Yale Psychiatry Grand Rounds: April 9, 2021

April 09, 2021

Information

"Socioeconomic Adversity and Deaths of Despair"

Mark Olfson, MD, MPH, Elizabeth K. Dollard Professor of Psychiatry, Medicine, and Law; Professor of Epidemiology, CUMC, Columbia University, New York State Psychiatric Institute

ID6408

To CiteDCA Citation Guide

00:00And send my regards to Mark Mark's
00:03Mark and I met on the interview trail
00:08for residency and were resident
00:11classmates here at Yale and have been
00:15good friends ever since that time.
00:18So it's really wonderful.
00:20Plus, as everybody on the call knows,
00:24his major contribution to the
00:27field of psychiatry is parenting.
00:30Emily are wonderful.
00:36Trainee here at Yale Emily Austin,
00:38who's who's been a wonderful
00:40addition to our Department as well,
00:43but I did want to share some very good news,
00:47which is that U.S.
00:48news and World Report issued their rankings
00:52of medical school departments of psychiatry.
00:55Which in which Yale Place #1.
01:00So. You know, these polls are
01:05always of of somewhat uncertain.
01:09Meaning and value,
01:10but it's always nice to be #1
01:13and so congratulations everyone.
01:15This is obviously an achievement of a
01:18community and and a mark of of really
01:22the incredible work that's going on here.
01:25So congratulations to everybody and Bob,
01:27I'll pass it to you to introduce Mark today.
01:31OK, so it's a pleasure
01:34to introduce Mark Olson.
01:36Who is Elizabeth K Dollard professor
01:38of psychiatry at Columbia and
01:41among his many accomplishments,
01:42he's listed as one of the most important
01:46scientific contributors in the world
01:48by some award system an I thought
01:51I'd say a little bit about Mark.
01:53He came from Chicago.
01:55And then traveled West to Reed College,
01:59where he missed Steve Jobs by a few years.
02:03He was a little after Steve Jobs,
02:06but he was just telling us how
02:09there was shenanigans going on
02:12near his room from the job's orbit.
02:15Steve Jobs, as you know,
02:17when East after after staying it
02:19read for six months when East to
02:22India to have his Guru experience.
02:24Mark also went E 1st to Chicago,
02:26stopping off briefly to get an
02:28MD and then finally arrived at
02:30Yale where he was a resident.
02:32One of our best residents at the time.
02:37And then he embarked on a remarkable journey
02:40where he went to Columbia to get an MPH,
02:44and then a postdoc with
02:46David Mechanic at Rutgers.
02:48And I would say he is now the leading mental
02:51health services researcher in America.
02:54And I would also say,
02:56I think,
02:57with not much exaggeration,
02:58that he knows more about mental health
03:02care in America than anyone else.
03:05And I would say both substantively,
03:08he has contributed to so many
03:11areas documenting the expansion
03:13of the use of antipsychotics.
03:15The limited use of ECT.
03:17He's followed the impact of insurance
03:20on the use of buprenorphine.
03:22He studied suicide nationwide,
03:24showing its limited use studied the
03:27impact of discharge planning on outcomes,
03:30the Association of Psychiatric
03:32diagnosis with deaths from Covid.
03:35And those are just his papers
03:37that are in press.
03:38Let alone the other 500 that are
03:40that have already been published,
03:42and that those who read.
03:44Of what's now called JAMA Psychiatry.
03:47Get to see what he's doing every every month.
03:52I just want to use this as a teachable
03:55moment to say something about.
03:57Mental health services research,
03:59which is the study of how
04:02services are actually delivered.
04:05It's a merger of psychiatry,
04:08mental health and Epidemiology.
04:10Anne relies on a 30,000 foot
04:14perspective of the whole country.
04:17And Mark has mastered this an I
04:20think is the leader in this area.
04:24And without further ado,
04:25I pass it on to him to talk about
04:28socioeconomic adversity and depths
04:30of despair. Take it away, Mark.
04:32Well,
04:33thank
04:33you Bob for that exceedingly generous.
04:37Introduction I mean, I'm tempted
04:38even not to talk after that one,
04:41but just like I said,
04:42but thank you so much, Bob and John.
04:45Congratulations on the number one ranking
04:47in in the US News and World Report.
04:50Psychiatry. That's that's really
04:51an impressive accomplishment,
04:52but not a surprising one.
04:54Given the many strengths of your Department.
04:56So it's, you know it's very nice
04:58to be be here with everyone.
05:00And as Bob mentioned,
05:02many years ago, I was a resident.
05:04Yeah, I had the really good fortune
05:06to receive great mentoring then.
05:08And it it gave me the confidence
05:10to pursue my developing interests
05:12in in medical services,
05:14research and and so now many years
05:17later is John has mentioned,
05:19you know, I my daughter Emily.
05:22I've had the gratifying vicarious
05:23experience through her of as
05:25she's received really wonderful
05:27mentoring from Tom Fernandez in for
05:30Michael Black from Larry Vigilon,
05:32Owen and many others within your Department.
05:34So I know that that old tradition
05:38of supportive mentoring.
05:39Is you know he's alive and well at.
05:41You know psychiatry.
05:42And actually before I get to my truck,
05:45I wanted to dress for a moment on
05:47this theme of mentoring because
05:48I've entered a stage of my career
05:50where measuring it really has
05:52become quite important to me.
05:56Those of you may have learned in about the
05:58ancient Greeks in high school or college.
06:01You may remember the word
06:02mentor descends to us.
06:03Actually from the Odyssey when Odysseus,
06:05the King of Ithaca, was preparing to
06:07set sail and joining the Trojan War,
06:10and he knew it would take him away
06:12from his family for several years.
06:14He turned to his old friend mentor and
06:16you trust him to care for his young son.
06:19Telemachus. Now in the Odyssey
06:21there's a Greek word called napeo's,
06:23which means disconnected,
06:24that is used to describe.
06:25Telemachus, and it turns out it
06:27turns out that Telemachus was
06:29disconnected from his ancestral past,
06:31so when in his father's absence the
06:34suitors started to pillage their estate,
06:36and so his mother, Queen, Penelope's,
06:38hand in marriage to him because
06:41he didn't really get it.
06:42He didn't really appreciate the gravity
06:45of this threat, but Fortunately,
06:46mentor, actually it's it's Athena,
06:48in the guise of mentor.
06:50Appears to Telemachus Anna part imparts
06:52to him that great Greek virtue of menace.
06:56Meanness,
06:56which means putting wisdom into action.
06:58Now I'm not gonna,
06:59I'm not gonna ruin the rest of this story.
07:01For those of you who haven't read the
07:03Odyssey or haven't read it in many years.
07:06But my point is the great mentoring,
07:08then,
07:08is now involves giving people
07:10the wisdom to become effectively
07:11engaged in the world around them.
07:13And so after all these years,
07:15and it's been quite a few,
07:18I'm still grateful for the mentoring
07:20that I received as a resident yell
07:22from Bob and from many others.
07:25So please, except for me a very,
07:27very belated thank you Now this morning,
07:30as you see from my title slide here,
07:33I'm going to be discussing connections
07:35between socioeconomic adversity.
07:37On the one hand, in deaths of despair,
07:39by which I mean suicide,
07:41drug overdose in chronic liver disease,
07:43on the other,
07:43because of my interest in
07:45Epidemiology and medical services,
07:46I'm going to be showing you many,
07:48many numbers this morning, but no,
07:50but I want you to know this.
07:54These numbers can tell us what it feels
07:56like to be poor or to be unemployed.
07:58They can't tell us what it feels
08:00like to be in a suicidal crisis,
08:02or to lose a loved one to drug overdose.
08:06At at past I,
08:07I hope that my presentation and
08:09will give you a sense and maybe a
08:12quantitative sentence only of how
08:14important it is to consider the
08:16difficult socioeconomic circumstances.
08:18Under which many of our patients
08:21live out their lives. So here is.
08:26Here's a list of my potential conflicts.
08:29These are organizations, Playstations,
08:32from which I've received income
08:35over the past few years now.
08:38I will consider for related themes
08:40that you see on this slide here.
08:43First,
08:43I'll briefly provide some historical
08:45context for the general topic.
08:47Then I'll review some recent analysis
08:49of associations between markers of
08:51socioeconomic adversity and risks
08:53of death from suicide poisoning
08:55and chronic liver disease course.
08:57Most of those chronic liver disease
08:59deaths involved alcohol use disorder,
09:01and then I'm going to consider
09:03the extent to which overdoses
09:05overdose deaths involving opioids
09:07are actually suicidal intent.
09:09And finally, I'm going to close with
09:11a few words about some work that
09:14I'm hoping to carry out over the
09:16next several years now in awareness
09:18that health is intimately linked to
09:21social and economic circumstances.
09:22It wasn't introduced by Anne Case
09:24and Angus Deaton and their well
09:26known article in the Proceedings
09:28of National Academy of Sciences
09:30Back in 2015 on rising mortality
09:33among middle aged white Americans.
