1
00:00:00,530 --> 00:00:04,520
Welcome to the season of Future Makers Brain and Mental Health.

2
00:00:05,120 --> 00:00:09,980
I'm Professor Belinda Lennox. I'm a psychiatrist and a researcher here in Oxford.

3
00:00:10,430 --> 00:00:17,120
And this season, you'll be joining me as we demystify the science behind the most complex object in the known universe,

4
00:00:17,330 --> 00:00:23,900
our brains, and look at the wide reaching impacts of mental illness on individuals and society.

5
00:00:24,860 --> 00:00:32,120
I'll be introducing you to some of Oxford's best academic minds working to solve the greatest challenges in brain and mental health.

6
00:00:32,690 --> 00:00:36,979
And I'll also be speaking to guests from beyond the university to bring their

7
00:00:36,980 --> 00:00:41,750
perspectives and lived experience and get a sense of the impact of what we're doing.

8
00:00:42,170 --> 00:00:47,720
Join us as we discover how Oxford is shaping the future of brain and mental health research.

9
00:00:50,860 --> 00:00:54,939
In this episode of Future Makers, I'm delighted to be joined by my colleague,

10
00:00:54,940 --> 00:01:00,370
Professor Saeid Fazal, to talk about current research around suicide and suicide prevention.

11
00:01:00,940 --> 00:01:09,910
Senior is professor of psychiatry, consultant, forensic psychiatrist and the new director of the Centre for Suicide Research in Oxford.

12
00:01:10,840 --> 00:01:21,159
Oxford has a really proud tradition of suicide research over many decades, doing research into the nature causes and of suicide,

13
00:01:21,160 --> 00:01:26,860
and translating those findings into prevention, treatment and improved support.

14
00:01:26,860 --> 00:01:32,059
And this was pioneered by the work of Professor Keith Horton. In fact, actually,

15
00:01:32,060 --> 00:01:36,620
the rates observation of suicide amongst university students was one of the

16
00:01:36,620 --> 00:01:41,120
reasons that the University Department of Psychiatry was founded 50 years ago.

17
00:01:42,050 --> 00:01:52,490
But the work of Keith was really important. In the 1990s, he revealed the extent and characteristics of suicide following paracetamol overdose.

18
00:01:52,820 --> 00:01:58,760
And this work led to the UK legislation that restricted pack size in 1998.

19
00:01:59,540 --> 00:02:07,219
And he and colleagues then showed that this reduction in pack size was followed by a substantial reduction of over 30% in the

20
00:02:07,220 --> 00:02:15,920
number of deaths from paracetamol poisoning and actually work ten years ago now showed that that benefit had been sustained.

21
00:02:15,920 --> 00:02:20,390
Although maybe we might come on to discussing whether that's actually still the case nowadays.

22
00:02:21,410 --> 00:02:29,780
But further work from from Keith and from this department showed the particular fatalities and the

23
00:02:29,780 --> 00:02:36,800
danger of co proximal which is another painkilling drug which is a paracetamol plus another compound.

24
00:02:36,920 --> 00:02:40,670
And it showed that it was particularly toxic in overdose.

25
00:02:41,240 --> 00:02:45,740
And that work led to the withdrawal of CO proximal in 2008.

26
00:02:45,740 --> 00:02:51,740
And again that that group showed that that corresponded with a reduction in deaths due to poisoning with that drug.

27
00:02:51,890 --> 00:02:59,360
So basically the research, the observation, the description of the means of suicide and restricting those means is actually a

28
00:02:59,360 --> 00:03:03,979
really powerful public health intervention that could be used for suicide prevention,

29
00:03:03,980 --> 00:03:12,090
I think is almost the most powerful way that research can influence reduction in suicide rates going forward.

30
00:03:12,140 --> 00:03:20,200
I'm sure we will discuss it, but to bring you in, since your work particularly focuses on high risk populations, doesn't it?

31
00:03:20,230 --> 00:03:24,440
I want to describe a little bit about the work that you've done over many years now.

32
00:03:24,530 --> 00:03:34,760
Yes. So I've focussed, as you say, mainly on high risk populations and they include people with certain medical conditions,

33
00:03:35,240 --> 00:03:40,610
usually neuropsychiatric conditions like severe mental illness such as schizophrenia,

34
00:03:40,610 --> 00:03:51,410
bipolar disorder, but also what I think are relatively neglected conditions, but also very common conditions like traumatic brain injury, epilepsy.

35
00:03:51,800 --> 00:04:01,190
And then we've also recently looked at a large number of people with quite chronic physical health problems like diabetes,

36
00:04:01,520 --> 00:04:05,720
lung diseases, heart diseases, and looked at their suicide risk.

37
00:04:06,860 --> 00:04:13,669
And we've been interested in those groups in particular because they are groups of people that do access medical services.

38
00:04:13,670 --> 00:04:18,709
So there is a potential to intervene and that's important.

39
00:04:18,710 --> 00:04:26,870
And we also know that there are good quality interventions that can be offered to those people.

40
00:04:27,590 --> 00:04:33,560
The second group, high risk group, you call them high risk group, are people in touch with the criminal justice system.

41
00:04:34,070 --> 00:04:38,540
So a lot of people every year arrested. Most of those people aren't charged.

