235
00:22:16,600 --> 00:22:23,440
Why people, teams, organisations, policies, systems and cultures change over time.

236
00:22:24,130 --> 00:22:27,910
We don't have to reinvent this. There are excellent theories we can go to.

237
00:22:28,270 --> 00:22:33,820
And it's those why questions that these theories have been grappling with that help us understand mechanisms.

238
00:22:34,210 --> 00:22:38,950
Those are the causal links that we can go to and actually test and look for in our implementation work.

239
00:22:39,640 --> 00:22:48,100
So I'm a big fan of going back to disciplinary traditions. I teach a graduate course kind of on this around how you use this to pick the right theory,

240
00:22:48,100 --> 00:22:54,760
model or framework for a particular challenge by recognising what is your big lever of change in the implementation you're doing?

241
00:22:55,000 --> 00:23:02,740
Sometimes we know, sometimes we don't. Um, in the example of kind of pharmacists taking on new, um, uh, new prescribing behaviours.

242
00:23:03,100 --> 00:23:06,180
There's like a couple of big levers we can think about, right?

243
00:23:06,190 --> 00:23:09,009
There's political levers, which you noted, uh,

244
00:23:09,010 --> 00:23:15,160
being able to share that role is another big lever that being clinicians reading power to another clinician.

245
00:23:15,400 --> 00:23:21,010
That's a big lever in each of those levers kind of hook into these different kinds of traditions of understanding change.

246
00:23:21,490 --> 00:23:26,319
And that's how you would go ahead and pick a theory to help you understand the challenge change that's

247
00:23:26,320 --> 00:23:30,670
happening and something you can actually test and use and maybe appropriately rebuilding over time.

248
00:23:31,480 --> 00:23:33,550
So I'm going to go through a few of them and we'll go through all of them.

249
00:23:33,760 --> 00:23:39,520
I'll go through a few that we've been able to look at and get some coherence level understanding of mechanisms and,

250
00:23:39,520 --> 00:23:45,639
and some of the work that we've done. And hopefully this sparked some, some ideas in the example you've been working on, um,

251
00:23:45,640 --> 00:23:50,950
for mechanisms that might be underlying the stuff that you're seeing or in any projects that you have coming up.

252
00:23:52,230 --> 00:23:56,970
So let's start with cognitive mechanisms. It's often when we think about brain because decisions are made by humans.

253
00:23:57,810 --> 00:24:04,620
Um, and these can be things like habit formation, mental models, identity from an individual standpoint.

254
00:24:04,620 --> 00:24:13,590
We'll talk about identity from a global standpoint. Um, and there are different theoretical approaches that try to understand cognitive behaviour,

255
00:24:13,800 --> 00:24:19,800
things that we use in implementation science or things like they can be or they capability,

256
00:24:19,800 --> 00:24:26,459
opportunity, motivation, behaviour and related theory or plan behaviour and um behaviour change.

257
00:24:26,460 --> 00:24:28,410
We'll a lot of those behaviour change theories.

258
00:24:28,680 --> 00:24:35,700
And if you go into their like foundational papers and how they were written first they'll tell you the assumptions that they're built on.

259
00:24:35,700 --> 00:24:38,730
They'll tell you the causal mechanisms that they're resting on.

260
00:24:39,000 --> 00:24:45,510
And that really helps you understand why they believe these things are important to a shift in behaviour over time.

261
00:24:45,870 --> 00:24:49,710
I often encourage people try to find, like the first paper that talked about the theory,

262
00:24:49,980 --> 00:24:54,330
because they don't always carry the assumptions of the theory forward as they continue to write about it.

263
00:24:54,330 --> 00:24:59,280
And this becomes really important for us to understand the assumptions that are underpinning our thinking.

264
00:25:00,120 --> 00:25:03,120
Um, so one example I am very happy to share.

265
00:25:04,080 --> 00:25:09,270
And it's a little this one's a little sad. Um, like I actually passed just two months ago very tragically.

266
00:25:09,600 --> 00:25:18,450
Uh, and this is one of the projects we got to do together. Micah was a brilliant clinician, um, family doc, and worked in the ICU at our hospital.

267
00:25:18,810 --> 00:25:26,850
And he'd been working in rural, remote parts of Ontario for a long time, and he really believed in GLP ones for diabetic patients.

268
00:25:26,850 --> 00:25:32,309
He'd seen so much evidence for their effectiveness. He saw that they were helping his his patients.

269
00:25:32,310 --> 00:25:37,590
He was deep into the literature, but his colleagues weren't using it and they weren't prescribing it.

270
00:25:37,800 --> 00:25:42,120
Certainly in the North, where diabetes was a significant problem for the patient population of their,

271
00:25:42,120 --> 00:25:45,300
particularly our First Nations and indigenous populations.

272
00:25:45,870 --> 00:25:50,810
And so he came to me, he's like, you know, I know you do a lot of co-design work. Let's co-design a new training platform.

273
00:25:51,210 --> 00:25:54,450
They do this nice digital training platform and that's going to solve this problem.

274
00:25:54,990 --> 00:26:02,070
And I said, are you sure it's a training problem? Like is it that they don't know what it is and they don't know how to prescribe it?

275
00:26:02,670 --> 00:26:05,880
And then he kind of thought about that and said, no, well, maybe, maybe it's not.

276
00:26:05,970 --> 00:26:09,060
So we took a step back and used some funding.

277
00:26:09,060 --> 00:26:14,610
He got through one of his many career awards that he was getting at the time to do a deep dive.

278
00:26:14,610 --> 00:26:19,890
It was qualitative work and did some interviews with the clinicians that in his environment,

279
00:26:20,190 --> 00:26:25,770
and we applied calmly to it to understand what was shifting or not shifting their prescribing behaviour.

280
00:26:26,280 --> 00:26:30,270
And some really interesting things came out of that. Things like their clinical identity,

281
00:26:30,360 --> 00:26:36,930
who they understood themselves to be and what their job was as a physician was changing whether they were prescribing things or not.

282
00:26:37,560 --> 00:26:46,379
Um, their beliefs in GLP one and the social construction around Ozempic as a concept was shifting their behaviour,

283
00:26:46,380 --> 00:26:52,170
whether there was clinical evidence or not, that social influence was really important and meaningful to them,

284
00:26:52,620 --> 00:26:56,879
their past experiences and their risk assessments, all of those things mattered a lot.

285
00:26:56,880 --> 00:27:01,050
And if we'd gone ahead and just built a new technology enabled training platform, sure,

286
00:27:01,050 --> 00:27:05,640
it had accessibility to them, but it wouldn't have hit on these really key things.

287
00:27:06,180 --> 00:27:10,440
And what that meant was different kinds of relationship building, different kinds of sense making.

288
00:27:10,830 --> 00:27:15,900
And although we won't get to do it with Mike, we will try and carry this forward to, to, um,

289
00:27:16,260 --> 00:27:20,310
to do some interventions with those clinicians so we can kind of see his vision forward.

290
00:27:22,620 --> 00:27:27,060
So those are some cognitive mechanisms. There are lots though. You can think about lots of behaviour change stuff.

291
00:27:27,390 --> 00:27:30,200
Let's talk a little bit about sociological mechanisms.

292
00:27:30,210 --> 00:27:38,670
These are a little trickier because they exist both within individuals and within our social systems I think for complexity conversation.

293
00:27:38,670 --> 00:27:44,579
Right. All of these things are like a big blanket of interwoven fibres where our individual

294
00:27:44,580 --> 00:27:50,250
behaviour is shifted and shaped and moved by our social systems and fabrics.

