1
00:00:00,330 --> 00:00:07,710
Now, about ten days ago, the revolt, the ten year revolt for the Nelson lung cancer screening trial came out.

2
00:00:08,640 --> 00:00:14,970
All right. And you're in the midst of me writing an article that's going to go out in the BMJ, hopefully by the end of the week.

3
00:00:15,600 --> 00:00:20,309
So everything I say now, if it sounds good, will go in there. And if you don't understand it, I'll have to rethink it.

4
00:00:20,310 --> 00:00:24,990
For a woman to start with is going back a year and that's what was plain.

5
00:00:24,990 --> 00:00:28,620
I want you to just watch this two minute video. Let's see if we can make this work.

6
00:00:29,420 --> 00:00:37,860
The disease is early. It's a no brainer. It might be regular techniques and standards, but it's not that simple.

7
00:00:38,580 --> 00:00:44,490
The latest screening rollout testing them today comes in the form of lung health checks.

8
00:00:45,360 --> 00:00:54,480
These checks started out in trucks in a supermarket car parks in Manchester, where the incidence of lung cancer is around double the national average.

9
00:00:54,630 --> 00:01:07,290
Your arms away from the chest. NHS England says the Manchester projects on 2541 patients and found 61 lung cancer picking them up at an earlier stage.

10
00:01:08,640 --> 00:01:17,400
Now the trucks will visit ten new areas at a cost of £70 million over four years, targeting people at high risk of lung cancer.

11
00:01:18,210 --> 00:01:21,360
They say it will save hundreds of lives across the country.

12
00:01:21,960 --> 00:01:23,640
But not everyone agrees.

13
00:01:25,530 --> 00:01:34,170
NEWSNIGHT has seen a letter from senior doctors and public health officials concerned even the chief executive of NHS England.

14
00:01:34,980 --> 00:01:38,760
It raises substantial objections to lung health checks.

15
00:01:39,420 --> 00:01:44,100
They say proponents identify the homes and overplay the benefits.

16
00:01:44,520 --> 00:01:50,790
They point to Canadian data which attempt to quantify deaths for every thousand people

17
00:01:50,790 --> 00:01:57,990
scans annually for three years 391 who showed apparent indications of lung cancer.

18
00:01:58,710 --> 00:02:03,060
However, only 40 of those positives are genuine.

19
00:02:04,440 --> 00:02:11,430
Of those, 47 would not have died of the disease, 30 would have died anyway.

20
00:02:12,120 --> 00:02:21,600
And just three lives are saved. Meanwhile, the false positives all suffered major complications as a result of treatment.

21
00:02:22,200 --> 00:02:28,740
And one of those four will die. But.

22
00:02:30,460 --> 00:02:41,770
So that was interesting, wasn't it? We spent £70 million on trucks that now go into supermarkets and case find.

23
00:02:41,770 --> 00:02:49,180
People don't screen their case, find people for lung cancer, inviting people at high risk in high risk areas.

24
00:02:49,530 --> 00:02:52,720
One of them is is North Manchester, where I'm from originally,

25
00:02:53,050 --> 00:02:59,550
and this is a copy of the letter to Simon Stevens and you can see it coming from my email account.

26
00:02:59,920 --> 00:03:08,560
So I'm one of the then senior doctor people who like to make a nuisance of myself by making us think about what we're doing in health care.

27
00:03:09,190 --> 00:03:16,870
Very interesting to me how little people actually think about and go behind what we actually do in the system.

28
00:03:17,560 --> 00:03:22,870
We tend to do things and roll them out. And here's one other point, too.

29
00:03:22,870 --> 00:03:26,380
There is concern that there is no benefit to all cause mortality.

30
00:03:26,620 --> 00:03:31,090
What they would do now come back to the National Lung Screening trial with an American trial.

31
00:03:31,510 --> 00:03:38,590
But in Europe, there's been a trial called the Nelson, and it was published in a conference and it wasn't quite clear what was going on.

32
00:03:39,070 --> 00:03:46,750
We are concerned that enthusiasm for screening feeds a deliberate and independent effort to understand the benefits relative to the cost.