42
00:04:38,540 --> 00:04:45,439
But we know that the whole process from arrest all the way through to being in prison and

43
00:04:45,440 --> 00:04:50,510
then being released from prison is associated with quite high risk of of suicide death.

44
00:04:51,860 --> 00:05:00,349
And again, because they're in contact with services, sometimes not very good contact with medical services,

45
00:05:00,350 --> 00:05:02,870
but they're in contact with some sort of services from the state.

46
00:05:03,800 --> 00:05:10,040
They are interventions that can be offered and provided, and there are areas which can be improved.

47
00:05:10,040 --> 00:05:17,540
So these are all areas where there are very discrete changes that can be made that can improve things.

48
00:05:17,930 --> 00:05:24,709
And the point you make about paracetamol is a very good one because restricting access to means has been the most powerful,

49
00:05:24,710 --> 00:05:28,850
I think, approach to suicide prevention in the world to date.

50
00:05:29,240 --> 00:05:32,270
And there are simple things you can also do.

51
00:05:32,270 --> 00:05:36,589
For instance, in prisons where I do some clinical work, you know,

52
00:05:36,590 --> 00:05:44,660
removing ligature points are points that people can use to attach things like sheets

53
00:05:45,320 --> 00:05:52,340
that they used to sleep in or they they create types of ropes from sheeting.

54
00:05:53,090 --> 00:06:02,360
I mean, just removing those is a very simple intervention, very scalable that you can intervene and and administer widely,

55
00:06:02,360 --> 00:06:08,959
implement widely so that they're there are some examples of where, you know,

56
00:06:08,960 --> 00:06:15,650
the high risk groups have been identified in the work we've done and then the and then what we've been trying to do is,

57
00:06:15,890 --> 00:06:26,390
is not only point out the these the risk is high in these groups, but also what you can do to improve your assessment and then also your treatment.

58
00:06:26,390 --> 00:06:29,820
So we the approach we've taken is really a.

59
00:06:30,580 --> 00:06:37,090
Holistic. If you want to be looking at the problem where the problem really exists and what can

60
00:06:37,090 --> 00:06:41,350
be done to improve the assessment and then the treatment to management of the problem.

61
00:06:41,710 --> 00:06:44,980
Yeah. So there's so many things I want to discuss from that.

62
00:06:45,010 --> 00:06:50,440
I suppose one thing that immediately arises, you say something simple like reducing, taking away look at you.

63
00:06:50,440 --> 00:06:54,970
I mean, it's just seems such common sense, doesn't it? It also seems so simple.

64
00:06:55,330 --> 00:06:57,850
And. And I suppose just thinking through, you might think, well,

65
00:06:57,850 --> 00:07:03,400
why does removing the means if you have somebody that really, you know, wants to die is suicidal.

66
00:07:03,520 --> 00:07:08,680
Why does such a simple thing actually, I mean, wouldn't somebody then go and find another way.

67
00:07:09,250 --> 00:07:19,120
Yeah. Well, that doesn't seem to be the case. I mean, usually when people have a very acute suicidal thoughts and it doesn't last very long,

68
00:07:19,120 --> 00:07:23,200
and and so sometimes it just goes away, you know, with time,

69
00:07:24,070 --> 00:07:27,959
and sometimes it goes away with with some sort of contact with, let's say,

70
00:07:27,960 --> 00:07:35,080
a family member or friend or, as I say, just goes away because it's a temporary event.

71
00:07:35,380 --> 00:07:44,320
And so that's one of the key things, is that because it's very time limited, then having interventions like this, it's quite important.

72
00:07:44,320 --> 00:07:52,959
I mean, interesting enough in the US, you know, the the means that people use to die from suicide are guns.

73
00:07:52,960 --> 00:08:02,080
And so actually having restrictions on on on guns among people who mental health conditions actually reduces suicide risk.

74
00:08:02,080 --> 00:08:06,399
I mean, a lot of the focus is being in legislation on people being dangerous.

75
00:08:06,400 --> 00:08:09,490
But actually that's not really where the the focus should be.

76
00:08:09,490 --> 00:08:11,530
I think the focus should be on suicide prevention.

77
00:08:11,530 --> 00:08:20,589
And the and that's really where the research should be points towards really that by restricting guns,

78
00:08:20,590 --> 00:08:25,270
you can reduce suicides among people with mental illness because again,

79
00:08:25,570 --> 00:08:32,200
it's the it's having the means having it easily available, which is is the thing to address in the UK.

80
00:08:32,200 --> 00:08:40,969
I mean there was quite a innovative in a way, but also evidence based intervention where they remove ligature points from hospitals,

81
00:08:40,970 --> 00:08:48,100
from psychiatric hospital wards, inpatient wards, and that led to a reduction of suicide in inpatients.

82
00:08:49,570 --> 00:08:55,629
And that's been shown to have been, you know, quite, quite a large reduction in suicide rates.

83
00:08:55,630 --> 00:09:02,230
And that, again, it's it's an example of where good quality research can lead to a simple,

84
00:09:02,230 --> 00:09:06,850
scalable intervention which can then lead over time to reduction in rates.

85
00:09:07,060 --> 00:09:11,440
Yeah, Yeah. I want to ask you a bit about methods of before people.

