1
00:00:00,330 --> 00:00:04,410
Welcome, everybody. It's interesting. We're going to talk about the news.

2
00:00:05,100 --> 00:00:10,950
This gentleman here actually has a newspaper. Many people don't know what one of them is, actually, I'm pretty sure.

3
00:00:11,670 --> 00:00:18,750
So I'm going to talk about I do a lot of work in evidence based medicine and I do a lot of work with the media.

4
00:00:18,930 --> 00:00:23,100
And I'm going to talk about why you all should do a bit more yourself.

5
00:00:23,370 --> 00:00:30,720
So this is all aimed at you to get involved. Okay. So first thing is, there's a lot of out there.

6
00:00:32,190 --> 00:00:35,970
These are the views per month individual three shooting fans and four.

7
00:00:36,180 --> 00:00:44,130
But 80 million people, the number one paper in the UK, The Daily Mail, followed by the Sun, The Mirror and down here.

8
00:00:44,640 --> 00:00:50,010
And then you get to the express. And in the middle is all the for the Sunday Times and in pen for lots of news.

9
00:00:50,010 --> 00:00:52,560
Our consumption of news is actually going up.

10
00:00:53,490 --> 00:01:01,110
To talk about how it was 20 years ago with much more print and with the advent of computers and their mobile phones,

11
00:01:01,110 --> 00:01:11,399
how we consuming more news than ever. But my some of my favourite paper paper if I start with the Daily Mail and the Daily Mail's a

12
00:01:11,400 --> 00:01:19,410
great paper for health care headline loads of really interesting stories about processed meat,

13
00:01:19,440 --> 00:01:26,100
heavy web use harms a child's health poisoned by everyday life from bad, lethal cigarette.

14
00:01:26,940 --> 00:01:32,040
Now, it's interesting to note, why does this appear and why does it keep happening?

15
00:01:32,550 --> 00:01:42,720
It's incredibly prevalent to have health care news headlines is the editors know that you will click on these headlines.

16
00:01:43,770 --> 00:01:48,690
So you are part of the problem in that you know what it's like.

17
00:01:48,690 --> 00:01:53,040
You know, you know this is fake, but you can't wait to get in there and find out.

18
00:01:53,070 --> 00:01:59,400
So on the portable on the phone, we all go in and this increases in go zone.

19
00:02:00,030 --> 00:02:05,450
And even in Oxford we contribute to some of these these stories that obesity in women is dangerous.

20
00:02:05,460 --> 00:02:12,810
The terror threat and this is an Oxford study don't drink more than three glasses of wine a week.

21
00:02:13,080 --> 00:02:16,860
And so what's also interesting is how do these stories get interpreted?

22
00:02:17,850 --> 00:02:21,120
And I'm very fascinated by health care headlines.

23
00:02:21,810 --> 00:02:29,970
And when I first did this talk, I even sent out on Twitter a request for people for their favourite headline.

24
00:02:31,320 --> 00:02:42,780
And this is David who's running this week, his David favourite health care headline and so together.

25
00:02:43,140 --> 00:02:51,820
So David has an interest in exercise and lifestyle and this is a sort of evidence that he spends his time looking up about this.

26
00:02:54,660 --> 00:03:01,110
But, you know, it's interesting, isn't it, as we look at this stuff, you think, well, how does this impact on the decisions we make?

27
00:03:03,120 --> 00:03:06,480
And as I've shown you, he's my favourite.

28
00:03:09,090 --> 00:03:12,420
Your fridge is a major killer. You just don't know it.

29
00:03:13,380 --> 00:03:17,640
But here's my other favourite paper in the whole world. The Daily Express.

30
00:03:18,270 --> 00:03:21,510
The Daily Express definitely trumps the Daily Mail,

31
00:03:21,960 --> 00:03:28,680
and I'll show you what I'm going to come to a headline, which I know, but this is just about dementia.

32
00:03:28,680 --> 00:03:35,700
I started collecting news stories about dementia from the Daily Express, thinking, What could I learn?

33
00:03:36,430 --> 00:03:40,590
Simple way to fight off dementia. Actually, it's a stressful lifestyle.

34
00:03:41,190 --> 00:03:46,800
And then I started to think, Well, this is a bit odd. Chocolate can halt dementia, but lose weight to beat dementia.

35
00:03:47,250 --> 00:03:51,330
So how can anybody know you can eat lots of chocolate and lose weight?

36
00:03:51,810 --> 00:03:55,629
It's quite an interesting issue. You know, actually, you don't eat anything else,

37
00:03:55,630 --> 00:04:01,830
so you're going to go on the chocolate diet to stave off dementia and it keeps going daily drink aspirin,

38
00:04:02,700 --> 00:04:06,900
new way to fight dementia and even red wine pills stopped dementia.

39
00:04:07,800 --> 00:04:11,040
And so I was going I was getting really excited. You know, what you're like is a researcher.

40
00:04:11,430 --> 00:04:14,550
You like collecting data, don't you? We like to collect things.

41
00:04:14,880 --> 00:04:22,230
But what happened is this started to get a bit of a series addiction for me in that they started to increase it increase.

42
00:04:22,560 --> 00:04:26,700
Now the qualitative research is in love.

43
00:04:26,820 --> 00:04:31,920
This I thought, well, actually what we'll do is we'll take the headlines and try and create some themes.

44
00:04:32,490 --> 00:04:34,140
So I started putting them into theme.

45
00:04:34,150 --> 00:04:43,139
So here there's something about healthy living over here in lifestyle, something about drugs and chocolate, simple ways and to fight dementia.