33
00:03:47,770 --> 00:03:51,160
Because this is not going to cost 70 million and this is only in a few areas.

34
00:03:51,460 --> 00:03:54,700
This is going to cost in excess of a billion because the.

35
00:03:54,700 --> 00:03:58,440
NELSON As your theme of annual screening,

36
00:03:58,480 --> 00:04:05,110
in fact it involved four screens within the first two and a half years and it's not even clear what happens beyond that,

37
00:04:05,110 --> 00:04:12,870
whether you should keep screening. So I wrote a blog last year called Understanding Lung Cancer Screening, and in their report,

38
00:04:12,910 --> 00:04:17,410
some really interesting point for the point about what what you've just seen in the programme

39
00:04:17,860 --> 00:04:25,800
that the screening reduce lung cancer mortality by 26% in men and between 39 and 60% in women.

40
00:04:25,810 --> 00:04:30,820
That's what within the come with them. But remember it's this thing called lung cancer mortality.

41
00:04:32,230 --> 00:04:40,840
Now immediately we read I think what happens is we remove these words in our mind and the something comes into our psyche that reduces mortality,

42
00:04:41,260 --> 00:04:44,860
because if you reduce mortality by that much, that would be really impressive.

43
00:04:44,860 --> 00:04:48,490
Wouldn't you be like, what's the on about? Okay.

44
00:04:49,870 --> 00:04:56,589
And what they do is when they do these projects, you get this stage shift of the people who are turning up.

45
00:04:56,590 --> 00:05:07,780
In Manchester, about 18% were diagnosed at stage one and after the scans turned up, more people were diagnosed with stage one, nearly half the people.

46
00:05:08,050 --> 00:05:14,410
So immediately we go, that's a good thing. Surely if you get people earlier, you can intervene.

47
00:05:15,010 --> 00:05:19,450
And that going back to that chart, remember the chart, we said there are some fast and some don't progress.

48
00:05:20,050 --> 00:05:23,860
This makes an assumption that everyone is going to progress. Therefore it's a good thing.

49
00:05:25,210 --> 00:05:33,010
Okay, now here's the analysis. This was this is I love this stuff and I think we need more of this type of stuff because it's really helpful.

50
00:05:33,250 --> 00:05:40,240
This is the Canadian Task Force, which was in the Deborah Cohen article in NEWSNIGHT, and she taken this and broke it down.

51
00:05:40,540 --> 00:05:44,290
And it's very helpful to look at this type of stuff. Okay.

52
00:05:44,680 --> 00:05:51,249
But one of the arguments in the next why it didn't apply in the UK is they had a much more inclusive early diagnosis.

53
00:05:51,250 --> 00:05:56,920
So the size of the nodule matter. So if you take very small nodules, you end up in everybody.

54
00:05:57,310 --> 00:06:00,640
If you take larger modules, you can reduce this number.

55
00:06:01,210 --> 00:06:03,910
And the Nelson said we're going to reduce this number.

56
00:06:04,300 --> 00:06:10,570
So what they've come out with really, Matt, is they're going to say we reduce the false positives and the risks of overdiagnosis.

57
00:06:11,470 --> 00:06:18,400
Okay. So remember, and this is one of the things so when you look at this, this is a thousand people.

58
00:06:18,400 --> 00:06:26,890
Yeah. And I'll come back to some of seven of the 40 diagnosed lung cancers would not have caused illness or death overdiagnosis.

59
00:06:27,160 --> 00:06:31,360
So it's seven per thousand. Yeah. So you identify a thousand people,

60
00:06:31,360 --> 00:06:39,150
you screen them and of them 40 people will have a diagnosis of lung cancer after they've had the but seven of them wouldn't

61
00:06:39,160 --> 00:06:46,660
have progressed whatever that's the overdiagnosis in 33 was that's a sort of estimate of how much overdiagnosis is going on.

62
00:06:47,470 --> 00:06:50,500
Okay. So the Nelson trial comes out.