86
00:09:11,440 --> 00:09:19,419
So sort of thinking, oh, this is going to be I think this is really actually of key importance because describing as you do a lot,

87
00:09:19,420 --> 00:09:29,409
you look across whole populations, don't you really comprehensive data sets and you look to go beyond just the associations of what's

88
00:09:29,410 --> 00:09:35,980
associated with suicide in a particular population or particular characteristic to really get it causation.

89
00:09:36,760 --> 00:09:41,920
And that's really absolutely key, isn't it? Because otherwise you get kind of wrong messages.

90
00:09:42,190 --> 00:09:50,470
For instance, you know, a lot of noise about antidepressants make people suicidal, for instance, a sort of fallacious argument.

91
00:09:50,770 --> 00:09:55,590
But actually that's why you need high quality epidemiology, isn't it?

92
00:09:55,600 --> 00:10:00,370
I think so. And I think that's the area where I think we've done the most.

93
00:10:00,580 --> 00:10:11,410
I suppose the the most innovative work is around methods, is trying to work out how to disentangle really quite complex.

94
00:10:13,400 --> 00:10:22,460
Series of issues and suicide is usually caused not just by one factor, but an accumulation of multiple factors over a life course.

95
00:10:23,060 --> 00:10:32,540
So how can you disentangle the relative strength of one particular risk factor that you want to isolate and ideally then prevent or treat?

96
00:10:33,270 --> 00:10:36,829
It requires quite a lot of thinking and careful work.

97
00:10:36,830 --> 00:10:45,110
And actually one of the I think one of the things that we've done very well is we've we've harness the power of large data, big data,

98
00:10:45,890 --> 00:10:50,150
but not just that having good quality methods is key because you can have large data,

99
00:10:50,150 --> 00:10:56,420
but if you analyse it superficially, you get the wrong, you can get the wrong answers.

100
00:10:56,660 --> 00:11:01,970
But what we've tried to do is be very careful about how it's analysed.

101
00:11:01,980 --> 00:11:07,670
We've used new methods such as for instance, one of the methods we've used is using sibling controls,

102
00:11:07,670 --> 00:11:14,700
where this is using a brother or sister to someone who has, let's say, a disorder of interest.

103
00:11:14,720 --> 00:11:21,709
So for instance, you could look at someone with a head injury and then you compare the sibling who doesn't have a head injury,

104
00:11:21,710 --> 00:11:29,870
this same sex sibling who doesn't have a head injury. And then you follow those two people up over time and ideally,

105
00:11:29,870 --> 00:11:36,920
you really follow them for a long period of time because you need that longitudinal perspective to make sense of the information.

106
00:11:37,340 --> 00:11:44,749
And that enables you to disentangle someone's background that you know, that the child had background that they share a family background,

107
00:11:44,750 --> 00:11:50,210
the family influences that they share from the impact of a of a head injury.

108
00:11:50,570 --> 00:11:55,340
In this case, it may also be another piece of what we did was an epilepsy.

109
00:11:55,340 --> 00:12:00,350
And that was, again, I think at the time it was a new piece of work.

110
00:12:00,350 --> 00:12:02,360
People hadn't really studied epilepsy in this way.

111
00:12:02,360 --> 00:12:10,819
And again, it's very, very common condition can be treated, very important condition to to treat effectively.

112
00:12:10,820 --> 00:12:16,700
But we highlighted the importance of the suicide risk and the importance of of of being

113
00:12:16,700 --> 00:12:21,379
aware of that and addressing it over the period of the whole period of the illness,

114
00:12:21,380 --> 00:12:24,540
not just right at the beginning or in the middle of the illness.

115
00:12:24,540 --> 00:12:29,450
So when someone has an acute episode, a relapse of some sort.

116
00:12:29,810 --> 00:12:32,720
So methods are important. And I would say the other thing about methods,

117
00:12:32,730 --> 00:12:39,070
I think the department again is really a leader in this is that we've we triangulate the evidence with systematic review.

118
00:12:39,230 --> 00:12:46,730
So systematic review is a way of bringing together in a transparent, comprehensive way information about.

119
00:12:48,060 --> 00:12:53,130
Let's say, risk factors or treatments, bring it together, making sense of it.

120
00:12:54,000 --> 00:13:00,149
And what we've tried to do is, is is use that approach with all these epidemiologic approaches and sort of triangulate

121
00:13:00,150 --> 00:13:06,430
those there's those different pieces of information because it's a complicated jigsaw.

122
00:13:06,450 --> 00:13:11,070
You really need to triangulate the information to make sense of it.

123
00:13:11,970 --> 00:13:15,810
And good quality reviews can be very difficult to do. People think it's simple.

124
00:13:15,810 --> 00:13:22,560
Oh, you just going away of a week and and, you know, and a search on the search engine and put the information together.

125
00:13:22,600 --> 00:13:27,480
Actually, my my experience has been that, you know, good quality ones take years to do.

126
00:13:28,500 --> 00:13:33,960
They need to be updated also regularly. They require quite a lot of thought and care.

127
00:13:34,890 --> 00:13:36,930
And you've got to get the question really right.