46
00:04:43,140 --> 00:04:46,170
It's somewhat about tests and spicy diet.

47
00:04:46,170 --> 00:04:49,770
So started to organise them into is it a female tick analysis?

48
00:04:50,040 --> 00:04:56,400
Is that fair? What would you call that? Is that it's the beginning the beginning of being a qualitative.

49
00:04:56,580 --> 00:04:59,610
And when it did that I noticed is started to be able to fill in the.

50
00:04:59,940 --> 00:05:09,270
Even more and ended up like this with these eight different ways of looking at how we're going to beat dementia so you can lose weight,

51
00:05:09,960 --> 00:05:17,400
eat chocolate, keep busy, and have a healthy life. You can do puzzles, take some drugs and have some tests and watch your diet.

52
00:05:18,030 --> 00:05:23,550
And the Daily Express tells you how to do that. But, you know, this is all [INAUDIBLE].

53
00:05:25,260 --> 00:05:29,460
When you look at all of them and you know this is what's going to happen in ten years time.

54
00:05:30,510 --> 00:05:34,360
And that's the issue here, isn't it? You start to go and this is how this is mine.

55
00:05:34,380 --> 00:05:46,740
Anybody can add in another one. There are loads of really interesting headlines that we're all, all the time looking at and pressing.

56
00:05:47,430 --> 00:05:55,170
Let's have a look. And I know that because the editor of The Express, his favourite topic is statins.

57
00:05:57,180 --> 00:06:04,500
His wife took a statin and his wife then got some muscle pains and decided to stop taking the statin.

58
00:06:05,010 --> 00:06:06,960
And so he became obsessed by statins.

59
00:06:08,100 --> 00:06:15,210
And what happened in doing that is a I think about tripled their readership when they put a story about such things on the front page.

60
00:06:15,660 --> 00:06:21,030
So they're always looking for statins, but what interested me is the circularity of all this.

61
00:06:22,170 --> 00:06:30,660
So this is the one headline Take Statins to Save Your Life, and then you go, Oh, let's put that in a health alert.

62
00:06:31,410 --> 00:06:37,560
Interesting. And part of that is the next one is it causes diabetes.

63
00:06:39,030 --> 00:06:42,090
However, in causing diabetes, we're going in that direction.

64
00:06:42,480 --> 00:06:43,860
It's going to cause chaos.

65
00:06:45,330 --> 00:06:52,799
But in causing chaos, we back to prove statins save millions and we come back in a circularity again and we'll get the whole thing.

66
00:06:52,800 --> 00:06:55,590
And this is what happens in about a year.

67
00:06:56,550 --> 00:07:04,770
And these news headlines keep getting recycled in a way if a piece of evidence emerges that actually fits this narrative.

68
00:07:06,810 --> 00:07:12,000
And you can keep going. You can upload the form with the express. You can do cancer, you can do whatever you want in headlines.

69
00:07:12,810 --> 00:07:19,850
And some of these are verging on the ridiculous drug to cure.

70
00:07:19,860 --> 00:07:28,200
Nine in ten cancers, rhubarb can save your life and they will feel a bit like fun until you meet the first patient who says,

71
00:07:28,590 --> 00:07:32,010
I stopped taking treatment and now I'm taking rhubarb instead.

72
00:07:33,150 --> 00:07:42,870
And I was describing a patient I've seen who basically came off his chemotherapy because his real life case started to take coffee enemas instead.

73
00:07:44,040 --> 00:07:47,879
And that will not anybody in medical care at some point is involved,

74
00:07:47,880 --> 00:07:54,030
will find somebody who goes on on one of these news headlines to take an alternative path.

75
00:07:55,170 --> 00:07:59,100
And I can come back to what I think issues usable. So it's interesting.

76
00:07:59,110 --> 00:08:05,420
So Google's official blog, about 1% of all searches on the Internet now.

77
00:08:07,140 --> 00:08:14,220
And that's you know, you don't have to think that's millions. If not, God knows how many that are about people just looking at symptoms.

78
00:08:16,380 --> 00:08:24,420
And that includes you. When you have a symptom, you go, ooh, I'm going to look it up and see if there's a miracle cure or what's going on.

79
00:08:24,780 --> 00:08:32,010
But interestingly, they sort of started to take a position to start to think about their role in health care.

80
00:08:33,450 --> 00:08:40,620
Now, you think it's because they're going to do health care for they want to do the right thing.

81
00:08:40,860 --> 00:08:47,670
They've realised like the Daily Express, this is a business and in doing it better you'll come to Google more often.

82
00:08:48,450 --> 00:08:56,370
But again, the evidence of the self-check is and there is this paper which is in the BMJ, that shows when you look at the symptom checkers,

83
00:08:56,370 --> 00:09:06,810
there were 23 symptom checkers provided the correct diagnosis was only available a third of the time in these online forums.

84
00:09:07,830 --> 00:09:11,010
And you start to think, what's going on here?

85
00:09:11,670 --> 00:09:17,159
If you think about the drivers of health care, when is this appropriate for people to check something and go,

86
00:09:17,160 --> 00:09:19,020
Well, that means I need to go and see my doctor.

87
00:09:19,080 --> 00:09:27,150
Or actually I can stay at home interested in its own right, the potential to carry off more work or inappropriate work.

88
00:09:29,310 --> 00:09:33,750
So I'm going to take you back in time. Here's my guilt again. Those in the MFC told them this here.

89
00:09:34,050 --> 00:09:38,610
This is going back to 2009 and 2009.

90
00:09:39,510 --> 00:09:43,470
There are move people when we're all going to die. Okay.