295
00:27:50,640 --> 00:27:55,690
Right. That that individual clinician isn't necessarily just working as an individual clinician in a vacuum.

296
00:27:55,710 --> 00:28:05,940
There's all these forces acting on us. Things like norms, things like power networks and then identity from a shared sense,

297
00:28:06,360 --> 00:28:10,940
things about not just who am I as a person, but who am I as a physician?

298
00:28:10,950 --> 00:28:15,030
Who am I as a parent? Who am I as a child?

299
00:28:15,270 --> 00:28:20,500
All of these kind of simple constructions within shared groups that shape our behaviour as well.

300
00:28:20,520 --> 00:28:24,030
So identity is a sticky one that lands in a couple of places.

301
00:28:24,660 --> 00:28:28,380
And then the theoretical approaches we want at this level are a little different.

302
00:28:28,500 --> 00:28:34,800
They're not necessarily about individual cognition. They're about like how the individual interacts with their social environment.

303
00:28:35,130 --> 00:28:38,640
So then you get into like some classic social interactions in theory.

304
00:28:38,820 --> 00:28:42,830
I'm always a big fan of Goffman and he's like front stage backstage kind of stuff.

305
00:28:42,840 --> 00:28:48,390
I think that really resonates well. You can get into critical theory and think about power and, um,

306
00:28:48,570 --> 00:28:53,820
kind of structures of power environments and how that shifts our ability to move or not.

307
00:28:54,300 --> 00:28:59,820
Um, and then socialisation processes, if you're a fan of Foucault, this one is always really fun.

308
00:28:59,820 --> 00:29:07,230
And if you can think a lot about Inter in terms of professional identity creation, that as people go through the training of their profession,

309
00:29:07,500 --> 00:29:14,459
they're socialised into that identity and that shifts their behaviour and what they think is the right or wrong way of doing things,

310
00:29:14,460 --> 00:29:24,480
and it also shifts the environments around them. So mechanisms that might emerge from this level might be things like identity and normative threat.

311
00:29:24,780 --> 00:29:28,020
So what is threatening my power as a clinician?

312
00:29:28,380 --> 00:29:32,670
What might be threatening my role as a, um, a caregiver?

313
00:29:33,480 --> 00:29:42,360
Um, there might also be legitimacy construction. So what is a legitimate or illegitimate form of action within what is understood within my society?

314
00:29:42,960 --> 00:29:46,350
Um, and there's also things like network contagion that always sounds kind of fun,

315
00:29:46,350 --> 00:29:53,310
but it's basically that when something happens in one part of a network, it can trickle and move towards other parts of a close knit network.

316
00:29:54,030 --> 00:29:59,460
Um, we've looked at this a little bit in the shared identity space of professional identity.

317
00:29:59,850 --> 00:30:03,030
Um, and this is work by my lovely colleague, uh, Jennifer Schuldiner.

318
00:30:03,210 --> 00:30:06,870
Um, in 2021. And then we did some work in 2024. Um,

319
00:30:06,870 --> 00:30:16,049
all the years leading into it that found in virtual care environments that things like professional identity were really important to use or not use,

320
00:30:16,050 --> 00:30:22,540
and for virtual care. So doctors, nurses work, looked at emergency virtual emergency departments.

321
00:30:22,560 --> 00:30:28,380
Uh, during Covid they were setting these up all over Canada, definitely in Toronto and in her hospital in Toronto.

322
00:30:28,890 --> 00:30:32,250
And they were finding some clinicians loved it and some clinicians hated it.

323
00:30:32,250 --> 00:30:33,980
And they were trying to find out why.

324
00:30:33,990 --> 00:30:40,770
And they found that the clinicians who really hated it said things like, this isn't what it means to be an emergency room doctor.

325
00:30:41,640 --> 00:30:46,170
And then we did some work and social representation and virtual care and primary care.

326
00:30:46,170 --> 00:30:51,090
Did some deep ethnographic work, talk to patients, talk to their caregivers, talked to providers.

327
00:30:51,090 --> 00:30:57,120
After watching an interaction in a virtual environment. And lots of clinicians used it and lots of clinicians didn't.

328
00:30:57,120 --> 00:31:00,930
And the ones who didn't use it said, this is not what it means to be a family doctor.

329
00:31:01,410 --> 00:31:05,520
This is not how I was trained. This isn't how I understand what my role is.

330
00:31:05,940 --> 00:31:12,780
And so virtual care was a very immediate and real threat to what it meant to them to be a good doctor,

331
00:31:13,140 --> 00:31:18,120
which for them was being in a room with another person. But that wasn't everybody.

332
00:31:18,150 --> 00:31:23,670
There were lots of clinicians in the study for whom it resonated strongly with their identity as a clinician,

333
00:31:24,150 --> 00:31:30,060
and it meant that they could do a better job of connecting, particularly some of the social workers and um,

334
00:31:30,840 --> 00:31:36,450
um, nursing staff who were doing much more kind of mental health and social support pieces,

335
00:31:36,840 --> 00:31:42,420
and probably because the patients were more vulnerable in a virtual environment than they could be in person.

336
00:31:42,660 --> 00:31:48,600
So that allowed them to tap into their personal identity of being a good social worker that could

337
00:31:48,600 --> 00:31:53,940
connect with somebody emotionally because there was now availability to connect to them emotionally.

338
00:31:54,240 --> 00:31:59,880
And a lot of that was wrapped in the construction of how they were trained and what their expectations were over time.

339
00:32:01,450 --> 00:32:07,430
So those are very different kinds of mechanisms we think about. I'll talk about how that changes strategy in just a moment.

340
00:32:08,870 --> 00:32:13,240
There's also organisational and team mechanisms so important when we think about these ones.

341
00:32:13,250 --> 00:32:20,900
Yes, the individuals still acting, but individuals can also act on behalf of their team or on behalf of their organisation.

342
00:32:21,290 --> 00:32:29,770
So for example, when the leaders or directors are acting, they're not necessarily making decisions because of what they, as an individual want to.

343
00:32:29,780 --> 00:32:33,469
They're making it a kind of, um, for their organisation,

344
00:32:33,470 --> 00:32:37,370
and they're kind of embedding what their organisation wants to do and what their priorities are,

345
00:32:37,370 --> 00:32:41,719
and all of those things that they can, you know, constructed in. Um, so that's how they're acting.

346
00:32:41,720 --> 00:32:48,590
So we need a different kind of theoretical lens to understand that kind of change that's happening, things we can look at.

347
00:32:48,590 --> 00:32:55,640
Now back to normalisation process theory, which talks of teams changing behaviour over time and and broadly team science,

348
00:32:55,820 --> 00:33:03,620
lots of organisational behaviour theory like ecology and institutionalisation, lots of things that talk about why an organisation changes or not.

349
00:33:03,650 --> 00:33:08,090
Tons of little literature there and then you can get into some shared mental model stuff.

350
00:33:08,090 --> 00:33:11,000
So mental models can like how an individual understands the world.

351
00:33:11,450 --> 00:33:18,170
Um, when we share that amongst a group, that becomes something within teams or organisations, um, that you can work with,

352
00:33:18,830 --> 00:33:25,159
mechanisms we might see here are things like collective sense making, role renegotiation, trust,

353
00:33:25,160 --> 00:33:30,050
collaboration, things that are kind of happening at a team interactional, larger group level.