128
00:13:36,930 --> 00:13:44,070
You've got to get the your approach really right, how you deal with all the potential spurious associations that you find.

129
00:13:45,240 --> 00:13:54,209
So I think that that's also part of this is not just using good quality methods to interrogate new large datasets,

130
00:13:54,210 --> 00:13:58,210
but also situating that in the context of good quality reviews.

131
00:13:58,230 --> 00:14:06,270
And I think that's something that we've been doing for 20 years now really well, and that by using this method of sibling control,

132
00:14:06,270 --> 00:14:12,690
so really controlling for absolutely everything apart from as well as much as you possibly can other than the event,

133
00:14:12,690 --> 00:14:15,750
the head injury or the epilepsy and then looking for outcomes.

134
00:14:16,110 --> 00:14:20,819
Has anything surprised you? I mean, I mean, did anything fallout,

135
00:14:20,820 --> 00:14:28,590
anything novel that you found that previously perhaps might have been associated with outcomes of head injury or epilepsy?

136
00:14:29,880 --> 00:14:38,670
Well, I think I mean, I think we were able to, I think, confirm what some studies had suggested.

137
00:14:39,180 --> 00:14:44,700
So for instance, and we were able to give some sort of magnitude to the risk of the estimate.

138
00:14:44,700 --> 00:14:52,770
So it's quite important. So you can say something is a risk factor but doesn't necessarily really help you if it's a very small impact or very low.

139
00:14:52,920 --> 00:14:56,160
But you want to know if it's small or very large because that will influence.

140
00:14:57,280 --> 00:15:01,870
You know, to what extent, you know, you focus your resources on that particular risk factor.

141
00:15:02,260 --> 00:15:08,680
So in the case of head injury and epilepsy and these co-morbid physical health,

142
00:15:08,680 --> 00:15:16,240
chronic physical health conditions and all of those, we were able to provide some some precision to the magnitude of the effects.

143
00:15:16,630 --> 00:15:24,250
We also able to clarify that there was an effect because I think in the past, for some of them, there's been a bit of uncertainty.

144
00:15:24,580 --> 00:15:31,680
People weren't sure. Is it the people who have head injury anyway because they're impulsive and maybe these people are just impulsive by nature.

145
00:15:31,690 --> 00:15:37,180
It's got nothing to do with head injuries who are able to to, in a sense clarify that question.

146
00:15:38,530 --> 00:15:41,589
For schizophrenia, bipolar disorder, we've done similar work as well.

147
00:15:41,590 --> 00:15:48,820
And again, what that does is it help it helps provide some precision of the match, matches the effect and highlight the man's effect.

148
00:15:49,990 --> 00:15:57,250
The one area I think there was a bit surprising for us was that we looked at the very large number of people who get community sentences.

149
00:15:57,250 --> 00:16:05,080
So these are people who don't get very serious, who haven't committed very serious crimes, but nevertheless have convicted of a crime.

150
00:16:05,500 --> 00:16:11,740
And there we just published a study actually last week where we found really quite high rates of suicide in people.

151
00:16:12,390 --> 00:16:15,400
And by high just mean relative to what you would expect.

152
00:16:16,330 --> 00:16:23,469
So, you know, two or three times more than you'd expect. And that was new again, I think, and and important,

153
00:16:23,470 --> 00:16:28,810
because that isn't something that people even think about when they're looking at how to manage

154
00:16:28,810 --> 00:16:33,670
people in the community who've had these these type of sentences they think very much around.

155
00:16:35,030 --> 00:16:39,499
The risk of committing a new offence rather than the wider question,

156
00:16:39,500 --> 00:16:44,809
which is the health needs and often the mental health needs and substance misuse needs,

157
00:16:44,810 --> 00:16:51,500
which are the two main drivers, I think of suicide risk in that particular high risk population.

158
00:16:51,800 --> 00:17:05,590
Yes. Can I ask you about the tricky relationship, an association between mental illness and suicide and I suppose the treat ability.

159
00:17:05,600 --> 00:17:08,710
So what happens if you treat the mental illness?

160
00:17:08,750 --> 00:17:12,980
Does the suicide risk go away entirely? What's that? What's the link between the two?

161
00:17:14,180 --> 00:17:18,110
Well, I think part of it, it depends on the particular diagnosis.

162
00:17:18,710 --> 00:17:27,200
So I don't think it's easy to say whether there is a straightforward link between all mental illnesses or mental conditions,

163
00:17:27,200 --> 00:17:30,980
health conditions and suicide. Part of it depends on the diagnosis.

164
00:17:31,400 --> 00:17:36,559
And I think you've got a little bit of a hierarchy in the sense of which has the strongest effects,

165
00:17:36,560 --> 00:17:41,370
and it's usually the most severe end of the spectrum that have stronger effects.

166
00:17:41,410 --> 00:17:53,389
So by this I mean schizophrenia, spectrum disorders, bipolar disorder, they have the the the strongest effects with with suicide risk.

167
00:17:53,390 --> 00:18:02,450
And what we know, at least in bipolar disorder, from very good quality evidence, this is trial evidence, is that lithium in particular,

168
00:18:02,450 --> 00:18:13,329
which is commonly used in people with bipolar disorder as a medication to help treat the condition is associated with reducing suicide risk.