91
00:09:44,970 --> 00:09:49,740
Remember that Daily Express, we're on alert for the new super flu.

92
00:09:52,170 --> 00:09:59,820
Don't know. It's not about Jonathan Ruffin. It is swine flu already here now the battle to contain.

93
00:09:59,940 --> 00:10:04,680
It starts, then it gets really serious.

94
00:10:04,710 --> 00:10:09,870
Now, you really interested in the news? Because it's going to kill 350 people every day.

95
00:10:11,160 --> 00:10:19,500
Yeah. We are in panic mode and you can't go on your holidays.

96
00:10:20,170 --> 00:10:23,670
It's just like when you think about Brexit, you won't be able to go on holidays.

97
00:10:23,670 --> 00:10:29,549
You couldn't go the holidays in 2009. What was the problem? And then if you really think about it, it got even worse.

98
00:10:29,550 --> 00:10:33,450
When are you going to go? We're all going to die. It's interesting.

99
00:10:33,570 --> 00:10:35,219
So what's happening now?

100
00:10:35,220 --> 00:10:45,150
Interestingly, some of that was based on fake and some of it's based on real experts and what we do now going back to the swine flu,

101
00:10:45,150 --> 00:10:54,870
remember, some of you should now be dead. Donald Rumsfeld, remember the known unknowns and unknowns, whatever you know, you don't know.

102
00:10:55,290 --> 00:10:59,549
Well, he actually this is an article in The Independent.

103
00:10:59,550 --> 00:11:06,120
He actually had shares in a company called Gilead who had the intellectual property to Tamiflu.

104
00:11:07,110 --> 00:11:16,530
And he this is an article about him saying he ordered this UN report into how many predicted death would bird flu cause now.

105
00:11:17,870 --> 00:11:23,560
You'd like to have a go? A number? How many deaths do you think would be by bird flu?

106
00:11:25,090 --> 00:11:32,760
You can go either UK based or worldwide. 1020.

107
00:11:33,950 --> 00:11:38,550
A couple of hundred. A couple of thousand. You're going to go 5000.

108
00:11:39,030 --> 00:11:44,310
That feels pretty high, doesn't it? And would you like to go higher or lower?

109
00:11:45,840 --> 00:11:51,060
Because this is important. If you're going to think about stockpiling drugs and thinking about we're all going to die.

110
00:11:51,330 --> 00:11:55,010
How many people in a UN report were predicted to die?

111
00:11:55,010 --> 00:12:01,410
We have 5000. How many of you think lower? And how many of you think higher?

112
00:12:02,220 --> 00:12:07,460
Oh, you're all a bit higher. How many lady there with a green scarf?

113
00:12:07,470 --> 00:12:11,160
170 million. 170 million.

114
00:12:11,160 --> 00:12:14,990
Bloody [INAUDIBLE], man. I knew it.

115
00:12:15,900 --> 00:12:23,730
Not too bad, actually. In fact, we were some of the 700,000 deaths that might be expected in the UK alone.

116
00:12:26,430 --> 00:12:29,790
So it was 700,000 deaths worldwide were predicted.

117
00:12:30,600 --> 00:12:33,660
So you actually although it sounds laughable, you weren't that far out.

118
00:12:34,420 --> 00:12:42,840
Yeah. So we predicted so. No wonder everybody goes into panic mode, but this stuff really gets into our psyche and have impact on policy.

119
00:12:43,110 --> 00:12:46,820
So you can calculate what's called the news to death ratio.

120
00:12:46,830 --> 00:12:57,570
So in the first 24th of April, 2009, in the first 15 days of the swine flu outbreak, you can look at the number of deaths in terms of swine flu.

121
00:12:57,990 --> 00:13:07,470
So there were 31 death for swine flu in the same 13 days for TB, there were 63,066 death.

122
00:13:08,970 --> 00:13:15,510
The number of news articles were for swine flu were two five 250,036 and 8000 for TB.

123
00:13:16,530 --> 00:13:23,070
So that gives you a news to death ratio of 8176 news articles for one death.

124
00:13:23,760 --> 00:13:27,330
But actually, every time you see a news article on TB, there'll be ten death.

125
00:13:29,490 --> 00:13:33,360
Now, that was done by a guy called Hans Rosling.

126
00:13:33,360 --> 00:13:39,960
I pinch that often, but it's got a very important message in the distortion about how the news comes at you,

127
00:13:40,320 --> 00:13:47,340
because this is a major killer and it's really important in terms of its health impact.

128
00:13:47,850 --> 00:13:53,089
But nobody gives. A monkey study about it.

129
00:13:53,090 --> 00:13:58,250
Nobody cares about it in any way. And I think all of us go into this area and think where the excitement.

130
00:13:58,430 --> 00:14:02,510
But this is a huge killer. Now, we published the Cochrane Review in this area,

131
00:14:02,720 --> 00:14:10,390
got the unpublished data and did lots of fun and published all sorts of stuff that says, I wouldn't give this stone to 12.

132
00:14:10,400 --> 00:14:11,560
It's got huge problems.

133
00:14:11,570 --> 00:14:22,670
We published all of the research and I thought we would change practice and we ended up in 2014 £550 million we spent on stockpiling this drug.

134
00:14:23,120 --> 00:14:29,810
And it's a waste of time. The reason it's a waste of time, it does have a symptomatic effect, a bit like paracetamol.

135
00:14:30,020 --> 00:14:34,520
The evidence shows it reduces your temperature, reduces some of the symptoms of flu,

136
00:14:35,180 --> 00:14:41,180
but it doesn't reduce the important complications of flu like bacterial infections,

137
00:14:41,480 --> 00:14:45,710
hospital admissions, all of the things you would want it to do if you were going to stockpile it.