09:34Many of the concepts undergirding when I
09:37call their deaths of despair hypothesis.
09:40Have deep historical roots and sociological
09:42sociological research and have even
09:44found their way into popular culture
09:46over the past couple of centuries.
09:48In Europe and here in the United States.
09:52So let's begin with the
09:54consideration of this man here,
09:56Rudolf workout.
09:57Now he was he was a German physician.
10:00And there's a part he was a pathologist,
10:03a politician, a public health
10:04reformer and a social activist,
10:06and he had very.
10:07He's also an outspoken advocate for the poor.
10:11His writings are filled with
10:12recommendations on how to improve
10:14people's health by improving their
10:16economic and material circumstances.
10:17He advocated things like rationalizing
10:19the food distribution and
10:20reforming the educational system.
10:22He wrote about expanding the political
10:24enfranchisement and reducing military
10:26expenditures which irritated the
10:27heck out of Chancellor Bismarck
10:29and actually lead almost to a dual.
10:31There's a funny story about
10:33that that if we have time,
10:35I'll tell you,
10:36and also a variety of other changes
10:38in social policy,
10:39all in the name of improving health.
10:43And behind Virchows thinking was this
10:45concept that adverse economic conditions.
10:48Are the first or fundamental cause
10:50of many of the medical diseases
10:52that he saw in his practice,
10:54and therefore the practice of medicine
10:57involves seeking to improve the social
10:59and economic realities of patients lives.
11:02So here,
11:02here are some quotes,
11:04a couple of quotes that give you a
11:06sense of his visionary aspiration's.
11:08He said once he met Medicine,
11:10established this anthropology and once
11:12the interests of the privilege no longer
11:15determine the course of basic events,
11:17then the physiologist and practitioner
11:18will be counted among the elder statesman
11:21who support the social structure.
11:23Medicine is a social science
11:24in its very bone and marrow,
11:26and he goes on in this same essay
11:29with the famous lines that are
11:31often repeated associated with them.
11:33Positions are the natural attorneys
11:35of the poor and social problems
11:38largely fall within their scope.
11:43But my point, my broader point is it
11:46wasn't just medical scientist rate.
11:48Medical scientists like Virgo but also
11:50men and women in the arts and letters
11:52who long drawn connections between
11:54social economic circumstances and health,
11:56particularly between things like poverty
11:58and alcoholism as well as financial
12:00crises or romantic losses and suicide.
12:02One can think of like the sorrows of Werther.
12:05Here, a couple of illustrative quotes
12:07from 2 great 19th century novelists,
12:09a young Charles Dickens writes of
12:12the allure of alcohol is a bomb.
12:14The misery of poverty rates pieces
12:16maybe actually from his sketches,
12:18opposes the first thing he published his
12:21young man until you improve the homes of
12:24the poor or persuade 1/2 Amish stretch
12:26not to seek relief in the temporary
12:29oblivion oblivion of his own misery.
12:32Gin shops will increase in number
12:34and splendor, and then it below.
12:36There used to quote from emails.
12:39So let's masterpiece lesson water.
12:41And yeah,
12:41in which he portrays the fate of survey.
12:45Alondras from working class Paris who
12:47struggles to work her way up into the
12:49middle class and eventually to own
12:51and operate their own small laundry,
12:53only to suffer financial crisis
12:55when her husband is injured.
12:56Work and then both of them descend
12:58into poverty and alcoholism and
13:00near the end of the novel,
13:02the narrator opines they never
13:04knew exactly what she died of.
13:05Some people spoke of a chill,
13:07but the truth is was that she died
13:10of poverty from the filth and
13:12wretchedness of her weary life.
13:13She rotted to death.
13:17So these quotes from bestselling
13:19novelists suggests that connections
13:21between poverty and alcoholism were
13:23part of the public's awareness.
13:24In the 19th century that case, indeed,
13:27it is all described momentarily in our day.
13:30They don't exactly argue that poverty per,
13:33say, drives alcoholism, alcoholism,
13:35and the other deaths of despair,
13:37but rather that these threats to health
13:40are connected with its wording of an
13:42individual's needs and expectations,
13:44and that's an idea that.
13:46Was foreshadowed in the academic suicide
13:48literature and the work of Emile Durckheim,
13:51who was one of the founders
13:53of modern sociology.
13:54Durkheim wrote about the imbalance between
13:56personal needs and available resources.
13:58Here's a here's a quote, actually,
14:00a couple of quotes from his chapter on anomic
14:04suicides from his famous book on Suicide.
14:07Uh, the late 1800s.
14:09He said he wrote that no living being can
14:11be happy or even exist and less his needs
14:14are sufficiently proportion to his means.
14:17And in other words,
14:18if his needs require more than can be granted
14:21or even merely something of different sort,
14:23they'll be under continual friction
14:25and can only function painfully.
14:26And then he goes on.
14:28And here he's rebutting the idea
14:30that we can understand suicide
14:31simply by looking at poverty levels.
14:33He says, what proof?
14:35Still more conclusively,
14:36economic distress does not have the
14:38aggravating influence often attributed to
14:40it's not saying just economic distress.
14:42Is that very little suicide in Ireland?
14:44There's very little suicide in
14:46Ireland with the peasantry police
14:48arrest a life and it goes on to
14:50talk about collaborating in Spain,
14:51France and a variety of other places.
14:54So here's a nice.
14:55Also a nice illustration by the way
14:57of using statistical epidemiologic
14:58data very early to make the case that
15:01suicide rates can't be explained simply
15:03by appealing to rates of poverty.
15:06So.
15:08So Durckheim is arguing for focus
15:10on desires and expectations and
15:12the risks that occur when they're
15:14frustrated out of balance.
15:15And as I mentioned,
15:16we kind of hear an echo of this idea in
15:19case and Edens Dessa despair hypothesis,
15:22people despair when their material
15:24and social social circumstances
15:25are below what they expected.
15:27This despair leads people to act in
15:29ways that simply harm their health,
15:31and this may have a direct impact on
15:33death through suicide and indirect impact.
15:35Through heavy drinking smoking.
15:37And alcohol abuse and it rude is
15:41economic and social breakdown now.
15:46It's this morning, a personal expectations.
15:48The case, indeed, move you as really
15:51as as as part of the mechanism.
15:54That sort of leads to despair.
15:57The what they write about the erosion
15:59of longstanding economic opportunities.
16:01Again, the erosion of longstanding
16:03economic opportunities.
16:04People have these opportunities,
16:06and they've slipped away and that that
16:10process plays a prominent role in their
16:13view in precipitating these deaths.
16:16Of despair.
16:16So if we think about this this
16:19concept schematically,
16:21it might look something like this.
16:25Economic stagnation driven by globalization,
16:27rapid technological change,
16:28loss of union power that shielded
16:31workers in the US from competition,
16:33low wage labor both here and abroad,
16:35has over a period of time,
16:38narrow job prospects and accelerated
16:39declining rates of marriage that
16:41contributed to social isolation and
16:43loneliness and these processes,
16:45in turn, may contribute to something
16:47that we might think of.
16:49Diseases of despair,
16:51and ultimately to deaths of despair.
16:54And of course,
16:55over the last year or so,
16:56one would think that all of this has been
16:59exacerbated by the economic disruption
17:01and the loss of 20 million jobs.
17:03Only about 10 million of which
17:05should come backed.
17:06It's been brought on by the covid pandemic,
17:08and if we have time,
17:10we can talk about the CDC's release
17:12of numbers, preliminary numbers,
17:14provisional numbers just last week,
17:15showing that in fact we we didn't
17:17in 2020 have an increase in
17:19suicide as many people had feared.
17:21But,
17:22but that's the that's sort of the general.
17:25As I say,
17:26a schematic view of the case and Deaton.
17:30Hypothesis,
17:31and at the very center of
17:33their hypothesis is this idea,
17:35or this notion of despair.
17:36And yet nowhere in their writings.
17:39And I've read the I read their
17:41book in their articles.
17:43Do we encounter a measure of despair?
17:45So so there's real empirical work to
17:47do if we're interested in this area,
17:50I work that I'm I'm not skilled,
17:53I have the skill set to pursue,
17:55but I would encourage others to to do,
17:58and that is work that involves
18:00kind of understanding.
18:01The psychopathology one could
18:03imagine a small field.
18:05Devoted to understanding the
18:06psychological consequences of
18:07financial strain and social isolation,
18:09and that that how they those
18:11things might mediate maladaptive
18:12behaviors that drive Destin despair.
18:14And it might in fact be somewhat similar
18:16to what is already well developed.
18:19Field of depression and and in that area.
18:22Of course we think in terms of not only
18:24behavioral and cognitive emotional,
18:26but also biological.
18:28Aspects that may bear and be relevant
18:31for understanding the etiology
18:34or onset of how these financial
18:37strain is used is experienced by
18:40people and results in putting that
18:43increased risk for these diseases
18:46that can be life threatening.