169
00:18:13,330 --> 00:18:18,530
And that's completed suicide, not just self-harm only, but but also completed suicide.

170
00:18:19,220 --> 00:18:28,760
For schizophrenia, it's less clear because there is some evidence and it's looked at mortality and it shows a reduction in mortality,

171
00:18:28,760 --> 00:18:38,240
overall mortality. But it's less clear because the trials haven't really had the length of time that's needed to follow people up.

172
00:18:38,540 --> 00:18:45,800
Sometimes. Charles don't even look for the outcome of the suicide risk, and sometimes they're not large enough to look at it.

173
00:18:46,220 --> 00:18:54,930
So in that case, you really need to triangulate that evidence with other forms of evidence using real world, what we call real world evidence.

174
00:18:54,930 --> 00:19:02,509
So you look at populations that are being treated with, let's say, antipsychotic medication and see what happens.

175
00:19:02,510 --> 00:19:09,860
There is some work done in Sweden and in Finland which shows that antipsychotics are associated with reduction suicide risk.

176
00:19:11,510 --> 00:19:14,690
So I think in some conditions we have trial evidence.

177
00:19:14,690 --> 00:19:20,600
In other conditions, we have what's what's called observational epidemiological evidence.

178
00:19:21,320 --> 00:19:22,940
In some conditions we don't really know.

179
00:19:24,200 --> 00:19:31,820
And you mentioned in your introduction you had difficulty with anti depressants making sense of that information.

180
00:19:32,690 --> 00:19:37,249
And the difficulty there is that when people are prescribed antidepressants,

181
00:19:37,250 --> 00:19:46,370
often because there's an acute increase in their suicide risk, and so that if you do see any people who self-harm or die from suicide,

182
00:19:46,370 --> 00:19:48,140
particularly early on in their treatment,

183
00:19:48,560 --> 00:19:54,340
it may be because they're on a trajectory which is towards increasing suicide risk because their illness is worse,

184
00:19:54,350 --> 00:20:01,819
a depression is worse and that's why they're getting the medication. So you can that's a very difficult area to disentangle.

185
00:20:01,820 --> 00:20:09,889
And there you really need to look at, you know, long term outcomes of people that are 92 presence.

186
00:20:09,890 --> 00:20:17,060
And there again, the trials are suggestive, but they you know, they they don't usually live long enough.

187
00:20:17,060 --> 00:20:22,130
They they're necessarily they're quite short term because they're looking at symptoms of depression.

188
00:20:23,540 --> 00:20:30,590
But overall, my impression from that evidence is that it's depressants are associated with a decrease in suicide risk.

189
00:20:31,310 --> 00:20:38,900
That's my impression. And but I think the difficulty is how you make sense of the first few months and the first

190
00:20:38,900 --> 00:20:43,640
few months is really complicated by lots of other things happening at the same time.

191
00:20:44,900 --> 00:20:52,020
So you will find some papers and some research findings to show a slightly increased risk in the first few months

192
00:20:52,020 --> 00:20:57,950
and others that trying to account for it and don't find that and some that actually find a reduction in risk.

193
00:20:57,950 --> 00:21:00,230
So you can find all three.

194
00:21:00,830 --> 00:21:09,350
And I think probably my view is that it's better to look longer term because then you can see a clearer pattern emerge, which is the decreased risk.

195
00:21:11,520 --> 00:21:15,870
Hello. I hope you're enjoying this episode of Future Makers Brain and Mental Health.

196
00:21:16,470 --> 00:21:22,470
If you'd like to learn more about our work here in Oxford, head to OCS, dot, AC, dot, UK,

197
00:21:22,470 --> 00:21:27,930
forward slash brain or let us know what you think on social media using the hashtag.

198
00:21:28,200 --> 00:21:34,259
Oxford Brain. Can we move on to the area of risk prediction,

199
00:21:34,260 --> 00:21:41,639
which is another really difficult area because I suppose on the one hand I would say

200
00:21:41,640 --> 00:21:48,000
this is absolutely a core part of what a mental health service should be doing,

201
00:21:48,000 --> 00:21:58,650
should be risk assessing and providing tailored individualised treatments that reduces people's risk and keeps them safe and keeps other people safe.

202
00:21:59,160 --> 00:22:08,219
It's a very difficult thing to do and it's a you can get in our area with other mental health clinicians and researchers.

203
00:22:08,220 --> 00:22:15,150
You can get people getting quite shouty about it, can't you, about whether they're actually useful, whether we should throw them all out the window.

204
00:22:16,020 --> 00:22:27,780
What's your view? Well, I think, um, I think that, um, my, my view is they have a role to have a role to support, to enhance clinical decision making.

205
00:22:28,920 --> 00:22:35,520
And I think that very much aligns with the rest of medicine where we use these sort of tools in,

206
00:22:35,520 --> 00:22:40,530
let's say, cancer medicine or cardiovascular medicine to support clinical decision making.

207
00:22:42,090 --> 00:22:49,490
We know that clinicians are not perfect at this sort of thing, that they are necessarily optimistic.

208
00:22:49,500 --> 00:22:53,190
They have to be, I suppose, in order to work in this quite difficult area.