63
00:06:50,920 --> 00:06:55,329
So this is some of the information I'm now going to provide you and try and walk

64
00:06:55,330 --> 00:06:59,260
you through how I think about trying to tease out the important information.

65
00:06:59,890 --> 00:07:06,940
Now, one of the things about evidence based medicine, which is really interesting, is you really can make a name for yourself if you can be bothered,

66
00:07:08,200 --> 00:07:11,950
because nobody else can be bothered to look at this stuff and try and understand it.

67
00:07:12,430 --> 00:07:17,740
There's only a few people who really can say, I am going to try and understand this and share that information.

68
00:07:18,010 --> 00:07:24,879
So if you want a route to a really impressive career, just having to look at sort of spend an hour of my time and try and understand

69
00:07:24,880 --> 00:07:28,570
it so I can do a talk like this in your own institution and maybe write it up.

70
00:07:30,710 --> 00:07:34,340
So so this is a bit which I wrote the year before.

71
00:07:34,370 --> 00:07:39,770
This is the press release and when it came, but they said we doing the screening and the abstracts.

72
00:07:39,800 --> 00:07:47,090
This was at the conference reported there were 214 lung cancer deaths in the male control lab and 57 deaths in the screen arm.

73
00:07:47,660 --> 00:07:52,070
Yeah. See that thing happening again? The lung cancer disappears.

74
00:07:52,880 --> 00:07:58,520
Yeah. So suddenly. Well, okay. And so actually, it reduces and it comes back again.

75
00:07:58,520 --> 00:08:05,570
Mortality rate ratio for men is 26% reduction in death or 26% reduction in lung cancer death.

76
00:08:06,230 --> 00:08:13,010
Yeah. So and this is how it how it then gets interpreted decrease mortality by 26% in high risk men.

77
00:08:13,820 --> 00:08:18,469
Why are we not getting on with it? Then you see what really matter.

78
00:08:18,470 --> 00:08:26,330
These language matters could get missed out and then this gets interpreted into this and therefore we end up with a screening program.

79
00:08:26,780 --> 00:08:35,660
Now, this is a Nelson trial that came out February six, half the reduce lung cancer mortality volume 50 screening in a randomised controlled trial.

80
00:08:36,330 --> 00:08:44,780
Okay, now how long for mortality was significantly lower among those who underwent volume screening?

81
00:08:44,780 --> 00:08:49,390
Among those who underwent no screening. Okay. All right.

82
00:08:49,400 --> 00:08:56,900
I'm going to try now break that down, because when I read this stuff, I never feel that informed.

83
00:08:57,980 --> 00:09:01,580
I read these abstracts and I try and go, What is this trying to tell me?

84
00:09:01,850 --> 00:09:08,160
And look at the way the data is presented. Things like 4.91 cases per 1000 person years.

85
00:09:08,180 --> 00:09:13,850
What does that mean? What's a difference between 4.90.91 of a person?

86
00:09:14,420 --> 00:09:16,730
Is that important? How are we supposed to interpret that?

87
00:09:16,940 --> 00:09:23,710
Then you flip to relative rate, cumulative rate ratios, and then you come back to another one and then you've got the rate ratio.

88
00:09:23,720 --> 00:09:29,360
So we flip in math all the time. And so even I find this really difficult to understand.

89
00:09:29,900 --> 00:09:33,680
And the only people who don't are the ones who don't read it so soon.

90
00:09:33,730 --> 00:09:38,780
If you read these and you go, This is difficult, you should do. Because it's not clear to me if this is deliberate,

91
00:09:39,890 --> 00:09:50,840
an attempt to make sure nobody quite understand this so you can make this or it's just incompetence or it's deliberate about everything, isn't it?

92
00:09:51,620 --> 00:09:54,950
But it certainly isn't wrote in, in a way, to understand the information.

93
00:09:55,430 --> 00:09:59,260
So let me take you through the information and you can see.

94
00:09:59,300 --> 00:10:04,879
So the first thing is table two screening test results in each screening round for male participants.

95
00:10:04,880 --> 00:10:08,780
So the first thing is a go. Right, here we go. Round one, two, three, four.