138
00:14:47,240 --> 00:14:50,750
It doesn't actually prevent you from getting fully when you look in the back of the evidence.

139
00:14:51,020 --> 00:14:59,599
I went to a parliamentary committee in 2014 and this is when it hit home and we spent about 2 hours laying all this out and leaving that committee,

140
00:14:59,600 --> 00:15:02,270
thinking all of the people in the government get this issue.

141
00:15:02,840 --> 00:15:12,170
And a week later, we spent £50 million more stockpiling it because the stockpile we had was going out of date and they said, we can have it cheap.

142
00:15:12,560 --> 00:15:14,780
You can have it at the half price if you renew your stockpile.

143
00:15:16,010 --> 00:15:23,209
So I also think there's a huge agenda with our policy people thinking about how we inform policy in an evidence based way.

144
00:15:23,210 --> 00:15:31,610
That's part of our job in research is to think about how do we arm the next generation of policy people to look at evidence.

145
00:15:33,380 --> 00:15:40,880
So some people with money will have been on the MSE course and seen this graph where there are thoughtful steps to evidence based practice.

146
00:15:40,910 --> 00:15:46,130
Ask a question, acquire some evidence, appraise the quality and apply it in practice.

147
00:15:46,880 --> 00:15:53,090
Now, when I'm thinking about applying this to the media and thinking about how we go about a

148
00:15:53,090 --> 00:15:58,760
process of informing the news and what we do in health care is finally flip that round.

149
00:15:58,760 --> 00:16:04,700
And I use this in a different way. I have an extra step, so I focus.

150
00:16:06,230 --> 00:16:10,250
You can see instead of asking questions here, this is an extra step that's going in.

151
00:16:11,000 --> 00:16:16,580
So I think about acquiring the best evidence and assessing the effect all the time.

152
00:16:16,790 --> 00:16:24,380
When I get asked about media questions or healthcare evidence, I think, is this the best evidence and what's the size of the effect?

153
00:16:25,640 --> 00:16:30,080
This step is deliberately smaller because we never get to the biases,

154
00:16:30,500 --> 00:16:35,630
often because by the time you're in the biases, most people don't understand what we're on about.

155
00:16:37,040 --> 00:16:44,149
If I'm having a discussion in a newspaper about confounding and indication, bias confounding by indication,

156
00:16:44,150 --> 00:16:47,720
and the people going, We're not interested, nobody's going to understand this.

157
00:16:48,020 --> 00:16:51,499
So I focus on these two areas and I'll show you how I do it.

158
00:16:51,500 --> 00:16:55,820
So the first step is I always think about is when I see some evidence,

159
00:16:56,510 --> 00:17:01,430
I ask a general question, Is this the highest level of evidence to answer this question?

160
00:17:03,020 --> 00:17:12,860
That's my first step. And often what you'll see is, for instance, within the sort of process of evidence, you'll see lots of news headlines.

161
00:17:13,370 --> 00:17:20,749
But down here it might actually be just an animal study, but if it's an animal study,

162
00:17:20,750 --> 00:17:25,610
it'll take 20 years to get into practice and only about one in ten of them will get into practice.

163
00:17:25,880 --> 00:17:31,670
So when they say we've got a new cure for dementia and it's in mice, you're thinking, well, one in ten chance in 20 years.

164
00:17:32,210 --> 00:17:36,710
There's a quote where if what you're thinking is, where are we in the evidence hierarchy?

165
00:17:37,070 --> 00:17:42,920
And to do that and inform that you generally want to be hierarchy, don't you want to be open to the systematic review?

166
00:17:42,920 --> 00:17:49,670
And and to do that, you need a very good understanding of this one table, which is the levels of evidence.

167
00:17:51,050 --> 00:17:55,490
What type of question is it and where are we in this matrix?

168
00:17:55,790 --> 00:18:04,790
And if we down here, we know we've got some issues where if I want to be at the higher end of the evidence, to think about the systematic reviews.

169
00:18:05,210 --> 00:18:13,640
So I'll give you an example of how we applied that. We did this for evidence for what's called IVF add ons.

170
00:18:14,990 --> 00:18:23,390
We was approached by the BBC who said We're going to do a story about all of these add on treatments that you can get in addition to IVF.

171
00:18:24,530 --> 00:18:26,269
So we did two things.

172
00:18:26,270 --> 00:18:36,980
The first thing is we said, Well, let's look at the websites and look at what the claims are and how many times they actually when they stake a claim,

173
00:18:37,610 --> 00:18:45,080
they cite a bit of evidence and then ask a question is, is it the highest level of evidence?

174
00:18:45,680 --> 00:18:50,240
Is it actually a systematic review? Because if you were having a new treatment, you would.

175
00:18:50,290 --> 00:18:57,070
I want to be saying, is backed up by a systematic review. And in doing that, we showed that actually they were all miserable.

176
00:18:58,420 --> 00:19:01,569
That actually they often on the websites didn't find any evidence.

177
00:19:01,570 --> 00:19:05,230
And when they did, they actually went to lower quality or opinions.

178
00:19:06,190 --> 00:19:15,249
And when we went searching for the evidence, we could only find evidence five times where there was some evidence backed up by systematic reviews,

179
00:19:15,250 --> 00:19:20,020
where you think there might be an effect and then discussed the effect sizes.