209
00:22:54,090 --> 00:22:57,540
And so they tend to underestimate risk.

210
00:22:57,840 --> 00:23:03,809
They tend to rely on the most recent things that have happened to people rather than look at the sort of longer term perspective.

211
00:23:03,810 --> 00:23:08,730
And again, that's just a bias that human beings have on many levels when it comes to decisions.

212
00:23:09,240 --> 00:23:13,800
And there's a bunch of other biases also that human human beings do have.

213
00:23:14,580 --> 00:23:19,680
And I think then having tools which are structured provides some consistency.

214
00:23:20,010 --> 00:23:30,330
They can sort of help pull up the sort of basic level of quality of risk assessment and then allow

215
00:23:30,340 --> 00:23:35,130
for communication between different services because you're speaking about that same sort of thing.

216
00:23:36,780 --> 00:23:42,900
And they highlight important risk factors which sometimes you forget about to ask.

217
00:23:43,320 --> 00:23:47,610
And they also, I think, underscore what we call safety planning.

218
00:23:47,610 --> 00:23:54,630
And safety planning is where people come up with a a plan about what they would do if they have an acute crisis of some sort.

219
00:23:55,440 --> 00:24:00,239
So particularly the acute suicidal crisis, you know, who they would contact, what they would do in their midst,

220
00:24:00,240 --> 00:24:06,360
let's say, if they were medication, what other support they would try and use.

221
00:24:06,630 --> 00:24:09,870
So I think suicide risk prediction has a role.

222
00:24:10,950 --> 00:24:21,779
And I think a lot of the previous work, which has dismissed it, is based on really old times, old ways of doing things.

223
00:24:21,780 --> 00:24:30,329
So for instance, the most common way people used to do this is by using a checklist which was developed for another purpose.

224
00:24:30,330 --> 00:24:39,690
So for instance, you would use a depression symptom checklist and you would apply it to someone who you're worried about their risk of suicide.

225
00:24:40,140 --> 00:24:44,340
And then you test whether it does a good job at predicting what happens in the future,

226
00:24:44,340 --> 00:24:48,210
let's say future self-harm episodes or even future suicide risk.

227
00:24:48,510 --> 00:24:48,990
And of course,

228
00:24:49,170 --> 00:24:55,739
a depression checklist is was developed 20 years ago for the purpose of telling you how severe depression is isn't going to be very good.

229
00:24:55,740 --> 00:24:59,760
It's a completely different purpose, which was never developed for.

230
00:25:00,240 --> 00:25:08,129
And so a lot of the evidence on this question, a lot of the opinions, including this actually is about those type of tools,

231
00:25:08,130 --> 00:25:15,790
which I think quite rightly are not useful because the tools that were developed for other purposes.

232
00:25:16,230 --> 00:25:24,540
Nice. Nice being the National Institute of Clinical. And yes, so there are a committee of experts who review the evidence.

233
00:25:24,540 --> 00:25:32,639
And and I think that their review of the evidence is right when it comes to these old checklist tools which are poor quality,

234
00:25:32,640 --> 00:25:37,350
they're not validated properly. They wouldn't even develop for the purpose of suicide risk prediction.

235
00:25:37,830 --> 00:25:41,670
So they're right in the sense that, you know, we need to start again.

236
00:25:42,300 --> 00:25:44,879
And one of the things I've been trying to do is start again, actually.

237
00:25:44,880 --> 00:25:52,110
So I've I've developed a number of tools that can be used to inform suicide risk prediction.

238
00:25:52,110 --> 00:25:56,220
And we've developed the tool for people with severe mental illness called locksmiths,

239
00:25:56,820 --> 00:26:01,950
and just published a tool for people who self-harmed could all access.

240
00:26:02,700 --> 00:26:15,750
And then we we just use very modern methods, large data sets transparently reported, easy to score, easy to interpret,

241
00:26:16,650 --> 00:26:26,129
presenting the information in different ways, not just high and low risk, which is very old, I think an old fashioned way of doing it.

242
00:26:26,130 --> 00:26:29,160
But actually probability schools, which is a more modern way of doing it,

243
00:26:29,760 --> 00:26:37,319
akin to having like a probability score of, let's say, dying from another course.

244
00:26:37,320 --> 00:26:41,460
Let's say if you've got a cancer diagnosis, you're given a five year survival,

245
00:26:42,570 --> 00:26:49,170
or if you over the age of 50, in many high income countries, you get a cardiovascular risk score.

246
00:26:50,490 --> 00:26:54,120
So that's the way that things have been moving in the rest of medicine.

247
00:26:54,120 --> 00:26:59,099
I think mental health has been a bit slow to catch up and that's part of our research programme.

248
00:26:59,100 --> 00:27:02,589
And I think I mean, we we know that there's a lot of interest in this,

249
00:27:02,590 --> 00:27:08,669
so we can see from the website traffic on our website where we've got these tools up that we have quite a lot

250
00:27:08,670 --> 00:27:17,820
of take up of these tools and people are using them widely for research purposes as well as to help them.

251
00:27:18,150 --> 00:27:24,389
Think about groups of patients that they may see as well and indicated treatments as a result.

252
00:27:24,390 --> 00:27:30,270
So if you have an increased risk of cardiovascular events, you might start a statin or you might stop smoking.