354
00:33:31,130 --> 00:33:35,600
Um, I actually ended up adding this slide this morning because I forgot the study was that, um,

355
00:33:35,990 --> 00:33:41,809
and this was actually done by a postdoc of mine, and she moves so quickly into her scientist's role, I forgot how fun this was.

356
00:33:41,810 --> 00:33:50,660
And this actually came out last year where we did a realist synthesis to understand context mechanism outcomes for learning, um, organisations.

357
00:33:50,660 --> 00:33:54,440
She wants to go and figure out what makes a learning health system a learning health system,

358
00:33:54,770 --> 00:33:58,669
and what are the mechanisms that drive that learning health system to exist.

359
00:33:58,670 --> 00:34:02,569
Well, so she's building on this work, but this one came out, um, just last year.

360
00:34:02,570 --> 00:34:04,700
And this is our foundational work. There's Marissa there.

361
00:34:05,270 --> 00:34:10,639
Um, and where we wound up finding, you know, 60 different CMO groupings across the literature,

362
00:34:10,640 --> 00:34:13,940
but we were able to distil them down into six big areas.

363
00:34:14,300 --> 00:34:18,290
Um, and they kind of look like this as an example.

364
00:34:18,290 --> 00:34:21,800
So this is one of our six. So embracing risk and failure.

365
00:34:22,070 --> 00:34:28,820
Uh, a CMO in this space looks like when evaluation is promoted as a learning activity within an organisation.

366
00:34:28,820 --> 00:34:37,850
So learning health systems or learning learning organisations, um, that tolerates risk and does not shame and blame staff for evaluation failures.

367
00:34:37,850 --> 00:34:44,660
That's for context. Uh, staff involved in the evaluation are more apt to engage with and learn from that evaluation.

368
00:34:44,750 --> 00:34:53,840
There's your outcome. And because they do not fear your mechanism, disciplinary punitive measures associated with evaluation failure.

369
00:34:54,290 --> 00:34:58,069
So we have our context, our environment of the outcome we're trying to get to.

370
00:34:58,070 --> 00:35:02,420
And then our mechanism in the middle. So realist thinking really does help us unpack what the mechanism is.

371
00:35:03,050 --> 00:35:06,350
Uh, we were able to distil the 16 into these six things.

372
00:35:06,350 --> 00:35:12,079
And she's in the process of testing, um, on a now, what is really fun about this paper, um,

373
00:35:12,080 --> 00:35:21,740
that I reread and Marissa and I talked about so much is that each of those six kind of big buckets are really at this underlying level.

374
00:35:22,220 --> 00:35:26,209
That first one mechanism that's about fear, that's second one.

375
00:35:26,210 --> 00:35:32,150
If you go and read the paper, it's going to be all about provenance. That third one's about a sense of ownership and values.

376
00:35:32,480 --> 00:35:35,150
The fourth one is about meaningfulness of data.

377
00:35:35,270 --> 00:35:40,520
I love this one because we've been doing this in digital health for a long time, and meaning and data is so crucial.

378
00:35:41,090 --> 00:35:46,850
That fourth one, the fifth one is about motivation and the last one is about kind of contributed role.

379
00:35:46,860 --> 00:35:49,910
So we're back to like a shared role and share identity in that last one.

380
00:35:50,150 --> 00:35:54,470
So all six of them are operating at this lower level.

381
00:35:54,530 --> 00:35:56,299
Yeah. But yeah just a quick question.

382
00:35:56,300 --> 00:36:05,240
Um, so I guess several, but um, I'm looking at the learning system and I'm wondering where does the data come from?

383
00:36:05,420 --> 00:36:12,110
And with regards to the evaluation piece, has there been any conversations about, um,

384
00:36:12,110 --> 00:36:18,559
what forms that takes into or is, is is sort of, um, there specific teams that sort of do that?

385
00:36:18,560 --> 00:36:24,340
Is it. Everybody does it together. Is individuals having that reflective spirit within, uh, team.

386
00:36:24,350 --> 00:36:26,720
I'm really curious for that part of that question.

387
00:36:26,810 --> 00:36:31,639
Just to clarify, are you asking about like, data that informed building this framework or like data used in a learning health system?

388
00:36:31,640 --> 00:36:34,100
No, no, the data in the learning health system, in the learning health.

389
00:36:34,110 --> 00:36:40,489
And so, um, Marissa and that team over actually in health partners, they've been doing their entire system is learning health systems.

390
00:36:40,490 --> 00:36:43,610
So they have a really nice framework that talks about different sources of data.

391
00:36:44,120 --> 00:36:50,360
Um, it can it helps when if it's coming from something the organisation's already collecting anyway.

392
00:36:50,810 --> 00:36:54,140
Um, because trying to give them new stuff to do is very hard.

393
00:36:54,440 --> 00:37:02,000
But what you end up doing is shifting, um, their understanding of the data and presenting it in a different way in order to adapt these things.

394
00:37:02,450 --> 00:37:08,270
So, for example, you can get to a sense of ownership and value if you sit down with the clinicians.

395
00:37:08,380 --> 00:37:12,940
Who you'd be asking to do something different and say, how do you think about this data?

396
00:37:13,000 --> 00:37:16,270
What do you think we should be collecting? How should it be presented to you?

397
00:37:16,420 --> 00:37:20,010
And suddenly they're part of the conversation. It's making sense to them.

398
00:37:20,020 --> 00:37:27,099
So you're going to hook into the data is meaningful piece. And then they feel like it's not something that's been thrust upon them.

399
00:37:27,100 --> 00:37:33,460
It's something they actually went and asked for. Um, so often I doing it that way, it's helpful.

400
00:37:33,770 --> 00:37:41,020
Um, my other wonderful colleague, Harry Carlucci, does a ton in patient experience, and she took the patient experience data they were doing already.

401
00:37:41,020 --> 00:37:45,580
She tweeted to improve it. And Hannah has now expanded that into caregiver experience.

402
00:37:45,970 --> 00:37:52,450
And she worked with clinicians and worked with her decision and the kind of key decision makers to present data in real time,

403
00:37:52,450 --> 00:37:56,710
back to clinicians on tablets. And it just exists in their environment.

404
00:37:57,160 --> 00:38:02,590
And because they're kind of competitive people, they like to see how they're doing and they like to see how their neighbours are doing.

405
00:38:02,590 --> 00:38:11,140
So it's presented unity unit, but they also extracted really meaningful, um, quotes from the experience surveys,

406
00:38:11,530 --> 00:38:18,220
which hooked into the motivation thing and the confidence thing really nicely because suddenly they're like,

407
00:38:18,520 --> 00:38:24,250
wow, I really made a difference to this person. I'm motivated to do more things and I'm confident I'm doing a good job.

408
00:38:24,640 --> 00:38:30,880
So they were able to pull on a couple of different mechanisms, um, in order to present data that was meaningful and useful to them.

409
00:38:31,870 --> 00:38:35,599
At the end of the day, when it gets into kind of like meaningfulness of data and usefulness,

410
00:38:35,600 --> 00:38:42,040
and a lot of these components co-production of the process and understanding in that collective sense making thing that kind of talked about.

411
00:38:42,040 --> 00:38:45,309
I have a whole other lecture on citizens making I didn't get last week.

412
00:38:45,310 --> 00:38:54,130
I didn't bring it in here, but we can talk about it. Um, is really helpful to overcome things like threat and fear of data and new stuff.

413
00:38:54,820 --> 00:38:58,840
Because you're working together, you're often having to change with a lot of different people.

414
00:38:59,140 --> 00:39:03,280
And if you all understand that differently, it can feel very threatening and very scary.