180
00:19:20,290 --> 00:19:24,009
Now, in just doing that very simple bit,

181
00:19:24,010 --> 00:19:30,639
we got to a point where most of the Royal College of OBS in Guinea and there are some people in here went mad at us and said,

182
00:19:30,640 --> 00:19:35,980
you have no idea what you're on about. Well, remember, my job is I'm a GP, I am a GP.

183
00:19:35,980 --> 00:19:40,840
So my job is to try and inform people. If they come in and say, I'm thinking of having some IVF.

184
00:19:41,140 --> 00:19:47,050
What do you think about embryo glue? Well, I can tell you moved GP's have no idea what to do in that situation.

185
00:19:47,320 --> 00:19:52,420
So there has to be some way of saying where's the evidence? And said, what do you think you're doing?

186
00:19:52,420 --> 00:19:56,470
Explaining what the evidence is when you're not an obstetrician or a gynaecologist.

187
00:19:57,160 --> 00:20:03,850
When I said I thought, That's what we're trying to do. And all we do is say, What's the evidence in terms of its level?

188
00:20:04,900 --> 00:20:12,060
I'm not even gonna into the effect sizes. Two years later when you think about different this has been really interesting.

189
00:20:12,070 --> 00:20:20,860
So for about two years went round in circles and then early at the end of just the end of last year, the FDA,

190
00:20:20,860 --> 00:20:26,829
who the Human Fertilisation and Embryo Authority did a whole review where they brought

191
00:20:26,830 --> 00:20:30,790
together a system which I think is a huge step forward based on the research.

192
00:20:31,300 --> 00:20:35,230
And what they did, they said here we're going to provide you with the evidence.

193
00:20:35,230 --> 00:20:45,280
But they did some really smart, which I quite like. They have a traffic light system which is their traffic light system is red, amber, green.

194
00:20:46,240 --> 00:20:49,450
And the red there is this is assisted hatching. It talks.

195
00:20:49,450 --> 00:20:53,139
You are the rest. What's the evidence? And it says there is no evidence.

196
00:20:53,140 --> 00:20:56,980
You can't see that you need know that this ad is effective and safe.

197
00:20:57,650 --> 00:20:58,420
That's a red light.

198
00:20:58,420 --> 00:21:06,579
When they have a green, they say there is evidence from at least two randomised controlled trials and in the amber they go it's uncertain.

199
00:21:06,580 --> 00:21:11,770
The evidence is conflicting and it needs reducing in certain and I quite like

200
00:21:11,770 --> 00:21:16,210
that system because the traffic light system helps you really simply think,

201
00:21:16,450 --> 00:21:19,480
okay, what's the evidence? What's it like? Is there any evidence?

202
00:21:19,660 --> 00:21:24,250
If the answer is no, you've got a red. If there's more than two trials, at least you're on the pathway to somewhere.

203
00:21:24,670 --> 00:21:29,590
And you can start to filter out a lot of evidence really simply by just asking

204
00:21:29,590 --> 00:21:33,130
what type of evidence is it and is it appropriate to answer this question?

205
00:21:34,720 --> 00:21:44,010
Now the second step then is assessing the effect, and that's difficult because you're really thinking about patient benefit.

206
00:21:44,830 --> 00:21:55,450
So I often think about this is an equation in my mind that I think about, and each patient benefit is directly proportional to the outcome.

207
00:21:56,350 --> 00:22:06,130
The bigger the outcome, the more benefit, the better the outcome, the more benefit, the more relevant the outcome, the more patient benefit.

208
00:22:07,240 --> 00:22:11,950
So if I say you get a two year reduction in mortality, you might think, well, that's an important outcome.

209
00:22:12,760 --> 00:22:18,400
However, it's inversely proportional to the bias. The more bias, the less patient benefit.

210
00:22:19,750 --> 00:22:23,500
And if you introduce biases, you're overcoming the patient benefit.

211
00:22:24,160 --> 00:22:32,050
Now, that's really important in your thinking because if there's only a very small outcome benefit, any small bias will overcome the patient benefit.

212
00:22:32,890 --> 00:22:36,280
It is a rather large outcome and a big outcome.

213
00:22:36,280 --> 00:22:39,880
You might accept more bias. So it's a simple way.

214
00:22:40,060 --> 00:22:43,110
And then the third thing is you need the optimal information size.

215
00:22:43,120 --> 00:22:50,140
That's what systematic reviews are about, having sufficient information to be able to say with some certainty.

216
00:22:50,170 --> 00:22:51,340
This is a true effect.

217
00:22:52,150 --> 00:23:00,310
It's not the big data, you know, not it's not all the information because people get overwhelmed with we've got 3 million patient years.

218
00:23:00,790 --> 00:23:04,060
What you're interested in, have I got the right amount of information?

219
00:23:05,620 --> 00:23:08,440
And so one of the things to think about is to start with,

220
00:23:08,710 --> 00:23:18,970
I optimise the outcome and this is my second step and I ask a very simple question all the time Does this make a difference?

221
00:23:20,470 --> 00:23:23,920
And if it doesn't make a difference, why do you even need to care about the biases?

222
00:23:24,190 --> 00:23:26,860
Why do you even need to care about whether you've got optimal information?

223
00:23:27,310 --> 00:23:33,160
Why not just stop and you can stop and filter out a lot of information really quickly?

224
00:23:33,580 --> 00:23:38,560
Now is not quite there are there are bits of to does it make a difference?

225
00:23:39,060 --> 00:23:44,860
Oh yeah. And I think about three questions to that when I'm doing that.

226
00:23:45,190 --> 00:23:49,630
Yeah. And is assessing the effect I have these, these three questions.