96
00:10:09,110 --> 00:10:12,440
Okay. So there's 22,600. Yeah.

97
00:10:12,950 --> 00:10:17,960
And uptake of that I'm happy here is 90%.

98
00:10:18,260 --> 00:10:24,620
That's helpful information. 90%. That's qua what isn't it. In a screening round, this is a pretty compliant population, but that's helpful.

99
00:10:24,950 --> 00:10:31,130
But what I do then is try and break this down into useful information because I don't think these figures help me.

100
00:10:32,120 --> 00:10:39,620
So I break everything down to one thing. If I just take that thousand people a bit like Deborah and showed us what, and if you take a thousand people,

101
00:10:39,890 --> 00:10:46,560
71 people will have a positive test for screening round of which 31 will detect lung cancer.

102
00:10:46,560 --> 00:10:51,360
That's 44%. Yeah. And 40 will be falsely diagnosed.

103
00:10:51,620 --> 00:10:59,390
Yep. Happy with that. And so that's that figure there, the 44% I was round up because I don't believe there's such a thing as a half a person.

104
00:11:00,020 --> 00:11:04,310
Yeah. And, and so it's either 44 or 44, but that doesn't matter.

105
00:11:05,210 --> 00:11:11,570
So you happy? That's my first step. When I interpret the take a thousand people, how many will have a positive test?

106
00:11:12,020 --> 00:11:18,770
Now, if I took that and took that away, you can actually say anything sensible about that could be based on that table.

107
00:11:19,490 --> 00:11:22,700
But once you give a thousand, it becomes really simple. All right.

108
00:11:23,060 --> 00:11:30,920
But 71 people have a positive test and nearly a half of them will have that lung cancer and just over half won't.

109
00:11:31,070 --> 00:11:35,180
And that's that important figure there. Okay. So that's my first bit.

110
00:11:35,840 --> 00:11:44,600
My second bit then is to start to look at this piece of information which tries to tease out some of these issues of overdiagnosis.

111
00:11:47,120 --> 00:11:51,380
So what what this says, if you look at the whole number of cases here.

112
00:11:52,010 --> 00:11:57,650
Yeah. Any lung cancer, if 344 and in the control group it's 404.

113
00:11:58,220 --> 00:12:04,430
Yeah. So the difference between the two is what they're trying to say is the overdiagnosis component.

114
00:12:04,430 --> 00:12:08,140
You've got 40 excess cancers that you've detected. Yeah.

115
00:12:09,080 --> 00:12:13,040
And that's in the group they say an excess of was found among the male participants in the screening

116
00:12:13,040 --> 00:12:19,880
group ten years after randomisation a suggested excess in the in the incidence rate of 20%.

117
00:12:20,300 --> 00:12:25,230
That's what it's got in the paper. But again I can't understand that just doesn't make sense to me.

118
00:12:25,250 --> 00:12:28,880
What am I supposed to use that information. So again, I take.

119
00:12:29,180 --> 00:12:35,240
Okay if it 40 cases overdiagnosis per fictional thousand people screened that's six per 1000.

120
00:12:36,110 --> 00:12:40,160
Very similar to the actual number in the analysis, isn't it? But the same actually.

121
00:12:40,580 --> 00:12:44,790
That's helpful. It backs up and verified. It makes me feel that makes sense.

122
00:12:45,470 --> 00:12:50,060
So I find this information much more useful than that which is on interpret all to me.

123
00:12:50,060 --> 00:12:58,820
I don't know how I'm supposed to use that. Well, and then third, we come to table four, which is a incredibly interesting table.

124
00:12:59,060 --> 00:13:05,720
If a cause of death of deceased male participants at ten years follow all the important information.

125
00:13:05,870 --> 00:13:10,879
So let's just break this down. The first bit of information is the cause of death.

126
00:13:10,880 --> 00:13:19,940
And here's the lung cancer death. Okay. Hundred and 16 screening group versus 210 in the control group where you can see that death.

127
00:13:20,690 --> 00:13:24,740
And that's point 76. That's where you get the 24% reduction.