18:49Backing up one reason that this
18:51topic is so important is that
18:53over the past two decades or so,
18:55and I'm the last couple of years, well,
18:57you probably got to put an asterisk by,
18:59but if we step back and look it over the
19:02last, you know 20 years or so we've seen
19:04rising rates of suicide and drug overdose,
19:07especially of course opioid overdose deaths,
19:08chronic liver disease deaths, United States.
19:10There are all trending upward and that
19:12it stands in quite sharp contrast to the
19:14impressive declines that have been achieved
19:16in many of the other leading cause to death,
19:19some of which are on the slide.
19:21Here, things like ammonia and cancer.
19:23In HIV. So if we had a better
19:27understanding of the connections.
19:29Between the markers of social economic
19:31adversity and S to despair that might
19:34well have implications for the broad
19:36direction of social and health policy.
19:38If we were to find strong associations,
19:41it might help make the case that upstream
19:44interventions with the sorts that
19:46Doctor Burke I was writing about over
19:48100 years ago and advocating in Germany.
19:51And we recognize today is things
19:53like raising the minimum wage,
19:55increasing public support for job training,
19:57expanding investment in public education,
19:59loan forgiveness.
19:59Increasing the veil.
20:00Health insurance for low income
20:02people subsidizing child care.
20:04You know there's a?
20:05There's a long list,
20:06in fact,
20:07some of these policies resemble those sorts
20:09of measures that were recently included
20:12in the in this Big American rescue plan.
20:15Nearly $2 trillion plan that aims
20:17to improve the financial security
20:18of millions of people who both low
20:20and even middle income individuals
20:22in United States through things like
20:25extending unemployment benefits,
20:26providing direct payments,
20:27funding emergency pay leave,
20:29and several other measures.
20:30So again.
20:31This if you can demonstrate if one
20:33can demonstrate strong associations,
20:35it'll have important implications
20:37for social policy.
20:38On the other hand,
20:39if the week it if the linkages are weaker,
20:42it might suggest that the traditional
20:44mental health focus the narrower
20:46downstream focused on things like
20:48suicide screening and linking
20:49people to mental health services
20:51and evidence based treatments,
20:53building up and making more available
20:55crisis lines and trying to pass
20:57firearm safety policies and try to
20:59implement counseling safety planning
21:01with an emergency departments and
21:02expanding access to Medicaid.
21:04Medication treatments like
21:05buprenorphine for opioid use disorder,
21:07needle exchange programs,
21:08the locks,
21:09own availability so that people can
21:11be get reversed and saved in field.
21:13You know all these more sort of
21:16what I think of as downstream.
21:19Approaches and interventions would be,
21:21you know,
21:22we play a more central role if if the
21:25weaker associations between these
21:28broader socioeconomic determinants.
21:30So anyway to examine this issue.
21:34I worked with my friend Carlos Black.
21:37I was at night at and Melanie Wallop,
21:40Asstastic Colombian Sean Altekruse,
21:42who's at the NHL BI on a rather
21:45straightforward set of analysis and
21:47I will walk you through some of them.
21:50They're based on something called
21:52the mortality disparities in American
21:54communities data or EM back for short.
21:56And really,
21:57all it is,
21:58is taking the American Community
22:00Survey it from 2008.
22:01Huge survey representative sample
22:03of Americans.
22:044,000,000 which you see
22:05there and linking it to the
22:07National Death Index.
22:08It's been linked through 2015.
22:10Over 300,000 people died from that
22:12cohort of four and a half million,
22:14and in about a month or so,
22:16the Census Bureau is going
22:18to be releasing an update.
22:19So we'll be linking it through
22:212019 an you know that will give
22:23us more power and get a chance
22:25to look at the effects as the
22:27opioid epidemic really took off.
22:28Unfortunately,
22:29during that period between 2016 and 19,
22:31so I think there are results that
22:33were about to show you may change
22:35in a couple of months a bit.
22:37But still it's a.
22:39It's a very.
22:40It's an impressively large survey and
22:41just little bit about this survey.
22:43The American Community surveys,
22:45as I mentioned,
22:45fielded by the Census Bureau,
22:47and because it's legally required
22:49you got in the Mail.
22:50I happen to have got one last year.
22:54You know they've got this truly
22:55impressive response rate of 97.9%.
22:57You never see that in the literature in it.
22:59Any other context,
23:00and it sends a very broad sampling frame,
23:03so it's got people who are in
23:05regular housing units as well as
23:06wide range of residential facilities
23:08you see listed on the slide.
23:10There is a little bit of an imperfection
23:12in the linking to the National Death Index,
23:14so we're about 9:00 or so percent
23:16that can't be linked,
23:17and it's usually because an individual
23:19doesn't have a Social Security number,
23:21or it's these copied.
23:22Most of that occurs there and then to try to.
23:25Correct for that imperfection,
23:26they do some waiting,
23:28but it's I think it's it's a
23:30robust survey and it is.
23:31I think we can say with some confidence.
23:34It is truly representative.
23:37Of the country at the time of the survey,
23:39which is now of course quite
23:41some time ago in 2008.
23:43So the first thing we did is just sort
23:45of put this topic in perspective.
23:47So we looked at all the deaths
23:49from these three cars.
23:51Is the case in deep and identify suicide
23:53poisoning and chronic liver disease,
23:55and we see the the simple
23:56results here on this slide.
23:58Around 4% so one in every 25
24:00deaths are from Dessa despair,
24:02that means 96% of course deaths alright,
24:04and you can see that this these three causes.
24:07Make it roughly equal surprisingly equal
24:09contributions to this overall aggregate
24:11deaths of despair, and in each case,
24:13and this is not a surprise.
24:15Men are far more likely at far more prone
24:18to die of these causes than are women,
24:21so that's sort of just a basic frame
24:24to take with you into the next
24:27set of slides that I'll show you.
24:29And so here in this section I
24:32briefly want to focus on four
24:35markers up socioeconomic.
24:37A concern or adversity,
24:38but I'll talk about low
24:40educational attainment,
24:41unemployment,
24:41low income and then and marital status.
24:43Unfortunately,
24:44you know there isn't in the M deck.
24:46A good measure of wealth.
24:48They do have something which
24:49is another important aspect of
24:51socioeconomic status.
24:52They do have some things
24:53on whether you rent or own,
24:55whether you paid off your mortgage,
24:57I might dig into a little bit to try
25:00to pull something out about well,
25:02so you have that information as you'll see
25:05an income and employment on education.
25:08But so it's missing that piece.
25:10Of course. Most importantly,
25:11as a mental health person,
25:13the thing that's missing is it doesn't have
25:15information about psychiatric disorders,
25:17which really to get an understanding
25:19of these things is essential.
25:21But I maintain that there's still
25:24some things that you can learn
25:26even with that large omission so.
25:29The next four slides are organized
25:30in the same way,
25:32so we take a moment and tell you about
25:34them so you can understand them as
25:36they go by in the top panel of the slides,
25:39I'm showing you crude mortality
25:41rates per 100,000 person here,
25:42which is the way we present
25:44you know mortality,
25:45and this is over the period
25:47of follow up again.
25:48The classification,
25:48and here we're looking at education
25:50is done at the time of the
25:52American Community Survey in 2008,
25:54and in the bottom you're looking
25:55at at had results from Cox
25:57proportional hazards models.
25:58So these are hazards ratios that have
26:00been adjusted for all those things and.
26:02Fine print,
26:03which are a variety of socioeconomic
26:05measures as well as they've got a 6 item
26:08measure of functional disability now.
26:10So if we look at this at the level of
26:14education at the time of this survey.
26:18We can see that we can see these gradients,
26:21especially for poisoning and
26:22chronic liver disease,
26:23with the highest risk among
26:25people with lowest level of formal
26:27education and below when we look
26:29at the hazards of mortality again,
26:31adjusted problems factors we see
26:32that for suicide there's a slight
26:34increased risk associated with
26:36having less than a bachelors degree.
26:38It's similar and somewhat stronger
26:40for chronic liver disease mortality
26:42at the most robust gradient is
26:44for poisoning deaths.
26:45Across these different levels
26:47of of education.
26:48So in a similar way, turning to employment.
26:51We similarly see an Association between
26:53not being employed and the risk of
26:56each of these two causes of death.
26:58In the unadjusted,
26:59crude results above.
27:00And in the adjusted analysis,
27:02it's against stronger for poisoning in front
27:05of liver disease than it is for suicide.
27:08If we look at income,
27:10we see this now familiar pattern.
27:12With higher crude rates suicide.
27:16Poisoning and chronic liver disease
27:17deaths associated with lower income,
27:19specially high rates of poisoning deaths
27:22among people with net income losses.
27:24An in the adjusted analysis below.
27:27These associations have
27:28been largely absorbed,
27:29perhaps because of correlations
27:31among the social network variables.