415
00:39:03,700 --> 00:39:05,250
Same thing with presenting data.

416
00:39:05,320 --> 00:39:10,780
It can feel like it's being punitive if you're not having a conversation about what it means and why it's valuable early on.

417
00:39:12,130 --> 00:39:15,390
To help. Yeah. Yeah.

418
00:39:16,370 --> 00:39:19,970
Really quickly. And thanks. This is really fascinating.

419
00:39:19,990 --> 00:39:28,309
I'm wondering about the organisational preconditions, however. To what extent and to what level of success can you have?

420
00:39:28,310 --> 00:39:36,230
This framework placed atop an organisation that has not yet figured out the organisational preconditions?

421
00:39:36,950 --> 00:39:41,059
Yeah, I mean, it's it's such a wonderful question.

422
00:39:41,060 --> 00:39:47,690
And I think one we just are still grappling with because we don't have enough learning health systems right now to do comparative work.

423
00:39:48,110 --> 00:39:54,560
And the ones who are doing it already have the preconditions in place, for example, valuing data and valuing research.

424
00:39:54,890 --> 00:40:00,620
And, um, having a different view, like having that safety kind of culture and,

425
00:40:00,620 --> 00:40:03,919
um, kind of what Amy Edmondson talks about being like a fearless organisation.

426
00:40:03,920 --> 00:40:12,950
They already have that. Um, I have yet to see an example of an organisation go from being top down punitive into this model yet.

427
00:40:13,370 --> 00:40:21,349
Um, I think she's she's keen to find it. So one of the things she's been doing with some colleagues at McMaster is building a network of organisations

428
00:40:21,350 --> 00:40:26,660
that are either already identified as a learning health system or aspiring to be a legal system,

429
00:40:27,020 --> 00:40:31,429
and she's trying to they're trying to unpack that next piece. But you're right.

430
00:40:31,430 --> 00:40:35,390
Some of these like the context pieces and some of the preconditions, um,

431
00:40:35,720 --> 00:40:44,120
might rest on our organisational cultures that are very hard to move and sometimes very driven by particular types of leaders.

432
00:40:44,780 --> 00:40:48,680
Um, so they're also trying to unpack that piece as well. And it's tricky.

433
00:40:50,450 --> 00:40:54,570
Not mention. There's also political, uh. Um, so very cool work.

434
00:40:54,590 --> 00:41:04,040
I highly encourage you to follow rhythm. Um, another piece of work we did this is older, but this is also this is more of a team level, um, mechanism.

435
00:41:04,040 --> 00:41:09,739
And we actually did a trial of a technology that we had built and co-produced with patients,

436
00:41:09,740 --> 00:41:14,389
their families and primary care teams to do goal oriented care.

437
00:41:14,390 --> 00:41:18,920
So shifting a model of care to asking patients what matters to them,

438
00:41:19,190 --> 00:41:23,149
and then wrapping care plans around them and monitoring them against goals that they cared about.

439
00:41:23,150 --> 00:41:28,700
So that's goals like, I would like to walk my dog every day, not I need a lower HBA one C number.

440
00:41:29,090 --> 00:41:32,989
Um, and so we went through many, many iterations of this.

441
00:41:32,990 --> 00:41:36,980
And at the end of the trial, we had lots of people using it and lots of people not using it,

442
00:41:37,430 --> 00:41:46,850
and the people who really used it who really changed behaviour, both the clinicians and the patients had this very strong alignment of coherence.

443
00:41:47,570 --> 00:41:54,290
So that became a mechanism we care a lot about. And we have been building into kind of trials of our next rounds of technology.

444
00:41:54,890 --> 00:41:58,310
Um, because we often build a technology for like one user group.

445
00:41:58,850 --> 00:42:02,210
But a technology might actually be interactional and asking multiple people.

446
00:42:03,110 --> 00:42:11,419
And when we saw the highest amount of use from both the physicians and in clinical teams and patients like people we called super users,

447
00:42:11,420 --> 00:42:14,390
they were using this thing every day, whether they really needed to or not.

448
00:42:14,780 --> 00:42:20,910
They had this really strong alignment between their sense of coherence, both at the individual and team level.

449
00:42:20,930 --> 00:42:22,579
It made a lot of sense to them.

450
00:42:22,580 --> 00:42:29,090
The tool did, and it made a lot of sense to their relationships together, and that changed their behaviour really meaningfully.

451
00:42:29,510 --> 00:42:36,620
So this we've kind of reimagined normalisation process theory into this framework from a huge narrative analysis of a ton of data.

452
00:42:36,620 --> 00:42:38,690
And to actually get to something that looks pretty simple.

453
00:42:39,020 --> 00:42:43,850
But at the end of the day, that's moved us to think differently about technology, around what's meaningful to people.

454
00:42:44,120 --> 00:42:50,000
And that's a mechanism we kind of chase now, uh, we try to build it into the way we build the innovation.

455
00:42:50,330 --> 00:42:55,250
We build it into the way we do our implementation work as well. Meaningfulness really matters.

456
00:42:55,250 --> 00:42:58,250
Even in technology. They're not inert. They matter. They're things.

457
00:42:58,250 --> 00:43:03,170
They're real social artefacts. People have feelings. Lots of thoughts and feelings about our technology.

458
00:43:04,430 --> 00:43:08,930
Okay. The last one we'll talk about is kind of diffusion and adoption mechanisms.

459
00:43:08,930 --> 00:43:15,020
So this is where things get like really fun and sticky where things are just like dispersing out whether you can measure it or not.

460
00:43:15,020 --> 00:43:21,440
And um, there's so many fun studies of complex adaptive systems theory, um, where you can measure the first couple of stages.

461
00:43:21,440 --> 00:43:24,950
And as it gets further and further into embedding, it gets really, really hard to measure,

462
00:43:25,430 --> 00:43:30,650
um, which is the fun and challenging thing of a complex adaptive systems theory.

463
00:43:31,160 --> 00:43:35,629
But this is getting into bigger stuff like tipping points and betting and emergence social proof.

464
00:43:35,630 --> 00:43:38,870
And so diffusion of innovation is helpful for us here.

465
00:43:39,170 --> 00:43:45,079
The dynamic sustainability framework for us is helpful here. And of course, complex adaptive systems theory that's helpful for us here.

466
00:43:45,080 --> 00:43:48,950
All of them suggest why things go from one place to the other.

467
00:43:49,460 --> 00:43:57,680
So mechanisms here are social contagion, norm cascades, institutionalisation of change that occurs over time.

468
00:43:57,680 --> 00:44:01,549
So this is when our change is getting bigger and bigger. This is actually work from Scarborough.

469
00:44:01,550 --> 00:44:05,690
Um uh, care that says uh, from 2022.

470
00:44:05,690 --> 00:44:10,219
Joe, do you know what this is? I know Harry. Yeah. Oh, yeah. It's a really nice little piece of work.

471
00:44:10,220 --> 00:44:14,930
Yeah. What's a big piece of work? Uh, that they connected micro and macro level phenomenon.

472
00:44:15,260 --> 00:44:22,790
Um, so they were looking at mechanisms that were occurring locally that cascaded into mechanisms that were happening more broadly within the system.

473
00:44:23,180 --> 00:44:26,540
Um, which was super cool. It's a great paper. I highly recommend reading it.

474
00:44:26,840 --> 00:44:29,899
And they found things like learning, adapting,

475
00:44:29,900 --> 00:44:34,879
and institutionalising that occurred in a local space would also cascade into much

476
00:44:34,880 --> 00:44:38,300
bigger change of the social level over time when they were looking at the literature.