128
00:13:25,550 --> 00:13:29,629
Everybody happy with that? That's your lung cancer specific.

129
00:13:29,630 --> 00:13:34,490
Mortality is a really important piece of information.

130
00:13:35,090 --> 00:13:47,420
Okay. Which is down here. But again, you see how here we have the numbers and then suddenly we now go two deaths per thousand person year.

131
00:13:47,450 --> 00:13:51,260
Why did we do that? What's going on now?

132
00:13:51,410 --> 00:13:56,840
And in fact, when you do that, what you see here is actually there's no difference.

133
00:13:56,840 --> 00:14:02,990
And in fact, there's 1% more death in the screening group than there is in the non screening group,

134
00:14:03,920 --> 00:14:09,110
which is really interesting and I don't mean it with all this mortality and if you go on the table, you'll see.

135
00:14:09,110 --> 00:14:13,820
Well, I was speaking to Jeff Aronson today. I mean, are there any clinicians in the room?

136
00:14:15,080 --> 00:14:17,790
So on death certificate, I've done that with you.

137
00:14:18,020 --> 00:14:22,940
How many people would say the cause of death is symptom funds and abnormal clinical and laboratory findings?

138
00:14:25,340 --> 00:14:29,450
If you did the death certificate, you would be in front of the GMC in the beginning.

139
00:14:29,450 --> 00:14:32,989
You don't know what you're doing. That's a medical student potentially.

140
00:14:32,990 --> 00:14:37,730
He's got all these elevated tests. So I'm like, These people are supposed to be blindly allocating this,

141
00:14:38,030 --> 00:14:44,900
but somehow you're 86% more likely to die if you've had screening from symptom signs and abnormal clinical and lab finding.

142
00:14:45,890 --> 00:14:47,710
Just somewhat. Not right in my mind.

143
00:14:48,390 --> 00:14:54,590
And not only that, look at this you are twice as likely to die from integrated nutritional and metabolic diseases.

144
00:14:56,060 --> 00:15:03,260
So imagine if I released a paper thing. Screening increases double your risk of death from endocrine and nutritional and metabolic diseases.

145
00:15:03,560 --> 00:15:10,790
You'd be incredibly worried about that. So this part of the graph is extraordinary to me what's going on that actually,

146
00:15:10,790 --> 00:15:15,230
although you're talking about this reduction in lung cancer, we've actually seen some increases in other diseases.

147
00:15:16,190 --> 00:15:22,010
So what I do that is I try and bring everything together into one piece of information that help me understand it.

148
00:15:22,350 --> 00:15:30,440
Okay. And this is what it looks like. This is a flow chart, which I do for and I put the important information about who's in the trial.

149
00:15:30,680 --> 00:15:35,630
And so I take 2000 high risk individuals. A thousand have lung cancer in 2000 don't.

150
00:15:35,930 --> 00:15:43,430
Yeah. So the first thing, if there's 132 that that and hundred and 30 that there's a little bit more.

151
00:15:43,430 --> 00:15:51,440
But you can say it's no difference really easy. But there are some death there are 24 of these are caused by lung cancer as opposed to 32 here.

152
00:15:51,860 --> 00:16:03,800
But extraordinarily, there are 108 or the causes here compared to 99% more likely to die from another cause if you have screening.

153
00:16:04,550 --> 00:16:11,720
Oh, my God. And then under of the Fathom screen you have 71 will have a positive screening test of the 31

154
00:16:11,720 --> 00:16:18,200
will have lung cancer 40 will be both positive and of the 31 effect will be overdiagnosed.

155
00:16:19,430 --> 00:16:27,140
That's that's my one slide to explain to you about lung cancer and the effect of screening.

156
00:16:27,980 --> 00:16:35,030
Now, if I was in a decision making mode, if I explained and showed you people that you can come back to the end of this talk, we might put it up.

157
00:16:35,600 --> 00:16:40,760
It can help you really understand simply the numbers when you do it this way.

158
00:16:41,180 --> 00:16:42,649
But actually it takes a bit of time.