27:33But however,
27:33even after adjusting for
27:35all of these factors,
27:36the lowest income groups have significantly
27:39increased hazards of poisoning
27:40mortality in relation to highest groups,
27:43and finally.
27:45Considering marital status.
27:46Adults who are separated divorced
27:49have high crude mortality rates,
27:51higher rates than the other
27:53arrow statuses for all three
27:55causes of death, and widowed adults,
27:57likely because of their older age.
27:59I didn't mention this before, but.
28:02The chronic liver disease deaths occur
28:05at much older ages and this at most.
28:08With these reports related to alcoholism,
28:09and it takes many years, you know,
28:11for people that develop alcohol
28:12related hepatitis and cirrhosis
28:13can be stretched out over decades,
28:15and so that's an older group and
28:17that's why you see that as you see,
28:19the higher rates among those who are widowed.
28:22An in below the adjusted hazards tell the
28:25similar story with the highest hazards,
28:27especially for poisoning.
28:29Deaths were separated or sparseness
28:31compared to married individuals.
28:33So lot of numbers price,
28:36but to summarize them.
28:38Suicide poisoning a chronic liver disease.
28:41The so called death,
28:43despair made roughly equal
28:44contributions to mortality.
28:45Men are particularly vulnerable.
28:47Accidental poisoning appears to be the
28:49most sensitive and suicide the least to
28:51these markers of socioeconomic disadvantage,
28:53and again,
28:54it's separated.
28:55Divorced adults not employed individuals
28:57are at high risk and to a lesser extent,
29:00so we're low income people and
29:02those with less formal education.
29:04All of this.
29:07I think provides some support for the
29:09possibility and only the possibility.
29:11That that social policies that aim
29:14to improve occupational opportunities
29:16and financial security seek to improve
29:18educational attainment and try to
29:20diminish or lessen social isolation
29:22that they might and I emphasize,
29:25might have long term benefits in terms
29:28of lower risk of death to despair.
29:32Now.
29:35One related issue that dig into the
29:38weeds a little bit here that interests
29:40me is the extent to which Desta Dispara
29:43Flamel logically related to one another.
29:46Specifically, what is relationship
29:48between suicide and opioid overdose
29:50deaths at the most basic level?
29:52It's conceivable that many opioid
29:54overdose deaths are in fact
29:56intentional suicide events and that
29:58the opioid overdose crisis overlap.
30:00Apps with the suicidal crisis.
30:02When I say crisis,
30:03I mean these gradual upward trends
30:04that I showed you in the national
30:06data over the last 20 years was
30:08particularly steep for opioid
30:09deaths over the last several years.
30:12So people who had not fail opioid
30:15overdoses might be at increased
30:17risk for suicide and vice versa.
30:19And given you know there's some
30:22commonality in their social economic
30:24risk factors that I've just shown you,
30:27these aren't unreasonable speculations.
30:28So this issue kind of caught my interest
30:32went about a year and a half ago.
30:34The directores event IMHO and
30:37Night a issued a warning.
30:39Or not really a warning,
30:40but a message that you see here that
30:43suicide deaths are a major component of
30:44the opioid crisis that must be addressed.
30:47And this was a clear signal from the
30:49federal leadership that they believe
30:51that there was really a tight linkage.
30:53Between suicide and the opioid
30:55epidemic San space to be.
30:57If you read this thing,
30:59it's based primarily on the observation
31:01that opioids are often involved in
31:03non fatal suicide attempts that
31:05show up in emergency Department,
31:07so that isn't actually deaths but
31:10non fatal events and so it led
31:12me to wonder is it in fact true
31:15that that suicide opioid over
31:18deaths substantially overlap?
31:19And if you look at national data
31:21on that and trend data from the
31:24perspective of suicide deaths,
31:25you can see on this graph that is really an.
31:29These are certified deaths and of
31:31course can be misclassification,
31:32but there's really very little evidence.
31:35That opioids are involved in a significant
31:38or increasing share of suicide deaths.
31:40In United States.
31:41It's that thin orange line you see at the
31:44bottom of the graphs actually have to.
31:46Although it's an increasing number,
31:48it's a declining proportion
31:49because it's if you think about
31:51the rising base rate of suicide,
31:53and if you flip this around as we did,
31:56you can see that it's also true
31:58that suicide represents a small
32:00and declining proportion of opioid
32:01overdose deaths in the United States.
32:03So even if we were to.
32:06Include all of the deaths in which
32:08it's not possible to determine
32:10the intent of the overdose,
32:12and these are what are referred to therein.
32:15The Gray line is the Undetermined
32:17desk unit you would still see.
32:20It's still the case that a great majority,
32:23an increasing majority of these deaths.
32:25Opioid overdose deaths would be
32:27accidental or unintentional intent.
32:30Now, Interestingly,
32:30what I kind of went too far there there now.
32:34Interesting when we stratified
32:35by age group that last slide,
32:37we see that suicide deaths do
32:39represent a larger proportion
32:41of the female than male
32:43opioid overdose deaths.
32:44You see that there on the left
32:46hand side of the slide and a
32:48larger portion of the older than
32:50younger opiate overdose deaths,
32:52and this may be because females,
32:54as compared with master side S,
32:57generally involve ingestion's
32:58and less often involve firearms.
33:00And because suicide deaths among
33:02older people more often occur in the
33:05setting of chronic medical conditions,
33:07many of which are painful and therefore
33:09are often treated with opioids.
33:11So it's not surprising to see opioids
33:14show up in urine or blood of older,
33:17older suicide decedents now.
33:20Now just come.
33:21Just because most opioid overdose deaths
33:23are intentional doesn't mean that
33:25suicide and opioid overdose deaths are
33:27entirely distinct clinical phenomena.
33:29In fact,
33:30there's some evidence that these
33:32are related processes in the sense
33:33that they Co occur more often
33:35than one would expect by chance.
33:37So here results from an analysis
33:39of the nice arc data,
33:40and that's a large nationally
33:42representative household
33:43epidemiological survey that was
33:45performed by Julian Santiago Tenorio.
33:46He was at Columbia.
33:48And it reveals in this first slide.
33:51Here we're looking at the lifetime
33:53rates of suicide ideations
33:55and attempts among two groups,
33:57those with and without past year.
33:59Prescription opioid use disorder,
34:00and you see that they have those with
34:03the prescription abuse or have much,
34:06much higher lifetime rates of of
34:08reporting suicidal ideations and reporting.
34:10Having made an attempt,
34:12this is becomes more interesting
34:14when you look prospectively one
34:17of the things about the 1st 2.
34:20Sir, nice are surveys.
34:21Is there link?
34:22There are two waves of 1 sample three
34:24years apart and what we found here is
34:27that even after extensive adjustment
34:28for a wide range of factors that
34:30are listed below there you can see.
34:33That that as compared to
34:35those without pasture,
34:36non medical opioid use or disorder at
34:39wave one those with the non medical
34:42opioid use were significantly more
34:44likely to develop to develop suicidal
34:46ideations as you see on this slide
34:49here over the ensuing three years.
34:52Now we also looked at a suicide attempt here,
34:54and as you can see on the right,
34:57we didn't find a signal.
34:58But you can also see that they
34:59are very wide confidence intervals
35:01and it's my own sense that these
35:03analysis are underpowered,
35:04but I showed them to you so that
35:06you can make what you make,
35:08what you will of them in a.
35:11I'm on a related topic.
35:15And this is an analysis of the nested data.
35:18It's another large epidemiologic survey.
35:20It's of the civilian non
35:22institutionalized population.
35:22Hillary samples.
35:25This is a post actually is now an
35:27assistant professor at Rutgers.
35:28I'm pleased to say she similarly
35:30found that if you look at her,
35:33she's doing a propensity score
35:34weighting based on a wide range
35:36of clinical variables and social
35:38economic variables that adults
35:40with past year opioid misuse which
35:41I'm showing here on the right,
35:43were more likely to report suicidal
35:45ideations in the past year and plans then
35:48were people with no opioid use on the left.
35:51And in fact that no abuse in those
35:53the medical produced in the middle.
35:55Look very similar and again,
35:58it's similar to Julian's work.
36:00It's suggesting that there's a connection
36:03between opioid misuse and suicidal behavior.
36:07Now these sorts of findings.
36:11Razum an interesting conceptual issue.
36:17But should we embrace a unified perspective
36:20to intentional and unintentional injury
36:22that emphasizes there she shared social
36:25determinants like economic adversity,
36:27as I've been emphasizing this
36:29morning or maybe even biological
36:31risk factors like impulsivity and
36:33so should we kind of conception,
36:35lumped together into sexualize
36:37them as self injury or self harm?
36:40Or Alternatively,
36:41should we assume the traditional
36:43psychological perspective that distinguishes
36:45the motivation of overdoses and considers
36:47unintentional and intentional overdoses?
36:49As discrete or separate phenomenon.