227
00:23:49,840 --> 00:23:54,620
Who does. It applies to. Does it apply to the actual people we see or the patients we see?

228
00:23:55,280 --> 00:23:58,489
Does it make an actual difference? Both positive and negative.

229
00:23:58,490 --> 00:24:01,580
The benefits and the harms. And is it actually feasible?

230
00:24:03,090 --> 00:24:09,720
Yeah. And so you have to think of the three questions I'm often putting in at this step.

231
00:24:11,550 --> 00:24:14,610
So if we come back to the IVF example.

232
00:24:15,300 --> 00:24:22,140
Yeah. The first thing you would expect to happen is that anything that makes a claim for an intervention will be backed up by some evidence.

233
00:24:22,380 --> 00:24:26,340
Some of these treatments are of no benefit whatsoever and some of them are harmful.

234
00:24:26,730 --> 00:24:31,590
And that's going into the state. But it's going to. For instance, which ones?

235
00:24:31,590 --> 00:24:36,569
And only one treatment called an endometrial scratch was supported by moderate quality evidence.

236
00:24:36,570 --> 00:24:41,840
It would help it involve the procedure to scratch the womb lining to help an embryo successfully implant,

237
00:24:41,850 --> 00:24:44,820
although the evidence for this treatment was not itself beyond doubt.

238
00:24:45,090 --> 00:24:49,320
In fact, what's happened is this was the only one that was a moderate evidence of benefit.

239
00:24:49,530 --> 00:24:54,240
A large randomised controlled trial has come out in the last month that says it's of no benefit.

240
00:24:55,170 --> 00:24:59,670
So we often get this early stage where things look promising, but as you go along, nobody.

241
00:25:00,030 --> 00:25:05,339
So you get by asking, here's another one. I'm thinking about the outcome.

242
00:25:05,340 --> 00:25:10,710
Does it make a difference? So this is an interesting paper came out in the BMJ.

243
00:25:11,490 --> 00:25:15,780
I got us to quote on this so I don't know anybody here.

244
00:25:15,780 --> 00:25:23,490
It quoted for the media part. When you're on the Science Media Centre list, anybody else get asked to make a quote?

245
00:25:23,730 --> 00:25:26,820
Yeah. Do you actually quote? Yes, sometimes.

246
00:25:27,240 --> 00:25:34,140
Is it difficult or do you do it really quickly and integrative? I spent that time and you say, yeah, yeah, yeah, yeah.

247
00:25:34,320 --> 00:25:37,920
So it feels like sometimes it's going to take you forever. And that's why we don't participate.

248
00:25:38,610 --> 00:25:44,190
This is a real I'm going to be quoted in the news. I'm going to spend half a day read the evidence.

249
00:25:44,520 --> 00:25:47,580
Oh, my God. In fact, I haven't got a time to do that any more.

250
00:25:47,940 --> 00:25:54,930
So I actually think, oh, this is really interesting. I'll have a quick read while I'm having a cup of tea and to think about my framework.

251
00:25:55,230 --> 00:26:01,980
Okay. And I think I'm going to quote on outcome and here's here's me being quoted.

252
00:26:02,490 --> 00:26:07,020
Here's what I actually said and that's what I said. I wonder if these make a difference.

253
00:26:07,180 --> 00:26:09,540
And I wrote a quote back to the science media saying,

254
00:26:09,720 --> 00:26:15,330
It's hard to understand why half the drugs were approved in the first place if they provide no clinical benefit, he said.

255
00:26:15,750 --> 00:26:16,560
And that's it, Don.

256
00:26:17,610 --> 00:26:25,739
That took me about 5 minutes reading it because they're going with a very structured approach to what I'm looking for in the evidence,

257
00:26:25,740 --> 00:26:29,280
as opposed to just reading it and hoping some amazing quote will emerge.

258
00:26:29,700 --> 00:26:35,790
I go in and say, Does this make a difference today? And as I said to you, if it doesn't make a difference, why would you carry on?

259
00:26:36,540 --> 00:26:43,679
That's my quote. So there's this sort of approach to benefit.

260
00:26:43,680 --> 00:26:47,069
And I think if you think about outcomes, we haven't touched any of the other two bits.

261
00:26:47,070 --> 00:26:53,300
I've told you two things. Is this the appropriate evidence to answer this question and then these three questions?

262
00:26:53,580 --> 00:26:58,350
Does it make a difference? Does this apply to my patients? And is the treatment feasible in my setting?

263
00:26:59,520 --> 00:27:02,070
Okay. And you want to think about optimising them.

264
00:27:03,030 --> 00:27:13,470
So sometimes I get things like research and I dismiss it immediately because I'm asking this is a question about does this apply to my patient?

265
00:27:13,890 --> 00:27:15,960
And in thinking about it, I think does this apply to me?

266
00:27:16,770 --> 00:27:23,580
So effective breakfast on weight and energy intake, a systematic review and meta analysis of randomised controlled trials.

267
00:27:24,540 --> 00:27:32,850
This was a trial of whether you eat breakfast or should stop eating breakfast and how many people don't eat breakfast.

268
00:27:34,590 --> 00:27:40,230
Okay. Do you actually care about whether you should or shouldn't and whether it has an impact on your weight loss?

269
00:27:41,220 --> 00:27:47,070
Look, I don't care. I say it depends how much time I've got in the morning.

270
00:27:47,100 --> 00:27:51,959
If I if I lie in too late and I'm running late, I might not have breakfast at the weekend.

271
00:27:51,960 --> 00:27:55,980
I have breakfast. So I'm like, this doesn't even apply to the people I see.