477
00:44:38,840 --> 00:44:45,290
Um, and they embed, you know, Rogers diffusion of innovation theory and a couple of other theory which make the

478
00:44:45,410 --> 00:44:50,150
ultimate impact on the structure they build in cognitive and normative other,

479
00:44:50,180 --> 00:44:56,240
uh, theoretical frameworks to, um, to show how those kind of key mechanisms cascade forward into something bigger.

480
00:44:57,050 --> 00:45:03,170
Um, I have another wonderful student, um, Marsha Gupta, who, um, finished a PhD some years ago.

481
00:45:03,800 --> 00:45:06,710
She ended up writing like five papers on her PhD. And here's two of them.

482
00:45:07,130 --> 00:45:11,870
She looked at kind of the emergence of family medicine in India for her work.

483
00:45:12,230 --> 00:45:17,150
So fascinating. Um, I highly recommended your diffusion of innovation theory,

484
00:45:17,150 --> 00:45:22,490
for I was her theory nerd on her thesis because she was mostly supported by other family docs.

485
00:45:22,940 --> 00:45:26,059
Um, and she did some quality. These are her qualitative,

486
00:45:26,060 --> 00:45:32,360
descriptive and survey studies where she looked at emergence of family medicine over a 30 year period and was able to find

487
00:45:32,360 --> 00:45:39,469
this link between the work of innovators and early adopters to do institutionalisation work that allowed family medicine to,

488
00:45:39,470 --> 00:45:46,640
like, take root in the entire country. Um, so it the potential that so she didn't go at it from a mechanistic lens.

489
00:45:46,640 --> 00:45:52,549
But you can kind of apply some of that coherence, um, criteria, um, to understand these as mechanisms,

490
00:45:52,550 --> 00:45:58,100
things like having sufficient training programs that were widely dispersed was really critical,

491
00:45:58,100 --> 00:46:05,240
engaging a multidisciplinary model because of the very different kinds of discipline approach and disciplinary approaches to medicine in India.

492
00:46:06,080 --> 00:46:10,910
Also absolutely important to kind of embed into the norms of the society there.

493
00:46:11,180 --> 00:46:17,239
And then having increasing postgraduate training was a way of institutionalising it into the entire system.

494
00:46:17,240 --> 00:46:18,649
So over this 30 year period,

495
00:46:18,650 --> 00:46:26,390
they went from not knowing what family medicine was to having a burgeoning group of clinicians who were delivering care all over India.

496
00:46:26,750 --> 00:46:32,150
Really nice piece of work that shows this kind of how mechanisms can occur in these much higher order ways.

497
00:46:33,720 --> 00:46:39,720
So these things all matter and they're really cool. But they also shift how we think about our strategy.

498
00:46:40,260 --> 00:46:42,670
So not only understanding why something happened.

499
00:46:43,140 --> 00:46:48,360
Um, I like to talk to our leaders about they, they also tell us maybe which strategies you should target,

500
00:46:48,660 --> 00:46:50,970
because it's going to pull that lever of change for you.

501
00:46:51,390 --> 00:46:55,890
So a lot of leaders kind of intuit the different strategies you might use to activate a mechanism.

502
00:46:56,220 --> 00:47:01,320
But the difference between using theory and not is if you don't, you might be trying eight different things,

503
00:47:01,320 --> 00:47:05,480
and if you do, you might only have to try to in order to actually get something going.

504
00:47:05,490 --> 00:47:09,040
So I have an example from that technology example.

505
00:47:09,040 --> 00:47:18,000
I you the core of two ways we were thinking about this. So with when we were doing this at the primary care level, there's our goal or indicator tool.

506
00:47:18,420 --> 00:47:22,290
A really important enabler for us was getting clinician champions.

507
00:47:22,290 --> 00:47:25,500
So I'm sure you've heard this before. Clinician champions matter a lot.

508
00:47:25,890 --> 00:47:33,270
And at our primary care teams in Ontario, they kind of function somewhat independently, but they're very much a cohesive team.

509
00:47:33,660 --> 00:47:39,750
There's usually like a person not even with a title that's like a leader, but just a person in the team that everyone looks to for guidance.

510
00:47:40,050 --> 00:47:45,500
In some of our family health teams, it was the social worker and some of our family health teams, it was the admin.

511
00:47:45,510 --> 00:47:48,630
Sometimes it was a particular family doctor.

512
00:47:49,290 --> 00:47:52,970
Um, so we were getting them on board to this technology.

513
00:47:52,980 --> 00:47:58,049
They kind of came in and championed it, and then they got the whole unit on board so we could run the trial.

514
00:47:58,050 --> 00:48:04,350
So that was a really important mechanism for us to pull on. So, you know, fast forward five, six years later,

515
00:48:04,470 --> 00:48:09,620
and we're integrating this technology into a care planning tool at a hospital in two

516
00:48:09,630 --> 00:48:14,010
hospitals to support discharging from hospital to home for same kinds of patients.

517
00:48:14,010 --> 00:48:19,140
Very complex. This is work we did with Trillium, and we're doing it on my hospital at Finding Health.

518
00:48:19,680 --> 00:48:22,290
So we've gone through all of this co-production work,

519
00:48:22,290 --> 00:48:28,409
worked very closely with our hospital teams as well as our primary care teams, and we are ready to get this thing going.

520
00:48:28,410 --> 00:48:32,160
So we went to the back to the same strategy. We said, okay, here are our clinician champions.

521
00:48:32,160 --> 00:48:36,030
We identified them on the units across all of these different hospital spaces.

522
00:48:36,420 --> 00:48:40,350
And the champions were great champions and they continue to be great champions.

523
00:48:40,350 --> 00:48:46,410
They were rah rah, rah rah for this thing the whole time. But people really weren't signing up for a trial.

524
00:48:46,920 --> 00:48:51,479
And we were at this for months, and the champions did everything they could to talk to everybody.

525
00:48:51,480 --> 00:48:57,510
They invited us for lunch and learns. They invited my team to be on the units all the time, but nothing was really sticking.

526
00:48:58,320 --> 00:49:05,580
And then kind of one day the hospital decided to do a communication about the digital bridge project, and they have highlighted, um, us.

527
00:49:05,610 --> 00:49:10,200
They highlighted the wonderful students that were kind of all over the place, trying to work hard to recruit.

528
00:49:10,530 --> 00:49:18,540
They talked about its connection to Sinai Health and what was important to Sinai, and advancing compassionate care and connection to the community.

529
00:49:18,840 --> 00:49:26,790
And suddenly 20 people signed up for my. Yes. So that tells us a very different lever of change in primary care teams.

530
00:49:26,810 --> 00:49:33,830
It was about the person that they knew and their connection to that person that they knew, that made that behaviour change happen for others.

531
00:49:34,340 --> 00:49:35,299
For this environment,

532
00:49:35,300 --> 00:49:43,100
it was their connection to their hospital and their sense of being a team player with their hospital that hooked them in to change their behaviour.

533
00:49:43,580 --> 00:49:46,990
So, you know, we didn't know that before, but now we do that.

534
00:49:47,000 --> 00:49:54,979
Maybe before you go out and set up a strategy or an adoption strategy, you ask people like, what's the most meaningful thing to you?

535
00:49:54,980 --> 00:50:00,440
Like, who do you go to for help? What makes you feel like you're connected to this role and to this organisation?

536
00:50:00,920 --> 00:50:08,010
And if you ask those deeper questions about what drives people and what makes them want to do new things, then it changes your strategy.