36:50Now most mental health professionals,
36:52of course adhere to the latter perspective
36:54that psychological perspective and you know,
36:56if you think I'm from Columbia,
36:57so you know,
36:58I think about the Columbia suicide
37:00Severity Rating scale and that scale.
37:02Like all suicide scale places,
37:03great emphasis on intentionality as
37:04this in quinone of suicidal behavior.
37:06But I want to point out that there
37:08are dissenting voices out there and
37:10they're calling a little louder.
37:12I think over the last several years.
37:13And here's a quote in Rocket and
37:16his colleagues are one of them.
37:17And here's a quote from from his work.
37:20Suicide and lethal overdoses often share
37:22many of the antecedent characteristics
37:23in terms of emotional, behavioral,
37:25familial, and social disruptions.
37:27He goes on to say the label of accident
37:29is served as a default determination.
37:32When definitive approximal
37:34evidence remains uncertain, so.
37:38So again,
37:39there's this question about how much are
37:41these sort of all one phenomena versus
37:43discrete phenomenon and and this is,
37:45I apologize,
37:46somewhat of a confusing slide.
37:47I'll try to walk you through it.
37:49This is from our own work here.
37:52I'm putting together two different studies,
37:54two different studies involved
37:55linking Medicaid data claims to
37:56the National Index Death Index.
37:58So we're following people who are
38:00in the emergency Department and
38:02looking out over a year to see
38:04who dies and what we found.
38:05There is an what I'm bringing together again,
38:08or the risks.
38:09Of suicide in the year following
38:11suicide attempts an in the year
38:13following non fatal opioid overdose.
38:15And if you just focus on the female
38:17then those two histograms and blue and
38:19yellow there are roughly the same height.
38:22What that this in fact is showing even
38:24though it's from two different studies.
38:26The methods are similar.
38:28We are finding similar risks
38:30at the group level.
38:31So that is a woman showing up into the
38:34emergency Department with a non fatal
38:37suicide attempt is just as likely over
38:39the following year to die of suicide,
38:42as is a woman who shows up with a
38:45non fatal opioid overdose event.
38:48Some sense the opioid overdose
38:49and a suicide attempt for women,
38:52at least according to these analysis,
38:54are risks, risk equivalents.
38:55But let me emphasize this,
38:57showing similar risks at the group
39:00level like this isn't the same thing.
39:02Is showing at an individual level and
39:05that so it isn't yet demonstrated that
39:07people who make unintentional suicide
39:10overdoses are equally at risk for
39:12both fatal unintentional and suicidal
39:14fatal events, as might be expected under
39:17the assumption under the robust assumption
39:20of the unified self injury model.
39:22So to put that concept to the test.
39:27What we did is we found a cohort of
39:29patients who presented the California
39:31Emergency Department with either
39:33accidental or intentional non fatal
39:35overdoses to evaluate their risk
39:37of suicide and accidental overdose
39:39deaths over the following year.
39:41Sort of similar to the studies
39:43I've just shown you,
39:45but here we're following individuals now.
39:47If accidental and intentional
39:49groups have similar proportionate
39:50risks of these types of deaths,
39:52it would support the unified
39:54self injury perspective, right?
39:56People would travel along.
39:57With the same proportionality,
39:59they travel along that those.
40:00Blue lines as the black lines, right?
40:05But but if it were the other way around,
40:09and if fatal accident overdoses, right?
40:12Just say if non fatal accidental
40:14overdoses had greater risk
40:15of fatal accidental overdose,
40:17then of than those with non fatal
40:19overdose then it would support the
40:22traditional psychological perspective
40:23that emphasizes intentionality
40:25as a central defining feature.
40:27People would be much more likely
40:29to travel along the black lines.
40:32The accidental non fatal overdoses
40:34would be in greatly increased risk.
40:36Fatal accidental overdoses and
40:38those with intentional not fatal
40:40events would be a greatly increased
40:42risk for suicide deaths and not.
40:44Up for the related concepts.
40:46So if you as you can see from
40:50the results here.
40:53The findings provide support
40:55for both perspectives.
40:56Suicide risks are here in yellow
40:58and that they are greater for
41:00patients with non fatal intentional
41:03overdoses than accidental overdoses.
41:05While the risks of accidental overdose
41:07death risk circling blue were much
41:10greater for patients with non fatal,
41:12accidental and intentional overdoses,
41:14so this differential mortality risks with
41:16non fatal overdoses by intense supports.
41:18The clinical utility of distinguishing
41:21non fatal overdoses by intent.
41:24However, if you look at
41:25that findings on the right,
41:27which look quite the same in
41:29terms of their their shape,
41:31but the scale has changed and
41:33the scale here is standardized,
41:35mortality rates rate ratios.
41:36Those are the extent to which people
41:39are more likely to die of these causes
41:41of death than would be expected on
41:44the from the general population,
41:45based on their demographic characteristics.
41:47And here you see that all the groups
41:50have greatly elevated risks for external
41:52cause mortality across the groups.
41:54And that if you focus on the fact
41:56that there are much greater than one.
41:59Is innocence supportive of this unified
42:01self injury conceptualisation that
42:02emphasizes common underlying determinants,
42:04either social or biological,
42:06that might that might contribute to
42:09their risks or the other causes of death.
42:12So a clinical implication of all this,
42:15and I realize that it's it's a little
42:17bit in the weeds is that probably makes
42:20good sense to integrate substance
42:23use and mental health services,
42:25particularly in an emergency
42:27Department where these people present.
42:29And to carefully evaluate people
42:31making suicidal overdoses that involve
42:33opioids for the possibility that they
42:35actually have an underlying opioid,
42:37use disorder, for example,
42:38and to evaluate those making would
42:40appear to be accidental overdoses.
42:42People with opioid people,
42:43drug strong drug related histories
42:45for the possibility when they show
42:47up in your merchant with overdose
42:49that they're actually suicidal and
42:51may need mental health care.
42:53And that simply referring them to
42:55our back to a substance use center
42:58may not be sufficient.
42:59So you know in a sense.
43:02These results aren't too surprising,
43:04yet that in it because in addition
43:05to the shared socio economic risk
43:07factors that I focused on this morning,
43:10there's also evidence from a bunch
43:12of prior research.
43:13I won't go into detail that suicide and
43:15opioid overdose deaths have several
43:17similar or shared other risk factors,
43:19and you see some of them on the slide here.
43:22Things like depression and Evers child
43:24experiences, severe medical illnesses,
43:26especially those involving.
43:28Painful medical conditions.
43:31So.
43:33Before closing I I wanted to turn
43:36briefly to some planned research that
43:39I hope to be doing over the next.
43:42As it's several years several
43:44years now over the last year.
43:46So during the pandemic there's been
43:48considerable concern about the
43:50psychological fallout of the pandemic,
43:52and there's apprehension over
43:53consequences related not only to
43:55things like job losses and evictions
43:58and bankruptcies and high levels of
44:00personal distress and social and
44:02personal bonds that become afraid.
44:04There's grieving and mourning
44:06the loss of loved ones.
44:08Parents coming under great stress,
44:10having to home school children,
44:12and the difficulty set up that poses
44:15especially for people who don't have access
44:18or have limited access to the Internet.
44:21People living in congregate living
44:23situations are at particularly high risk
44:25and under great stress from the virus,
44:28and I'm an interested in understanding
44:31the effects of financial crises.
44:34On suicide and and
44:35substance related overdose.
44:37An in order to target efforts
44:38to rent or at least buffer or
44:41ameliorate the harmful effects at an
44:44individual level of suicidal crises,
44:46we need to understand something
44:48about their vulnerability and the
44:50vulnerability of economic strain
44:51to suicide and opioid overdose.
44:53And I can tell you,
44:55at a population level,
44:57it's not really very satisfying research.
44:59There tends to be an increase
45:01in suicide rates of roughly.
45:03.7 per 100,000 for every 1% increase there is
45:06an unemployment over the course of the year,
45:09but that's a very general observation
45:11from International Studies,
45:11and it's very imprecise,
45:13and it likely varies a great deal with things
45:15like the strength of the social safety,
45:18net,
45:18availability of formal informal supports,
45:19and a whole host of other factors.
45:23But it turns out we know very little
45:25about the fact of financial crises on the
45:28individuals risk for suicidal behavior,
45:30and that's the topic that I hope to
45:32be studying over the next few years.
45:35And one of the things that motivated
45:37me to study this with some work that I
45:40that was done again with that nice arc,
45:43first two waves that I mentioned before
45:45large epidemiological study done by an Hill.
45:47Garcia la Garza.
45:49Who is a doctoral candidate at Columbia?
45:52I've worked with me.
45:53What he did was have to say it
45:56was a real Tour de force on Hill.
45:58Took the entire nice arc and there
46:01are 3000 different variables in it.
46:03It's a long interview from Wave One and
46:05then he you use machine learning techniques.
46:08He tuned to balanced random forest
46:10analysis with ten fold cross
46:11validation and so in and in a way
46:14that doesn't impose any theory.