253
00:27:30,480 --> 00:27:36,180
If you have an increased risk of suicide, you might have help for your alcohol is a problem.

254
00:27:37,320 --> 00:27:44,250
And the arguments against that which which some of the some of the people who get some of you have written about this saves

255
00:27:44,550 --> 00:27:52,080
everyone should get the gold standard so that even if you end up with stratifying risk which is a more technical term,

256
00:27:52,470 --> 00:27:57,240
then that doesn't help you because we should be giving everyone the gold standard.

257
00:27:57,510 --> 00:28:03,479
But I don't think that's realistic. I mean, in a modern healthcare system, you can't give everyone the gold standard.

258
00:28:03,480 --> 00:28:09,420
I mean, everyone that turns up tonight in the John Ratcliffe Hospital, you know, who've taken has taken an overdose,

259
00:28:09,420 --> 00:28:18,180
can't get any sort of gold standard follow up, you know, ten sessions of cognitive behavioural therapy plus, you know, good quality follow up.

260
00:28:18,180 --> 00:28:21,569
And that's just not realistic in most countries in the world.

261
00:28:21,570 --> 00:28:25,880
It's not realistic. And so I think treatments have to be tailored.

262
00:28:25,890 --> 00:28:32,910
They have to be made more precise, but they also necessarily, I think, have to be there has to be some resource allocation and.

263
00:28:33,860 --> 00:28:38,080
One of the useful things about these tools is they can tell you who are the people.

264
00:28:38,090 --> 00:28:46,340
You don't need to do further assessments and further long, you know, treatments and assessments for.

265
00:28:47,240 --> 00:28:56,480
And so you can focus your energies on the group that does does have identified needs and will benefit from further assessment and treatment.

266
00:28:56,750 --> 00:29:03,890
Yeah. And if you are in control, if you are the Chancellor seen and you had your defined budget.

267
00:29:04,070 --> 00:29:06,710
Where would you put it for the most benefit?

268
00:29:06,740 --> 00:29:15,530
I'm just wondering, for instance, for people in prisons, for instance, would you give them a better quality mental health care?

269
00:29:15,950 --> 00:29:24,409
Well, I think I think from a if I was the chancellor, I think a lot of my attention would be on actually funding trials,

270
00:29:24,410 --> 00:29:29,629
because one of the things that struck me and I did a review is that the field for Internal Medicine,

271
00:29:29,630 --> 00:29:40,130
which is a well-known journal in our field, and when I did this review, I mean, I looked very hard for trials of suicide prevention.

272
00:29:40,910 --> 00:29:46,130
And I was surprised and disappointed, actually, that there weren't more trials.

273
00:29:46,160 --> 00:29:54,799
I mean, what seems to be the case is there are many, many trials looking at outcomes such as suicidal thoughts or ideas or even self-harm.

274
00:29:54,800 --> 00:29:58,970
And we know that they're quite different. So the risk factors are different.

275
00:30:00,020 --> 00:30:05,420
What drives people to to self-harm can be very different to what drives suicide mortality.

276
00:30:06,980 --> 00:30:15,050
And I think actually good quality evidence is actually lacking on the most effective treatments for suicide prevention.

277
00:30:15,530 --> 00:30:19,850
So that's where I think. So I would I would put a big investment in the R&D budget.

278
00:30:20,720 --> 00:30:24,860
Oh, thank you very much. Yes, Senior Chancellor.

279
00:30:25,250 --> 00:30:30,490
And I suppose that sort of connects to that. It is surprising to me how it's such a lead.

280
00:30:30,500 --> 00:30:33,770
I mean, this is a public health crisis, actually.

281
00:30:33,770 --> 00:30:38,910
Suicides particularly, You know, it's the leading cause of death in young people.

282
00:30:38,930 --> 00:30:48,050
Absolutely. Yes. And globally, I mean, so 70%, 77% of global suicides occur in low and middle income countries where,

283
00:30:48,500 --> 00:30:54,080
you know, absolute parlous level of research going on into how to improve this situation.

284
00:30:54,320 --> 00:30:58,010
But I mean, and it's linked to the sort of the stigma around it, isn't it?

285
00:30:58,300 --> 00:31:04,310
I don't it's not part of I don't think and sort of people that I talk to that it's seen as a,

286
00:31:04,720 --> 00:31:10,100
a solvable problem as it's something that actually research can provide the answers to.

287
00:31:10,550 --> 00:31:16,070
Maybe that is true. I mean, maybe people don't realise how preventable it really is and that, you know,

288
00:31:16,340 --> 00:31:22,790
these some of these interventions that have happened, you know, you know, simple interventions,

289
00:31:22,880 --> 00:31:27,860
know putting barriers up in bridges to buildings, railways, you know,

290
00:31:27,920 --> 00:31:34,270
making it more difficult to jump from high buildings, putting restrictions on pesticides.

291
00:31:34,280 --> 00:31:39,259
I mean, they're three that there is research, good research to show that they've made big impacts,

292
00:31:39,260 --> 00:31:46,760
actually, so that you can actually make an impact on reducing suicide rates.

293
00:31:46,790 --> 00:31:53,290
So it is something that is preventable. I suppose that's one of the messages that it's not inevitable, it is preventable.