159
00:16:42,650 --> 00:16:48,290
I've done this for years and it still took me a bit of time to tease around the paper, to pull them bit of information.

160
00:16:48,290 --> 00:16:52,730
But every table in the paper is unhelpful to me to get to this point.

161
00:16:53,780 --> 00:16:57,380
Isn't that interesting? And all the information we want is in that way.

162
00:16:59,570 --> 00:17:03,080
So I think this is incredibly interesting when it comes to decision making.

163
00:17:03,410 --> 00:17:04,639
And I want to point this.

164
00:17:04,640 --> 00:17:11,959
This is a systematic review that basically shows of all the studies, you put them all together that actually you come to the same example.

165
00:17:11,960 --> 00:17:18,740
It doesn't include the now for mortality data, but when you add that in, you'll see there's no mortality effect and it gets this pretty similar,

166
00:17:18,740 --> 00:17:24,770
about 20% reduction for GFR screening if you put in lung cancer survival.

167
00:17:25,010 --> 00:17:28,820
Lung cancer survival. Okay, now.

168
00:17:29,140 --> 00:17:34,150
For really interesting epidemiological concepts to think about lead time bias.

169
00:17:34,150 --> 00:17:38,380
I'm not really going to discuss, but that's why five year survival rates are a waste of time.

170
00:17:39,520 --> 00:17:43,360
The more you screen, the more people pick up earlier, the more they're going to live longer.

171
00:17:43,360 --> 00:17:49,150
But you might not affect overall survival. That's a very clear concept that we understand and trying to understand.

172
00:17:49,420 --> 00:17:51,940
But ten years ago, everything came in five year survival.

173
00:17:52,330 --> 00:17:58,660
And the classic one is in America, our five year survival for 80% for prostate cancer in the UK, the 5%.

174
00:17:59,350 --> 00:18:02,589
And that's why we can't have UK national health care.

175
00:18:02,590 --> 00:18:06,190
But actually the mortality rate and the survival is identical.

176
00:18:06,190 --> 00:18:09,550
You die at the same point. Both the US and the UK, you don't affect survival.

177
00:18:10,000 --> 00:18:16,329
But this is competing interests. Competing interests. The key diagnoses bias and slippery linkage bias.

178
00:18:16,330 --> 00:18:20,110
Isn't that interesting bias? If so, the first is competing risks.

179
00:18:21,670 --> 00:18:25,840
Competing ref means that things compete for your for our death simpler.

180
00:18:26,860 --> 00:18:31,899
And one of the problems is again is if you censor people when they've had the event of

181
00:18:31,900 --> 00:18:36,400
interest so if they die of some of the cause means they can't die of lung cancer either.

182
00:18:37,870 --> 00:18:44,859
So you also have a mathematical problem that you take them out of the analysis and therefore actually

183
00:18:44,860 --> 00:18:51,700
overestimate in the screening group the event rates if you don't account for competing interest.

184
00:18:52,450 --> 00:18:59,560
Yeah. So some studies don't take account of competing interests and prostate cancer is one of the best way we know you do that.

185
00:18:59,830 --> 00:19:02,860
People would effectively prefer a more likely outcome now.

186
00:19:03,130 --> 00:19:10,120
But in lung cancer is incredibly important. But there are also mathematical issues in terms of the censoring, which is important.

187
00:19:10,120 --> 00:19:17,920
And this paper is really neat because what it says is your kaplan-meier hazard ratios overestimate the effect

188
00:19:18,040 --> 00:19:22,570
if you don't account for the fact you're taking people out because you sense that them when they died.

189
00:19:23,650 --> 00:19:27,220
The second thing is these two concepts which are really interesting,

190
00:19:27,430 --> 00:19:33,700
we run about a thing called the catalogue of bias and IT Fund, and these are two new additions to the catalogue of bias.

191
00:19:34,360 --> 00:19:38,800
Sticky diagnosis bias occurs when death from other causes in the experimental group,

192
00:19:38,950 --> 00:19:44,229
in the screening group are incorrectly attributed to the disease is a target screening.