46:15He's looking to see at which of
46:18these 3000 variables that were
46:19collected at wave one predict.
46:21A suicide attempt over the following
46:23three years and then he and you know
46:26you come out with all these explanatory
46:27power in terms of the area under the curve.
46:30I won't bore you with the details
46:32of the model fit,
46:34but was interesting to me.
46:35I think relevant to my to my my
46:37interest was the ranking of variable
46:39importance and this is the right here.
46:42You see the list of the top ten of
46:44those things from this list of 3000
46:46variables and this is derived if
46:48by taking one variable at a time
46:50and looking at how it changes,
46:52the overall fit of the model, you can.
46:55Driver ranking it's not surprising.
46:58From a clinical perspective,
46:59that right at the top of the list is
47:02the people who three years earlier said
47:04they felt like they wanted to die.
47:05They would that that single
47:07variable was the strongest,
47:08predicting suicide attempts over
47:09the following three years and that
47:11you see the other two behind it or
47:13quite closely connected with suicide.
47:14And then there's a depression item age.
47:16We know that suicide attempts are
47:18highly dependent on age and they
47:20actually go down with age and the
47:22highest risks are among young people.
47:24Fortunately, most those are not hard,
47:27non fatal.
47:27And then there's an item about doing
47:29things less carefully because emotional
47:32problems and then directly following that.
47:34Is this issue or the one I've
47:36highlighted here that item dealing
47:38with financial crises and bankruptcy,
47:40and in fact,
47:41four out of the top 20 most important
47:43variables are related to jobs and income,
47:46and so this provides some of
47:48its at the individual level.
47:50And again,
47:50it's very crude 'cause we only have
47:53this one measure and then three years
47:55later we have a self report thing about
47:58suicide attempts and we don't have
48:00any information on people who died
48:02of suicide during that time frame,
48:04but nonetheless it's consistent
48:05with this idea that.
48:07Financial crises are one of the things
48:09that predict future suicidal behavior, so.
48:15And so here's what I've been up to.
48:17Some of the mischief I'm up to now,
48:19so I've been working with my colleague
48:21and close friend Steve Marcus at the
48:23Penn School of Social Policy and Practice.
48:25And a colleague, Molly candidate.
48:27But the Wharton school.
48:28And we've been working with Trans Union.
48:30There are consumer credit
48:31rating reporting company.
48:32It's amazing.
48:33I mean they've got data on the credit
48:35reporting scores that are updated
48:36all the time from the great majority
48:39of people in the United States.
48:40Well over 200,000,000 people.
48:42Basically anyone who's in the cache.
48:43Kind of anyone who's getting a paycheck
48:45or has a credit card or a debit card,
48:48they have your credit rating score and
48:50it changes all the time and we brought
48:52them together with the people that opt in,
48:55which is so large.
48:56Part of the United Health Group
48:58and and we've got them to agree
49:01to merge their data to create
49:03a deidentified merge data set.
49:05We haven't yet done it.
49:07When that's well still ironing out the
49:09details to do it on a small scale so
49:13we can write it a grant to support this work,
49:16but the idea is to take this integrated
49:19database that has at the patient level
49:21consumer credit rating scores that
49:23fluctuate with claims histories and
49:25link them to the national death in.
49:28And that will put us in position,
49:31we believe.
49:32To look at the effects of dramatic changes,
49:36dramatic down grading of people's
49:37credit rating scores, which are,
49:39in a sense an individual
49:41level of financial crisis,
49:43and to see whether that impacts
49:45people's future risks of drug overdose,
49:48deaths and suicide.
49:49And Moreover,
49:49we were particularly interested in is
49:52whether there are pre-existing factors
49:54that you can see in the claims data.
49:57So, for example, if a person has a history.
50:00Depression. Does that make them particularly
50:03vulnerable in terms of their suicide
50:05and overdose risk to a financial crisis?
50:08Those are the kinds of questions that
50:10haven't really that I've wondered
50:12about but haven't really been able
50:14to interrogate it empirically,
50:15because thankfully, these causes of
50:17death are sufficiently rare that is
50:19hard to get a large enough sample,
50:21and to get detailed personal level data
50:24to actually probe these associations so.
50:27You know, so after many months of working.
50:30With these two companies,
50:31we're gotten them to enter into an
50:34agreement to link their data at where
50:36the process of getting a small sample
50:38and building some pilot data for a grant
50:40that we will be writing over this summer.
50:44But there are some problems with this
50:46thing that didn't occur to me when I
50:48was thinking about this abstractly,
50:50and that is that when people undergo
50:52a financial crisis,
50:53they often sadly lose their health insurance.
50:55They lose their private health insurance,
50:57which is what we can see in the optim data.
51:00Some of them will stay on Cobra,
51:02but so we will only really have
51:04their death data.
51:05We won't be able to see whether
51:07after finish crisis,
51:08people have an increased risk of going
51:10second emergency Department within
51:11a non fatal suicide attempt, or it.
51:13Non fatal overdose and so forth will
51:15really only be able to study mortality,
51:18but nevertheless, you know,
51:19we hope to gain some insights on
51:21risks and protective factors.
51:23Based on claims histories of
51:25individuals prior to experiencing an
51:28individual financial crisis reflected
51:30in their claims scores.
51:33So if it works out, maybe.
51:35You can come back in a few years
51:38and tell you what we found.
51:41So in closing,
51:42let me reiterate that people
51:43who are financially vulnerable,
51:45those who have don't work and have
51:47less education appear to be an
51:49increased risk for overdose deaths
51:50and to a lesser degree chronic
51:53liver disease and suicide mortality.
51:55For these reasons, it it's you know,
51:57it's possible that social policies
51:59that support education and employment
52:01might yield long term benefits in
52:02reducing these deaths of despair.
52:04However, also, and you know, I.
52:06Realize all of you know this,
52:08but I'll say it anyway,
52:10that these deaths is fair.
52:12While they have some overlapping
52:13socioeconomic risk factors,
52:14there are also very distinct clinical
52:16phenomenon.
52:17They call for different treatment and
52:19different rehabilitation approaches,
52:20and For these reasons improving
52:21access to substance use and mental
52:23health services will obviously
52:25also play a critically important
52:27role in achieving these goals.
52:28So Lastly,
52:29let me just say that I am really fortunate
52:32in being able to work with and learn from,
52:35such as Sarah.
52:36Such a large and diverse group
52:38of exceptionally talented.
52:39And generous colleagues.
52:40And here the ones whose work I
52:42featured this morning and I want
52:44to thank you for your attention.
52:46I'm happy to take on other questions.
52:50Thank you very much Mark for incredibly
52:54stimulating and erudite presentation.
52:57And demonstration of the scope of your
53:00interests and ability to gather data.
53:03I wanted to start by asking you one question,
53:07a methodological question?
53:08You've looked at the data at the individual
53:12level and relationships between individuals,
53:15but much of the writing about
53:17it's true of case and Deaton.
53:20But also the writing about the opiate
53:24crisis is about the loss of community and.
53:29To their defense case,
53:31indeed, and you know,
53:32used deaths of despair as
53:34a kind of a shorthand,
53:37but they mainly were interested in
53:39geographic variability and locations,
53:41and I wondered if you thought of
53:43a way using your data.
53:45Most of these datasets don't
53:47identify communities,
53:48and I wonder if you've thought of a way
53:51of adding that dimension of Community
53:54integration or social capital in a community.
53:58As as as as a major risk factor.
54:00In addition to these
54:02individual characteristics,
54:02you know that's a wonderful
54:04question about and
54:05I and the short answer is
54:07somewhat a little bit you.
54:08This M DAG data that I showed those
54:11detailed slides you know on on on the
54:14different education and income and so forth.
54:16There is geographic information there,
54:18and I'm aware that people have looked at.
54:20You can look at I don't think you
54:23can go down to the census track,
54:25but you can go down to the zip code.
54:29And you can have County level data we
54:31did for example with that data set.
54:34And I know this isn't quite on topic,
54:36but it's it's structurally related.
54:38We looked at with that data set at
54:41firearm related suicides and looked at
54:43them in relationship to peoples where
54:45their state of the firearm ownership
54:47within this state and you could show
54:49very strong associations with people who
54:51live in states with higher firearm ownership,
54:54being much more likely to die of suicide,
54:57but also to die, I proportionately
54:59have firearm suicides, but I.
55:01Imagine the same kind of thing at a
55:03more granular level could be done.
55:06At the but part of the I think
55:08one of the challenges, though,
55:10is that some of these things change if
55:12you're looking over a 10 year period,
55:14you might have to update.
55:16You know that you might have
55:18to work with time dependent.
55:21Regional variables,
55:21but I I think you're right that that
55:23would really open things up to be
55:25able to look at the level of social
55:27capital within the communities that
55:29these people live as an additional
55:31dimension or explanatory dimension
55:32to their risks for these deaths.
55:33So I think that's an excellent point,
55:35but it's something if I if I had
55:37a student who is interested in it,
55:39I would be I could pursue.