537
00:50:08,030 --> 00:50:10,940
And we would have saved like eight months if you talk to someone.

538
00:50:11,570 --> 00:50:16,370
If we had just asked the communications team to write a little blurb about us much earlier.

539
00:50:16,670 --> 00:50:19,790
So the question that's just popped in that, I think,

540
00:50:19,940 --> 00:50:27,440
relevant to the whole question of saying when you're doing implementation work, it can be that multiple mechanisms are at play.

541
00:50:27,800 --> 00:50:34,850
So when that's happening, how do you how do you navigate appropriate models or do you incorporate multiple models and frameworks in this case?

542
00:50:35,120 --> 00:50:38,790
In other words, how do you work through the complexity of the real world messy implementation?

543
00:50:38,810 --> 00:50:42,110
Oh, talking about okay, what a wonderful question.

544
00:50:42,290 --> 00:50:50,419
I think I have a slide coming up on this. Um, I do have a slide coming up on this, but, uh, there are people who, um, remember the name.

545
00:50:50,420 --> 00:50:55,909
Anyway, the names on the slide, um, that you can embed lots of different, um, theoretical frameworks.

546
00:50:55,910 --> 00:51:00,560
They do need to be epistemologically aligned. Um, that becomes really important.

547
00:51:00,570 --> 00:51:07,790
Um, whether you're doing research or you're trying to guide mechanistic thinking, because some theories rest on some assumptions,

548
00:51:07,790 --> 00:51:12,890
and theories can rest on conflicting assumptions so important to see that they conceptually map together.

549
00:51:13,400 --> 00:51:17,450
Um, but you can kind of work them together if they're meeting,

550
00:51:17,450 --> 00:51:23,850
if they're not in conflict and they're meeting a gap that each other don't let each other aren't hitting.

551
00:51:23,870 --> 00:51:29,900
So if you want to like, layer cognitive and social theories together and are not really in conflict, that can help.

552
00:51:30,200 --> 00:51:37,009
And you see that happening in research, right. Um you'll see com being linked with for because com is not doing a good job of the environment.

553
00:51:37,010 --> 00:51:39,440
So they at C version get to the environment stuff.

554
00:51:40,040 --> 00:51:49,520
Um so you can the challenge and I don't know that I certainly haven't unpacked it and I haven't seen it unpacked yet is what you prioritise.

555
00:51:50,210 --> 00:51:56,360
Um, so if you have many mechanisms at play, which one you pick first, whether they want,

556
00:51:56,430 --> 00:52:02,720
there's evidence that they actually can chain and cascade and some mechanisms need to happen first and others have happened later.

557
00:52:03,080 --> 00:52:09,200
But it's still kind of a messy literature. Unless if somebody like read something that I haven't read yet and I'm not sure.

558
00:52:09,470 --> 00:52:18,230
Yeah, there is automatically. So we, we tend to go with, um, I like to go with talking to people and asking them like what matters to them the most.

559
00:52:18,620 --> 00:52:24,350
And then I prioritise that, like what they're thinking about in that moment, what the organisation cares about in that moment.

560
00:52:24,680 --> 00:52:29,989
And then I'll try to cascade, um, kind of strategies that linked to those mechanisms first, um,

561
00:52:29,990 --> 00:52:33,920
that can be aware because ultimately we're just trying to get people to do something different.

562
00:52:34,340 --> 00:52:38,480
Um, so you need to hook into the people. I am very possible.

563
00:52:38,840 --> 00:52:42,260
I do encourage going back to assumptions so we're not in conflict.

564
00:52:44,410 --> 00:52:50,170
All right, I'm gonna make you do more work. Okay, so here's a drawing of how you could potentially draw mechanisms.

565
00:52:50,470 --> 00:52:58,450
This is, again, Carroll Lewis's work. They're pretty intense, but if you're a visual person, they can actually be very helpful.

566
00:52:58,480 --> 00:53:03,580
You don't need to know all the components. So you had your, um, your list of thing.

567
00:53:03,910 --> 00:53:09,070
Your thing. The stuff in the middle and your outcome. You can actually logic model these guys out.

568
00:53:09,320 --> 00:53:13,910
You can see about how they're connecting. Um, here are I going to describe all the different.

569
00:53:13,960 --> 00:53:19,840
So you might just want to pick one. So here you might have either you're saying or your strategy whatever you're putting in place.

570
00:53:20,970 --> 00:53:27,270
This is maybe something like a preconditions. So like organisational preconditions we are sort of talking about from Maurice's work.

571
00:53:27,270 --> 00:53:31,650
So what actually has to be why maybe you might just need the funding to do the thing at all.

572
00:53:32,040 --> 00:53:34,140
Um, without it, nothing happens. Anyway.

573
00:53:34,650 --> 00:53:41,700
Uh, then you have your mechanism, but we are just chatting about important things that we need to think about with mechanism.

574
00:53:41,700 --> 00:53:49,320
So there are these moderators and mediators that Carol Lewis and company and other theorists in the space like to talk about.

575
00:53:49,860 --> 00:53:57,420
These are factors that can turn up a mechanism or turn down a mechanism or shut out the mechanism entirely.

576
00:53:57,900 --> 00:54:05,969
So sometimes, for example, something might be very meaningful to a person, but it's not an organisational priority.

577
00:54:05,970 --> 00:54:09,750
So they're not given any time. So that's shut down that mechanism entirely.

578
00:54:10,560 --> 00:54:13,830
Um, or sometimes there's no political will to do the thing.

579
00:54:13,830 --> 00:54:21,910
So shut down entirely. So these, these are things that can come in and just sweep away because they have more power or more, um,

580
00:54:21,990 --> 00:54:27,990
ability to either take away a precondition or kind of make the mechanism mute for that particular moment in time.

581
00:54:30,140 --> 00:54:34,700
Let's do some drawing or. Super fun.

582
00:54:35,420 --> 00:54:38,570
I won't do it. Um. Pick one. Just pick one mechanism.

583
00:54:38,780 --> 00:54:43,790
Chat with your friend again. Those online. Please go ahead and find a piece of paper and you will give it a go.

584
00:54:44,270 --> 00:54:47,719
Try to pick one mechanism. It can be at any of the levels we talked about.

585
00:54:47,720 --> 00:54:51,890
Could be a cognitive one. Could be a team one organisation social system.

586
00:54:52,340 --> 00:54:57,410
Um, the cognitive ones are sometimes a little easier to start with because you can feel them in your body.

587
00:54:58,070 --> 00:55:00,800
So those I don't see. So, um. Don't worry.

588
00:55:01,130 --> 00:55:09,650
You know, you can't put a determinant in there, but I think we want to focus on these areas outcome moderators, mechanisms and either strategy.

589
00:55:10,670 --> 00:55:15,380
So does any want to share anyone want to share with where they got you. There's no right or wrong here.

590
00:55:16,380 --> 00:55:19,440
So I'm going to do a little bit on what they offer us. And then I'm going to go to our wrenches.

591
00:55:20,310 --> 00:55:26,310
Um, so remember so I've already talked about this. Remember this study before because we know this now we can do different strategies.

592
00:55:26,320 --> 00:55:30,479
So I gave an example of that with our technology. But so say we're doing virtual care.

593
00:55:30,480 --> 00:55:33,600
If I know that virtual care hinges on professional identity.

594
00:55:33,600 --> 00:55:39,210
And they told me that they were training to be physicians in a particular way, well, then I need to train physicians in a different way.

595
00:55:39,300 --> 00:55:42,450
I need their identities to be not. I'm not a doctor in a room.