272
00:27:55,980 --> 00:27:59,100
Why would anybody even care about the effect of breakfast?

273
00:27:59,370 --> 00:28:02,510
Nobody's ever come into my clinic and said, I talk to Hannigan.

274
00:28:02,520 --> 00:28:05,880
I'm really thinking about whether I should give up breakfast or not. What's your advice?

275
00:28:06,120 --> 00:28:08,880
I've never seen anybody, so it doesn't mean you.

276
00:28:09,150 --> 00:28:17,490
But what I'm saying to you is this is a dramatic example of is this the sort of evidence that applies to the people you see in practice?

277
00:28:18,720 --> 00:28:23,100
If it is, no, I wouldn't waste your time. I've moved on now.

278
00:28:23,100 --> 00:28:26,220
You could read it if you're interested. Okay. Oral thought for a reason.

279
00:28:27,870 --> 00:28:34,490
Now, at the moment I am feeling a little bit stressed actually at the moment because I got involved in that.

280
00:28:34,530 --> 00:28:38,790
I'm not sure I should have got involved in what I have done because it's a really evidence based issue.

281
00:28:39,240 --> 00:28:46,440
This is a piece I did a piece with Panorama, which was about gender dysphoria and the drug treatments in children.

282
00:28:47,730 --> 00:28:55,450
And these things now in the world are incredibly emotive. And I don't my my question is not whether they're right or wrong.

283
00:28:55,620 --> 00:28:58,920
The one thing I do is say, what does the evidence look like?

284
00:28:59,580 --> 00:29:02,570
Have we got sufficient evidence for somebody to make an informed judge?

285
00:29:03,090 --> 00:29:07,530
And in terms of children, you'd think, well, actually, you would want robust evidence, wouldn't you?

286
00:29:07,920 --> 00:29:10,170
These are very powerful treatments we're going to use.

287
00:29:10,590 --> 00:29:19,140
So if we're going to go down this route, this is a good example of how we should frame the whole of medicine, how we should think about what we do.

288
00:29:19,410 --> 00:29:23,670
Now, the problem is in health care is, boy, do we jump on bandwagon.

289
00:29:24,180 --> 00:29:30,720
There's a huge history of just trying treatments and going down a route because we think they can work.

290
00:29:31,920 --> 00:29:40,470
And in this article I put I said, we've got ourselves into a complete mess here because about seven or eight years into this,

291
00:29:41,010 --> 00:29:47,820
we have got in a position where we've given out these powerful drugs and now we're going more people want them,

292
00:29:48,030 --> 00:29:55,200
and now we're going, Oh, what's the evidence? And I looked at the evidence and went, Oh, my God, this is abysmal.

293
00:29:55,650 --> 00:30:00,900
It's worse than the IVF. Now I can handle with the IVF because of the problem in IVF.

294
00:30:00,900 --> 00:30:04,770
If you get it wrong, it's okay. You don't you don't have a child. That's pretty bad.

295
00:30:04,770 --> 00:30:12,480
But if you get it wrong and somebody is having really powerful drugs, wouldn't we as a society want to be thinking about the difference we make?

296
00:30:12,930 --> 00:30:16,260
Now all I do is say what the quality of the evidence.

297
00:30:16,680 --> 00:30:20,310
Now, on one side of this argument, people think I'm amazing.

298
00:30:21,540 --> 00:30:28,469
On the other side of this argument, they think I'm an absolute nightmare, the devil himself.

299
00:30:28,470 --> 00:30:38,880
And he feels like that because of the ideology. And all I'm doing is I not said anything apart from this should be done in a research context.

300
00:30:39,780 --> 00:30:44,940
If we're going to go down this route, we have to develop the research evidence and we have to do it in a really robust way.

301
00:30:45,150 --> 00:30:48,150
And there are so many unanswered questions and so many uncertainties.

302
00:30:48,420 --> 00:30:51,870
We have to have a different approach and at the moment we seem to be going, it's alright,

303
00:30:51,960 --> 00:30:56,700
nobody wants to go there because we're all a bit scared of the stuff we'll get happening.

304
00:30:56,700 --> 00:31:01,380
But if we can't do it for this in children, what are we going to do in the whole of society?

305
00:31:01,650 --> 00:31:05,610
And I think these are really interesting issues we now face.

306
00:31:06,690 --> 00:31:13,050
We are bringing treatments and innovations and new devices and technologies in a way that we've never seen before.

307
00:31:13,380 --> 00:31:20,460
Innovation is the word on the street. And in doing that, we want to ignore the evidence, apart from when it suits us.

308
00:31:21,750 --> 00:31:29,640
So one thing is the debate now I've learned is when I'm going to go somewhere controversial, I always write a piece to go with it.

309
00:31:30,510 --> 00:31:35,770
So I've had some pieces today that have gone You've said this, and I go, I didn't say that.

310
00:31:36,600 --> 00:31:40,139
And so instead of trying to answer it back ago, here's what I said.

311
00:31:40,140 --> 00:31:48,060
If you go here, you can read my review of the evidence and what my conclusions are at the end of the article if you disagree with them.

312
00:31:48,060 --> 00:31:56,220
So some people right now disagree with systematic reviews are an important way of informing the public.

313
00:31:56,790 --> 00:32:01,740
They think they're terrible. And I go, Well, okay, that's alright. But we use systematic reviews and here they are.

314
00:32:01,980 --> 00:32:07,320
And in there I took the individual studies and said, What does it look like in terms of the effects and the quality?

315
00:32:08,400 --> 00:32:10,200
And I think this is something I've learnt.