193
00:19:44,230 --> 00:19:48,879
So you could do it both ways. You can say, oh, I think we're interested in lung cancer.

194
00:19:48,880 --> 00:19:55,810
We're going to attribute this to lung cancer because that's in our minds, it's a it's a thought heuristic that happens.

195
00:19:56,380 --> 00:20:02,200
But you can do it the other way when you flip it the other way, which biases the favour, the screening group.

196
00:20:02,200 --> 00:20:04,209
So you can do it the other way when you're slippery.

197
00:20:04,210 --> 00:20:11,860
And so there are different biases occurring which particularly can occur if you're really invested in screening.

198
00:20:12,760 --> 00:20:21,910
You might falsely attribute things to biochemical factors when actually they may have had a lung cancer, they may have had something happen.

199
00:20:22,420 --> 00:20:25,350
You've removed the lung cancer, but a year later they died.

200
00:20:25,360 --> 00:20:30,370
You don't go, well, actually, maybe that was attributable to lung cancer, so you split the diagnosis.

201
00:20:31,840 --> 00:20:37,960
So I'm going to finish here with this position statement that the World Health Organisation talked about.

202
00:20:37,990 --> 00:20:46,780
I think it's very important in terms of screening both lead to an improvement in any results defined in terms of mortality, physical,

203
00:20:46,780 --> 00:20:54,820
social and emotional function, pain and satisfaction among those in whom early diagnosis is achieved or in the other members of the community.

204
00:20:56,320 --> 00:20:57,220
Why is that important?

205
00:20:57,520 --> 00:21:09,490
Well, okay, I'm okay if you have a screen detected group that actually there's no overall mortality effect if you affect some of these other issues.

206
00:21:10,360 --> 00:21:19,270
So for instance, if you can remove my lung cancer at stage one and I have a better quality of life because of that, I still die at the same point.

207
00:21:19,840 --> 00:21:27,070
That's a worthwhile intervention. So all of the outcomes have health related quality of life in there,

208
00:21:27,550 --> 00:21:32,650
but nobody has reported the health related quality of life beyond the first six months or a year,

209
00:21:32,920 --> 00:21:37,180
because all they're interested in is the emotional impact of the first six months after screening,

210
00:21:37,360 --> 00:21:40,600
and they say everything's okay by the year you're okay.

211
00:21:40,810 --> 00:21:44,950
But actually we really interested in what is the overall impact of quality of life.

212
00:21:44,950 --> 00:21:51,609
But none of this has been reported and it's really quite an important issue that what we

213
00:21:51,610 --> 00:21:57,909
stuck on is this measuring screening is where it started with this idea of a cancer.

214
00:21:57,910 --> 00:22:06,040
Specific mortality is why we screen. But actually a common phenomenon is we don't improve overall mortality.

215
00:22:06,430 --> 00:22:11,140
And that's exactly what's happening in lung cancer. The question now is,

216
00:22:11,350 --> 00:22:23,110
and this is where the debate is is quite a significant number of people consider we should now roll out c.t screening based on that Neilsen data.

217
00:22:24,460 --> 00:22:28,930
Some people who keep writing to everybody saying this could be a.

218
00:22:29,080 --> 00:22:36,040
Significant error because it will divert resources so that when you need your diagnosis,

219
00:22:36,040 --> 00:22:40,270
there won't be enough radiography there to help you and you need to be treated quickly.

220
00:22:40,930 --> 00:22:45,280
You could invest the money somewhere else. For instance, an amazing smoking reduction.

221
00:22:45,490 --> 00:22:50,290
You've got a belly in pain to play with. You might actually, if you can get people to smoke less.

222
00:22:50,290 --> 00:22:53,500
That would be still the biggest prevention and public health issue you can deal with.

223
00:22:53,860 --> 00:23:01,510
So if you were the person faced with that graph, where would you spend your billion in the next year?

224
00:23:02,590 --> 00:23:06,250
Would you spend it on CT screening or would you spend it on something else?

225
00:23:07,090 --> 00:23:07,930
Thank you very much.