55:41But I'm so busy at this point
55:43will be tough for me to pick up.
55:45But thanks for pointing that out.
55:48So
55:49I see in the chat box, JP Daquino had a
55:53comment JPD want to make that out loud.
55:57Sure, thanks for the excellent talk I was.
56:01I was struck by your introduction of
56:04vertical as a diagnostician of these
56:07social ills and some of the things you
56:11said reminded me of another diagnostician.
56:14Emailed or climb the French sociologist.
56:18Who wrote the book called Suicide in late
56:2119th century as friends transition from a
56:24traditional society to an industrial society?
56:27And I'm I'm, I'm wondering what
56:30your thoughts would be on as we
56:32transition to a postindustrial society.
56:36Whether the same meals that he
56:40diagnosed Geneva, few individualism,
56:43loss of community, excessive hope.
56:48You know loss of organized religion
56:51and weakening of the nation in the
56:53family weather work. This is kind
56:56of history repeating itself in a
56:58way you know. There's a wonderful
57:01point and I I think sadly there are.
57:04There are strong parallels.
57:05If you step back and look at the changes
57:09that have a Kurd have hollowed out.
57:12You know whole communities in the
57:14in the Industrial Midwest places
57:16that used to provide steady places
57:19of employment allowed people with.
57:21High school level of education to work
57:24their way up and supportive family.
57:26Think it really that the changes in our
57:29labor markets and our economy shifting to
57:32a you know high information based economy.
57:36And you know the greater disparities
57:38in an income that have opened up
57:41over the last couple of decades.
57:43Have really taken a toll on people
57:44and we can see the resentment.
57:46We can see it in our politics over
57:48the last several years and so we are.
57:51You know it's we have a.
57:53A dynamic economy as they had in in
57:56France in in Durckheim's day and and
57:58there are some of the consequences that
58:00he wrote about I think are still app,
58:03so thanks for.
58:04For pointing that out.
58:12The floor is open for questions
58:15and comments. I believe
58:16Jank Tech has a question.
58:20Hi, thank you. You also conducted
58:23and published several studies on
58:25early mortality and serious mental
58:28illness and schizophrenia and and
58:30I mean I could with this like in a
58:35renewed focus on on socio economic
58:39economic determinants of health.
58:41Much of that was. You know poverty.
58:45In serious mental illness and
58:47how much of that was?
58:49I mean, do you have any ideas on that?
58:52Yeah, well, you
58:53know, I think.
58:53Sure, and I didn't really talk
58:55about schizophrenia this morning,
58:57but very happy too, I think.
59:00You know overall within people
59:02and also schizophrenia.
59:04You know you have a population that
59:06has lots of that that has health
59:10challenges across a number of dimensions.
59:13At the most basic level,
59:15you know we live in a society that depends
59:19upon people rationally seeking care.
59:22When they are ever health read
59:25and people schizophrenia.
59:27Are quite impaired in that place,
59:29so even relatively simple
59:31things like an appendicitis,
59:32which generally results in an appendectomy.
59:35That's not such a serious
59:37procedure can be life threatening,
59:39and someone with schizophrenia
59:40who delays treatment seeking.
59:42So there's thinking about the cognitive
59:44problems that many adults with
59:46schizophrenia have in negotiating properly,
59:48go shooting and being received in an
59:50unbiased way in our health system.
59:53There's a layer there that then
59:55below that there are.
59:57Aspects.
59:58You know up there.
01:00:00There are aspects related to
01:00:02health behaviors, you know.
01:00:03Within schizophrenia we know from
01:00:05characterizations of the Katie sample that
01:00:07very high rates of obesity and hypertension,
01:00:09hyperlipidemia,
01:00:10things that pose great risks.
01:00:11We have high rates of smoking,
01:00:13the people who's not only people
01:00:15discussing it more likely to
01:00:17smoke in the general public.
01:00:19They're more likely of the smokers.
01:00:21They smoke more heavily so that their
01:00:23risks of polarities are, you know,
01:00:25are greatly out of proportion.
01:00:26Even more greater portion,
01:00:28there are suicide and then
01:00:30you have the disorder itself.
01:00:32We have a paper in impress
01:00:34attempts to control it.
01:00:35Come out in a few months looking
01:00:37at the patterns of suicide in
01:00:39schizophrenia and unlike the general
01:00:41population where it tends to go
01:00:43up and peaks in late middle age
01:00:45or in early adult older adult age
01:00:48in schizophrenia goes down and it
01:00:49is very highest in those first few
01:00:52years so that the psychosis paranoia
01:00:54people perhaps aware of the illness
01:00:56and the life in front of them.
01:00:59And their lives. So the illness.
01:01:01How some people believe how we treat
01:01:04it with some of the anti psychotics
01:01:06and their and their metabolic
01:01:09problems that they pose as well as
01:01:11embedding these issues with in.
01:01:13Our health care system,
01:01:15in the challenges of accessing
01:01:17appropriate care health behaviors,
01:01:19it's really a multiply determined
01:01:21thing that results in the shorter
01:01:25longevity of people with.
01:01:27Serious mental illness and then
01:01:29on top of that there of course,
01:01:32are disproportionately represented
01:01:34in the lowest rungs.
01:01:35In terms of socioeconomics, with,
01:01:38you know the great majority of
01:01:41people with schizophrenia.
01:01:43More than 2/3 are in are in the
01:01:45Medicare or Medicaid programs, or both.
01:01:48So you know,
01:01:49it's a very difficult and complicated thing.
01:01:51You went back.
01:01:52One of the things that I became aware of,
01:01:55like I went into it, thinking,
01:01:57well, it's probably mostly suicide,
01:01:59is probably it's actually not
01:02:00much this suicide.
01:02:01It's mostly pulmonary and cardiovascular
01:02:03disease that's contributing to the
01:02:04shorter life expectancy in the
01:02:06premature mortality in that population.
01:02:07So thanks for bringing that up.
01:02:14There's a comment from Sandra Bacon.
01:02:22Sandra, do you wanna speak up or
01:02:25should I read your sure?
01:02:28Sure, so I'm particularly interested
01:02:30in the high utilization of patients
01:02:32in psychiatric emergency services and
01:02:35what strikes me is that these patients,
01:02:38despite the amount of time that
01:02:41they spend in acute services,
01:02:43their outcomes are really poor,
01:02:45and disposition planning is
01:02:47oftentimes stymied by the fact that.
01:02:50What they need is not available
01:02:52in the community either.
01:02:54It's not a service that's covered,
01:02:56or there's waiting lists,
01:02:58and so they patients continue
01:03:00to use the emergency service.
01:03:02So how do we address the fact that
01:03:05the care that people may need is
01:03:08not necessarily always available?
01:03:11That is a big and important question.
01:03:14You know,
01:03:14I think first I just want to
01:03:16validate your observation.
01:03:18If you look and see at people
01:03:20who who people with serious
01:03:22mental illness who are discharged
01:03:24from the emergency Department,
01:03:26the single most robust predictor of
01:03:28whether they're going to run into
01:03:31trouble and require an represent
01:03:32to an emergency Department or
01:03:34require hospitalization in the
01:03:36ensuing weeks is whether or not
01:03:39they're in care before they came.
01:03:41To the emergency Department,
01:03:43it's it also,
01:03:44it's the strongest terminal whether
01:03:46or not they will follow up with care.
01:03:48We ask an awful lot of
01:03:50emergency department's.
01:03:51They're not really well
01:03:52configured for the care,
01:03:54most of them for the care of people
01:03:56who are serious milleson an accept
01:03:58replaces psychedelic haven where you
01:04:00have a dedicated psychiatric service.
01:04:02Most emergency departments
01:04:03actually don't have.
01:04:05And this is a remarkable thing.
01:04:07They don't have a psychiatrist
01:04:09or a psychologist either on
01:04:11site or available on call.
01:04:12It's really primary care.
01:04:14It's really part of the General
01:04:16Medical sector and you know.
01:04:18So to expect that aspect of our
01:04:20General Medical sector to care for in
01:04:23a compassionate and effective way.
01:04:25People who are at a people who have
01:04:27the greatest level of psychiatric
01:04:29illness severity at a time in
01:04:32their lives when they are in a
01:04:34crisis is completely unrealistic.
01:04:36And then it further,
01:04:37there's the point that you make
01:04:39that there aren't available slots.
01:04:40Particularly I think for young people,
01:04:42children in many communities and for
01:04:44older adults they have long waiting lists.
01:04:46This isn't a group that isn't
01:04:48often internally motivated and
01:04:50has the persistence to follow up.
01:04:52An emergency Department also overwhelmed.
01:04:54They don't have.
01:04:54There are some things,
01:04:56and there's a nice review,
01:04:57but I that I refer you to if
01:05:00you're interested in this.
01:05:01By Stephanie Dupnik in JAMA Psychiatry.