596
00:55:42,450 --> 00:55:45,780
I am a doctor that gives care in multiple modalities.

597
00:55:46,200 --> 00:55:51,420
That's a different way of socialising people, and that's a different kind of strategy that you would have to think about kind of earlier days.

598
00:55:51,870 --> 00:55:55,050
Um, so mechanisms can help us tailor our strategies.

599
00:55:56,040 --> 00:56:01,199
Um, they can also help us understand failure. So this is a fun exercise.

600
00:56:01,200 --> 00:56:07,470
If it was to do in all your humility when something goes wrong in so many of our trials, do not land the way we do,

601
00:56:07,620 --> 00:56:12,660
we will have to do this for my trial of our digital bridge project, because there was a lot of implementation failure.

602
00:56:13,020 --> 00:56:19,970
Um, and this was one for a clinical decision support tool where they did all the right kinds of strategies.

603
00:56:19,980 --> 00:56:23,760
They did their clinical champion, like the example I gave, they did their training.

604
00:56:24,180 --> 00:56:29,030
Um, and yet it failed. And they were able to uncover mechanisms of failure.

605
00:56:29,040 --> 00:56:32,369
And in these cases, it was a lack of confidence in the tool.

606
00:56:32,370 --> 00:56:37,739
So similar to Michelle's example, um, that can lead you down a you thrive or survive, pass fail.

607
00:56:37,740 --> 00:56:41,480
I'll talk about in a second. Um, and competing clinical priorities.

608
00:56:41,490 --> 00:56:44,490
Both of those things got in the way. Um, this is a great paper.

609
00:56:44,490 --> 00:56:50,190
Um, that came out if you're if you're keen on, um, you know, failure from us that that came out just earlier this year.

610
00:56:50,520 --> 00:56:53,540
What you'll see in these mechanisms papers as you get into them.

611
00:56:53,550 --> 00:56:57,990
Um, and I think some of the conversations were coming out like this that mechanisms

612
00:56:57,990 --> 00:57:02,190
can get changed and there can be many of them within a particular intervention.

613
00:57:02,370 --> 00:57:08,999
So you can actually have five, 6 or 7 of these, um, all in one row, which is anyone have kids?

614
00:57:09,000 --> 00:57:13,920
I said 6 or 7. And I can hear my daughter, like all it's her whole life.

615
00:57:13,920 --> 00:57:18,209
Right now she's ten. Um, mechanisms can also cascade.

616
00:57:18,210 --> 00:57:22,650
So we have that. Ah, um, that example before um, this I academic,

617
00:57:22,650 --> 00:57:28,170
you and company did a lovely scoping review of employee driven innovation a little while ago, and they,

618
00:57:28,170 --> 00:57:36,210
um, I was very kindly invited to write a commentary on it, uh, where we, uh, I looked at the literature a little bit and showed how, you know, um,

619
00:57:36,480 --> 00:57:43,080
mechanisms happening within an employee driven innovation space can cascade into the same things that drive large system change,

620
00:57:43,350 --> 00:57:48,870
things like distributed leadership achieving better fit between innovation and local context in alignment,

621
00:57:48,870 --> 00:57:56,680
um, uh, around kind of what is perceived to be meaningful that if you do that is small, that can cascade into bigger and bigger places.

622
00:57:56,710 --> 00:58:00,630
Those are mechanisms that can potentially transfer into other spaces.

623
00:58:02,180 --> 00:58:07,820
This is a mechanism I'm a little bit geeked about. This is Alex Honnold climbing type 101 earlier this year.

624
00:58:07,850 --> 00:58:13,690
Anybody watch that while sweating? I was on I was on my bike trainer at the time which was not a good idea.

625
00:58:13,700 --> 00:58:20,200
My heart rate was too high. And so when I've been working on a little bit and here's a drawing I've done for you.

626
00:58:20,210 --> 00:58:29,260
Sorry. Online. No, I'm off camera. Um, is how threat that exists at different levels can lead to different kinds of responses.

627
00:58:29,260 --> 00:58:31,219
So some people have kind of theorised this,

628
00:58:31,220 --> 00:58:39,380
again from a cognitive psychology space where they've shown that and particularly in accelerated accelerated change environments,

629
00:58:39,560 --> 00:58:44,450
that we can elicit this threat response, which can lead people to survive.

630
00:58:44,810 --> 00:58:53,300
So that's like when you present a new idea in a room and either you or somebody gets like angry or shuts down or looks at their phone or leaves,

631
00:58:53,540 --> 00:58:59,810
that's like your fight, flight or freeze. Um, example of when something new was presented and it feels threatening.

632
00:59:00,200 --> 00:59:06,049
Or you can go into a thrive space where you embrace the fear and climb this crazy building that looks horrible.

633
00:59:06,050 --> 00:59:12,830
Then, um, in implementation science, it's looking like things like resistance or anxiety or loss or perceived risks,

634
00:59:13,280 --> 00:59:16,630
but they're kind of sitting in that space of threat a lot.

635
00:59:16,640 --> 00:59:23,750
And the reason I'm interested in this is it seems to be happening, um, in technology environments that when you present,

636
00:59:23,750 --> 00:59:32,659
especially when somebody has gone through a really bad technology implementation before you present something new and it's a total shutdown,

637
00:59:32,660 --> 00:59:36,560
like an actual threat response to something that they don't want to go through again.

638
00:59:37,040 --> 00:59:44,269
And so it's a mechanism I think it's under theorised. And I'm working on a little bit, uh, around the nature of threat in terms of where it sits,

639
00:59:44,270 --> 00:59:49,280
not just at the individual level but at multiple levels, and that it can look like different things.

640
00:59:49,280 --> 00:59:54,830
It can look like identity, it can look like norms, it could look like survival at an organisational level.

641
00:59:55,130 --> 00:59:59,060
And so if I were drawing a mechanism drawing, I mentioned one I've been working on for months.

642
00:59:59,060 --> 01:00:03,890
This is one of them you can draw. So say someone's going through tool training.

643
01:00:04,460 --> 01:00:12,110
Maybe they're a radiologist. And that is leading to some rule uncertainty, whether they're actual and uncertainty being a precursor to threat.

644
01:00:12,500 --> 01:00:17,720
And then they feel some professional identity threat. And now we're either going to go into survival mode or thrive mode.

645
01:00:18,290 --> 01:00:22,610
So our Meccan our moderators in this example would be their past experience.

646
01:00:22,910 --> 01:00:26,150
How have they seen something like this before and what are they likely to expect?

647
01:00:26,690 --> 01:00:32,420
And then how meaningful this thing is to them in relation to their perceived identity and threat.

648
01:00:32,840 --> 01:00:37,310
And so when I've seen meaningfulness turned up, that's a dial that can go up.

649
01:00:38,420 --> 01:00:44,030
Thrive happens because they're willing to tolerate the risk and the uncertainty, because the outcome of the outcome,

650
01:00:44,540 --> 01:00:50,180
when their past experience suggests this is going to go terribly, they're more they seem to be more likely to happen here.

651
01:00:50,540 --> 01:00:54,650
So this is a lot of what I'm lecturing on in the next couple of months at different places.

652
01:00:54,890 --> 01:00:59,930
And I wanted to share with you guys because it's a learning process, and I'm also keen on what you guys think about that.

653
01:01:00,650 --> 01:01:03,690
Come on, come on. Them. Yes, please. Since.

654
01:01:05,520 --> 01:01:08,670
I like it. Um, I think.