294
00:31:53,630 --> 00:31:58,960
And and I think particularly in high risk groups, we know there are things that can be done.

295
00:31:58,970 --> 00:32:05,209
There is some evidence in people with with underlying mental health and physical health problems

296
00:32:05,210 --> 00:32:10,040
and in certain other high risk groups like people who touch the career justice system,

297
00:32:10,280 --> 00:32:12,439
people who self-harm like we just mentioned.

298
00:32:12,440 --> 00:32:18,260
I mean, they're another really important group because their subsequent rates of dying from suicide are really very high.

299
00:32:19,370 --> 00:32:24,200
And one of the highest groups in terms of humans just take a subgroup that is the highest,

300
00:32:24,230 --> 00:32:30,290
you know, I think in terms of risk in the next year and all those groups,

301
00:32:30,290 --> 00:32:33,649
something, there are interventions that are available,

302
00:32:33,650 --> 00:32:40,490
but I think there's actually a lot of room for more good quality research to highlight other things that can be done.

303
00:32:40,790 --> 00:32:49,610
We don't know a lot about the best services, so mental health services, how they can be better provided, better reconstituted.

304
00:32:50,060 --> 00:32:53,240
You know, you work in first episode services.

305
00:32:53,480 --> 00:33:00,469
I mean, we know it was very interesting study done in South Korea where they looked at the introduction of these services and they showed that

306
00:33:00,470 --> 00:33:06,590
there was a reduction in suicide rates among that population will be very interesting to those that's validated elsewhere in the world.

307
00:33:07,850 --> 00:33:16,610
So there is I think, yeah, for me, the striking thing is that we know now quite a lot about risk factors.

308
00:33:17,000 --> 00:33:23,900
We know a little bit more about the high risk groups, but we don't know a lot about interventions.

309
00:33:25,070 --> 00:33:30,580
And that's, I think where a lot of research effort should be placed.

310
00:33:30,590 --> 00:33:34,390
I think the other thing I think we could do better on is. Assessment, actually.

311
00:33:35,470 --> 00:33:43,810
So I talk about risk factors, but actually what is the case is that it's accumulation of risk factors.

312
00:33:44,140 --> 00:33:50,800
They act at different points in one's person's life, in different strengths.

313
00:33:51,760 --> 00:33:55,480
A clinician can only probably cope with three or four risk factors simultaneously,

314
00:33:55,840 --> 00:34:01,000
but these tools can bring together, you know, ten or 15 simultaneously.

315
00:34:01,390 --> 00:34:05,770
And that's the added value of them, really. You know, when I'm seeing someone in clinic,

316
00:34:05,770 --> 00:34:14,140
I can just about three or four factors together and hold them together and make a judgement about the relative, you know, impact of these factors.

317
00:34:15,130 --> 00:34:21,820
But then anything beyond that, you know, I run out of the sort of computing power and that's where I think,

318
00:34:21,820 --> 00:34:27,910
you know, and, and therefore, you know, there may be developments in artificial intelligence, machine learning.

319
00:34:27,910 --> 00:34:31,629
And we hear a lot of hype about this, and most of it is hype, to be honest.

320
00:34:31,630 --> 00:34:41,290
But if it's done well, I mean, it may also have a role. I mean, how scalable it is remains a question, but I think it needs to be studied.

321
00:34:42,160 --> 00:34:43,330
We need to see what, you know,

322
00:34:43,470 --> 00:34:49,660
what added benefit it does give you and whether that and then we have to decide whether it's worth it with the benefits,

323
00:34:49,660 --> 00:34:53,830
worth it in the in the sense is it worth it because it will cost a lot of money to implement.

324
00:34:53,830 --> 00:35:04,659
But also, you know, it's not as transparent as having the type of tools I've been working on where you can see all the risk factors,

325
00:35:04,660 --> 00:35:09,520
what impact they have, how they interact. Since very transparent open system.

326
00:35:09,520 --> 00:35:12,580
You can see how we developed it. You can see what risk factors we tested.

327
00:35:13,720 --> 00:35:22,570
We can see you can see our methods. But the these new methods to do machine learning, it's it's much more difficult to interpret.

328
00:35:24,550 --> 00:35:35,830
So they may well have a role. And there I think we would have to work quite closely with, I think, medical ethicists, to be honest,

329
00:35:36,100 --> 00:35:43,420
to try and figure out how much added value they can provide and whether that's actually

330
00:35:45,040 --> 00:35:48,909
it's really valuable in the big scheme of things and the acceptability of that,

331
00:35:48,910 --> 00:35:52,360
actually. Yeah, that's a whole different podcast, I think. I think.

332
00:35:52,540 --> 00:35:55,760
Well, though, thank you. Seen a fascinating discussion.

333
00:35:55,780 --> 00:36:04,820
I think that's a good place to leave it. Great. Thank you. I hope you enjoyed this episode of Future Makers Brain and Mental Health.

334
00:36:05,330 --> 00:36:12,590
You can find more episodes of future makers wherever you get your podcasts and more on Oxford's research at Oxford.

335
00:36:12,620 --> 00:36:16,820
AC DC's UK Forward Slash Brain. Thanks for listening.