316
00:32:10,200 --> 00:32:15,779
If you're going to join the media and have a go and say things, you should always have a backup where you go,

317
00:32:15,780 --> 00:32:20,909
actually here's my viewpoint here are published in some blog or some place where actually this

318
00:32:20,910 --> 00:32:26,910
is what I said and that helps you a lot because it gives you a if somebody misquoted you.

319
00:32:26,910 --> 00:32:32,879
And I think that's what everybody's worried about. You join in and suddenly somebody will translate something.

320
00:32:32,880 --> 00:32:35,880
You've said something into something so real or something.

321
00:32:35,880 --> 00:32:38,730
You go, I didn't say that. And that's what happens.

322
00:32:40,020 --> 00:32:48,470
So if I'm thinking I'm just going to end up now and then we can have a few questions is there's a book,

323
00:32:48,480 --> 00:32:57,299
there's a whole new world out there, and that world is Facebook and social media and bypasses.

324
00:32:57,300 --> 00:33:03,150
So the newspapers have sometimes put really good stuff up there, sometimes, as you've seen shocking stuff up there.

325
00:33:03,660 --> 00:33:06,840
But a lot of the time they're sort of in the middle.

326
00:33:07,050 --> 00:33:15,959
But this type of stuff has a huge impact because a large majority of health news shared on Facebook is fake, is fake or misleading.

327
00:33:15,960 --> 00:33:20,520
And that's what everybody is in their ideology sharing information.

328
00:33:20,940 --> 00:33:28,410
So the question is in health care is what's the role of researchers in informing the public?

329
00:33:29,010 --> 00:33:33,150
And if you look last year in the ten most shared health articles on social media,

330
00:33:34,110 --> 00:33:43,259
his his and this was in a way of just a simple measure of trying to say most of them are low are very low in terms of scientific credibility.

331
00:33:43,260 --> 00:33:46,980
The three that weren't here were six, seven and eight.

332
00:33:47,400 --> 00:33:52,740
And six is about cold or flu symptoms is how to tell the difference.

333
00:33:53,820 --> 00:34:00,690
One about stem cell treatment and the other was about how cycling in old age can keep your immune system young shed a lot.

334
00:34:00,690 --> 00:34:05,770
So we all like cycling in old age. The other ones are this one fine.

335
00:34:06,790 --> 00:34:11,980
Marijuana is 100 times less toxic than alcohol and it's safer than tobacco.

336
00:34:12,550 --> 00:34:15,130
That's a move shared story on social media last year.

337
00:34:15,610 --> 00:34:21,850
That's because all the people, I guess, who want to take their marijuana quite like that message and want to pass it on to everybody else.

338
00:34:22,570 --> 00:34:32,530
So it's a really interesting arena in terms of what our role as researchers in informing the public and how might you go about that.

339
00:34:33,130 --> 00:34:42,460
And I think we do have a huge role to play. And this is Fiona Fox, who runs the Science Media Centre, who do a great job.

340
00:34:42,970 --> 00:34:47,530
They have a system where they recruit people who will comment on news stories.

341
00:34:47,530 --> 00:34:55,299
And David's on that list and I get stories. But what she said in there was really insightful and she was commenting on this lady, Jenny.

342
00:34:55,300 --> 00:35:01,450
Ronnie said, I'm leaving the thought to world because you can't beat fake news with science communication.

343
00:35:02,620 --> 00:35:10,660
And her arguments were trying to say that actually I am seeing plenty to reassure me that the public are more discerning.

344
00:35:10,990 --> 00:35:16,900
The latest trust rating showed once again that scientists remain near the top of the list of most trusted professionals,

345
00:35:17,170 --> 00:35:25,690
with 83% of the public trusting scientists to tell the truth, compared with only 17% who say the same about politicians.

346
00:35:26,680 --> 00:35:32,200
So if you are involved in research, the public trust you to say the right things.

347
00:35:33,100 --> 00:35:37,149
And in doing that, I think it's helpful to say what she said.

348
00:35:37,150 --> 00:35:44,800
He's finishing it. Say no one said it was going to be easy. And I'm open to Roland's claims that it might be getting harder or at least more bruising.

349
00:35:45,370 --> 00:35:49,270
I think that is true. The world is more breathing than ever now.

350
00:35:49,510 --> 00:35:53,110
If you just try and put out a very clear message about here's the evidence.

351
00:35:53,500 --> 00:35:57,160
But that's all the more reason for Roan and a fellow scientist to hang in there

352
00:35:57,460 --> 00:36:01,300
now would be the very worst time for scientists to return to their ivory towers.

353
00:36:03,840 --> 00:36:09,030
Now I'm just going to finish with something that it doesn't matter what I do.

354
00:36:09,420 --> 00:36:15,360
I can't do anything about it. And it doesn't matter what you try, what your strategy is.

355
00:36:15,360 --> 00:36:29,560
There are some things you just can't do something about, and there's one now, having finished that, just to say.

356
00:36:30,030 --> 00:36:33,749
One of the things I'm having a bit of trouble at the moment is, is sleeping?

357
00:36:33,750 --> 00:36:36,810
Anybody else got a problem with sleeping in the room?

358
00:36:36,990 --> 00:36:44,220
Yeah, it's sort of like I'm alright with it. But what I do is I rely on the Daily Express to inform what I might do about it.

359
00:36:45,090 --> 00:36:50,160
And those of you sleep too much, you say you're in trouble.

360
00:36:50,190 --> 00:36:56,130
I'm alright, I'll be fine. But it's actually I'm using the Daily Express to help me out.

361
00:36:56,400 --> 00:36:57,300
Thank you very much.