01:05:03Last year,
01:05:04a meta analysis and she's looking at
01:05:05things that emergency departments
01:05:07can actually do right in the
01:05:09emergency Department and actually
01:05:10improve the short term outcome
01:05:12for people who are discharged.
01:05:13And they are things like sending texts,
01:05:16supportive text reminders,
01:05:17an engaging family,
01:05:18and so forth in the follow up.
01:05:20But they require effort and they.
01:05:22Probably work best in communities
01:05:24that have accessible,
01:05:25timely accessible mental services.
01:05:26As you say,
01:05:27I in a course in many settings
01:05:30and particularly even
01:05:31in you know in urban settings that
01:05:33this can be quite a challenge.
01:05:36I think it's it you put your finger on
01:05:38one of the critical weakpoints in many
01:05:41local mental health systems, so thanks.
01:05:53Other comments. So so I
01:05:57have one. So first mark.
01:06:01A brilliant erudite rigorous tour.
01:06:04The first lecture we'd
01:06:06expect nothing less of you,
01:06:08but we're extremely pleased to see.
01:06:13You know it's a thrill to have you
01:06:15here and to have such a brilliant
01:06:17lecture. Really thank you.
01:06:18Thanks for your kind words.
01:06:21The second thing is that I would like
01:06:25to key off of of Sandy's last comment
01:06:29so we have health care systems.
01:06:33There are overstressed.
01:06:34We have treatments that are often.
01:06:37Not definitive, you know there
01:06:40there they have transient efficacy,
01:06:43at least over the short run an.
01:06:48And and even in the long run,
01:06:51often are not able to achieve all the aims
01:06:54that we hope to achieve in treatment.
01:06:57And there is an implication in
01:07:00some of the things that you said.
01:07:03How at a societal level,
01:07:06through economic and vocational?
01:07:11Creating vocational, economic and
01:07:12vocational opportunities for people.
01:07:14You might be able to prevent some
01:07:16of the downstream consequences of
01:07:18the social problems that we're
01:07:21essentially treating in psychiatry.
01:07:23And and I'm curious.
01:07:25I'm curious what what you mean by that,
01:07:29but you might mean by that given.
01:07:33For example,
01:07:34what we're seeing now a $1500 check.
01:07:38$1600 check being provided for for
01:07:43covid relief and things like that.
01:07:47What what?
01:07:48What do you think societies can
01:07:51do through through social or
01:07:54economic policy that would actually?
01:07:57Measurably had the possibility of
01:07:59measure Lee measurably decreasing
01:08:01both accidental death and suicide.
01:08:03Yeah, well, first of all,
01:08:04it's
01:08:05it's. It's anything I say on
01:08:07this point is conjectural.
01:08:09We don't really have strong evidence
01:08:11that work there being a tight
01:08:13connection between these things,
01:08:15but but but if you do step back,
01:08:18you see you see these
01:08:19correlations that I've shown you.
01:08:21My sense is that report is going to far more,
01:08:25much more than a $1600 check.
01:08:28It's going to require a
01:08:31kind of a reorientation.
01:08:34Uh, and a strengthening of the safety net.
01:08:38But you know, I'm not naive.
01:08:40You know, to the political realities
01:08:42we live in a society that's that's
01:08:45very fractured along cultural and
01:08:47ideological and political lines.
01:08:49Underneath it,
01:08:50there isn't as much solidarity as a
01:08:53people that we have as is enjoyed in,
01:08:56you know,
01:08:57in northern European countries that
01:08:59have a much stronger safety net.
01:09:02But I,
01:09:02I think it requires a much
01:09:05larger structural change that
01:09:06then providing checks for.
01:09:08A year, you know,
01:09:10they say that the recently passed,
01:09:11although it's massive in size,
01:09:13will lift the large number of children
01:09:15out of half of the people to have
01:09:18their children out of poverty.
01:09:19But it's only for a year, you know,
01:09:22until you have very basic,
01:09:23you know reforms of things like you know,
01:09:26providing childcare so that
01:09:27women and men are able to work,
01:09:29and subsidizing that for
01:09:31low income people and,
01:09:32you know,
01:09:33fortifying the schools in low
01:09:34income neighborhoods.
01:09:35I mean the whole variety of things,
01:09:37and it would take a long time.
01:09:40But it might well be that that that those
01:09:42kinds of things would relieve distress.
01:09:45They're never going to,
01:09:46you know,
01:09:47we would still live in a world with
01:09:49schizophrenia and bipolar disorder
01:09:51and severe mental disorders that are
01:09:53largely biological determinants.
01:09:55But for many of the substance
01:09:57use related problems.
01:09:58And for many of anxiety and depression.
01:10:01I can well imagine that that that is
01:10:03society that was more generous and
01:10:04its safety net that you would have
01:10:06ameliorate the course of these things,
01:10:08and it might well show up.
01:10:10In some of the broad databases that I
01:10:12talk that I've described to you this morning,
01:10:15but again,
01:10:16I don't want to oversell this.
01:10:17This is, you know,
01:10:19this is all conjecture.
01:10:22Except Posner had a comment, Seth.
01:10:38So Seth asked me to read this the
01:10:41early 1990s insurance parity laws
01:10:44required coverage for bipolar,
01:10:47schizophrenia, etc,
01:10:48but not substance abuse.
01:10:50This made it financially advantageous
01:10:52for patients and treaters to
01:10:55attribute drug overdoses to
01:10:57depression and suicidal thinking.
01:10:59Which would be covered as opposed
01:11:01to uncovered substance use?
01:11:03Do you have a way with your
01:11:05datasets to sort out a general
01:11:08change over the years in survey
01:11:10responses and clinical reporting?
01:11:13You know that's a great question, there are.
01:11:17There are examples you know through
01:11:19littered throughout the healthcare
01:11:21system where incentives are put
01:11:23on to to recognize your diagnose,
01:11:25people in different ways so you know
01:11:28if you back up a decade or so before.
01:11:32The Parity Act and you look in the
01:11:341993 changes in the disability
01:11:36in the SSI disability.
01:11:38They excluded substance use as a disability,
01:11:40and suddenly those people lost
01:11:42their claims and we had more people
01:11:45making claims for other things.
01:11:47You know you can look at inpatient care
01:11:49for children and adolescents and see that
01:11:52you have a great disproportion of children,
01:11:55analysis or diagnosis for the
01:11:57first time with bipolar disorder,
01:11:58is that something that will
01:12:00get you into the hospital.
01:12:02So you need to be whatever you're
01:12:04working with these claims data.
01:12:06Which are simply a reflection of the actual
01:12:08diagnosis that clinicians are making.
01:12:10In practice,
01:12:11you need to be mindful that
01:12:12it isn't the same thing as a
01:12:15structured clinical interview.
01:12:16It isn't these aren't skids.
01:12:19These do represent not only the
01:12:22clinical judgments but also the
01:12:24economic context under which.
01:12:26Missions are making those judgments,
01:12:28so I guess my short answer to that
01:12:31question is I'm not really aware
01:12:33of a good way with claims data of
01:12:36getting beneath these kinds of.
01:12:39A biases that are introduced by
01:12:41a differential reimbursement,
01:12:42and you know,
01:12:43and I think the best you can do is to
01:12:47look for natural experiments that occur.
01:12:50Look for people who transition it
01:12:52between insurance plans that cover
01:12:54different things to quantify them,
01:12:56but that doesn't allow you really
01:12:59to get underneath and fully
01:13:01understand these processes,
01:13:02but it does that question does.
01:13:06Give us,
01:13:06I think inappropriate sense of humility,
01:13:08about the validity and the accuracy
01:13:10of the data that we work with
01:13:12that extends all the way through.
01:13:14There's a lot of concern that and there
01:13:17was concern even back in Turkey Times Day.
01:13:20Derek,
01:13:20I made a big deal out of your little
01:13:23being differences in rates of suicide
01:13:25between Protestants and Catholics,
01:13:27saying that Catholics have much lower rates.
01:13:29Well,
01:13:29in the Catholic Church,
01:13:30you can't be buried on sacred ground
01:13:33if you have.
01:13:34If you're deemed a suicide and so
01:13:36that there may have been deliberate
01:13:38MIS attributions of causes of death
01:13:40going way back a long period of time,
01:13:42so it's not.
01:13:43It's not a new issue,
01:13:45and it's anytime you seek to
01:13:47work with large databases where
01:13:48you're not actually doing in.
01:13:50You know intensive independent assessments.
01:13:52You can fall prey to these things,
01:13:54so that's a very good point.
01:13:58So I think we've reached the
01:14:00end of our allotted time. Ann.
01:14:03Let me again, thank Mark for joining us.
01:14:07And for sharing both specific subject,
01:14:09but also his methodological approach to
01:14:12looking at mental health problems and
01:14:15service delivery and a scale that we
01:14:18need to pay more and more attention to.
01:14:21So thank you very much to everybody is
01:14:24joining to John Crystal and especially to
01:14:27Mark for returning to Yale for this time.
01:14:31Thanks very much.