655
01:01:08,850 --> 01:01:15,629
I think we we all can think of an example where somebody has gone into that freeze mode because of a

656
01:01:15,630 --> 01:01:21,110
bad experience with the technology that was introduced that then kind of completely wrecked that job.

657
01:01:21,450 --> 01:01:24,690
And then they were then doing things they didn't like and didn't.

658
01:01:24,960 --> 01:01:30,150
But also the effort you put into the technology and then the whole thing fails.

659
01:01:30,570 --> 01:01:35,500
Um. I'm not sure that meaningfulness is the only thing in that full.

660
01:01:36,580 --> 01:01:43,000
I remember we were doing a study years ago, uh, which was a seven site,

661
01:01:43,000 --> 01:01:51,040
randomised controlled trial of telehealth in heart failure and was going well in six sites and then site seven, which was a plane ride away.

662
01:01:51,520 --> 01:01:58,840
Um, nobody was being recruited. So Joan and I got on a plane and went off to this place where they weren't encouraging anyone.

663
01:01:59,110 --> 01:02:02,560
We had all these kind of unusual things about, like, what? You're recruiting.

664
01:02:03,040 --> 01:02:07,629
And then we picked up the stories, and I think we got the stories in a taxi from someone, you know,

665
01:02:07,630 --> 01:02:15,550
some of those informal things, the story of something that had been introduced four years ago, that was just a disaster.

666
01:02:15,790 --> 01:02:21,880
And then everyone got blamed and punished and and people were still in that freeze mode.

667
01:02:22,240 --> 01:02:31,060
Now the challenge for us was not just to make this meaningful, but also at a technical level,

668
01:02:31,360 --> 01:02:40,240
to reassure them that the the technology that we were introducing was not going to go as badly wrong as the technology that was in Chicago.

669
01:02:40,390 --> 01:02:47,080
And actually, one of the things was that the technology had moved on, but that it was a much, much better technology than the one.

670
01:02:47,080 --> 01:02:50,229
Yeah, that was being introduced. So it wasn't just meaningful.

671
01:02:50,230 --> 01:02:54,070
It was like, is this thing going to work because it didn't work last time?

672
01:02:55,660 --> 01:02:58,870
And value. So. Yeah. Yeah. Yeah. You know.

673
01:02:58,870 --> 01:03:03,579
Whatever. Yeah. That's great, I love that. Um, I have a whole lot of things I like.

674
01:03:03,580 --> 01:03:06,820
Narrative approaches is really important because stories matter.

675
01:03:07,210 --> 01:03:10,390
Um, I love that. I'm going to ask you more about that. Um.

676
01:03:11,590 --> 01:03:15,160
Thank you. I'm collecting stories right now. I'll come back to that.

677
01:03:15,760 --> 01:03:20,420
Um, yes, because that's looking back into past experience and that's changing how I kind of value meaningful,

678
01:03:20,440 --> 01:03:26,349
because there are some challenges for our field that we need to keep going. And I want to elevate these because you're probably going to experience.

679
01:03:26,350 --> 01:03:31,390
And if you're going out and trying to study mechanisms, one is the definitional clarity issue.

680
01:03:31,630 --> 01:03:37,720
Uh, we still are missing it. So if you're going out just just pick one that resonates with the the environment that you're working in.

681
01:03:38,080 --> 01:03:39,909
Um, if you're implementation, hang out with Lewis.

682
01:03:39,910 --> 01:03:45,820
If you're realist, hang out with POS mentally and um, kind of recognise that this is still an evolving space.

683
01:03:46,630 --> 01:03:53,050
Duncan also suggests to us that magazines don't just get turned on and off like layouts, but they're actually probably more like dimmer switches.

684
01:03:53,470 --> 01:03:59,140
That there's actually a continuum of mechanism that can actually exist in small ways and big ways.

685
01:03:59,140 --> 01:04:05,740
And then and the context matters in that in those instances and conditions by and all of it.

686
01:04:06,310 --> 01:04:12,850
It's a dial. Other work I did with Jay, um, some years ago, this is back when we were doing postdoctoral work.

687
01:04:13,000 --> 01:04:18,210
Yeah, I grew up in here. I know he and I chatted, but we were very, very close colleagues.

688
01:04:18,220 --> 01:04:25,600
We work together a long time. We did some work understanding kind of mechanisms or can milestones, as we call them, advancing integrated care.

689
01:04:25,990 --> 01:04:27,760
And because we looked at it over time,

690
01:04:28,060 --> 01:04:34,030
what we are finding was that something that was a mechanism at one point in time actually became a context at another point in time.

691
01:04:34,540 --> 01:04:43,029
So mechanisms actually have like a time limit, or they shift over time in the way they, you know, actually have their action shifts over time.

692
01:04:43,030 --> 01:04:50,139
So you have to consider the time component. Here's the study I was mentioning before about how you could layer on different

693
01:04:50,140 --> 01:04:54,190
kinds of mechanisms that are cognitive and institutional and process oriented.

694
01:04:54,670 --> 01:05:01,220
This is of course in all paper. This is them just looking at, um, uh, spread of an intervention.

695
01:05:01,240 --> 01:05:05,350
They looked at nutrition programs, just seven that that spread.

696
01:05:05,590 --> 01:05:09,670
What you see in the middle is what the tests are important for scale and spread.

697
01:05:10,210 --> 01:05:14,080
And then these are all the mechanisms around those seven things that mattered.

698
01:05:14,440 --> 01:05:18,580
Um, in order to make Scanlon spread happen, you can make it a little deeper.

699
01:05:18,610 --> 01:05:21,400
It's a very cool paper. Very. It's like a lot.

700
01:05:21,730 --> 01:05:28,300
But it does tell you kind of what how these things relate to each other and how they kind of come to some of those bigger order.

701
01:05:28,660 --> 01:05:33,940
Often our guidance sits at this really high abstract level, and you don't know why it matters.

702
01:05:34,210 --> 01:05:39,670
This is a nice paper that tells you kind of why it matters. And they use multiple theoretical lenses to help us unpack that.

703
01:05:41,500 --> 01:05:48,910
Um, a last piece is that there are often invisible and consequently very hard to measure.

704
01:05:49,210 --> 01:05:52,480
Um, this is a nice paper by Hull and Company.

705
01:05:52,810 --> 01:05:57,010
Um, where they were looking at mental models and implementation science.

706
01:05:57,310 --> 01:06:03,990
Mental models are very hard to uncover. We did some shared mental models work, and it's a lot to uncover in in this very deep enough idealogical.

707
01:06:04,570 --> 01:06:08,580
Um, but they were finding that it's really critical and implementation science.

708
01:06:08,580 --> 01:06:12,430
So we need to start thinking differently about how we unpack mental models.

709
01:06:12,440 --> 01:06:19,600
Well. So to summarise, cosmic uh, determinism does help describe terrain.

710
01:06:19,850 --> 01:06:24,440
Our strategies help describe action, but our mechanisms explain movement.

711
01:06:24,500 --> 01:06:30,959
Like my kids running around the beach in Belgium just as we connect. The three things you can think about going forward are.

712
01:06:30,960 --> 01:06:35,280
Implementation science might be at an inflection point, or we need to pay more attention to this.

713
01:06:35,910 --> 01:06:40,110
Mechanisms inform strategy. Design can be essential for system level reform.

714
01:06:40,110 --> 01:06:46,710
And finally, studying mechanisms requires theoretical pluralism and possibly methodological innovation.

715
01:06:47,750 --> 01:06:49,100
Reminder. Thank you.