0:01

Ted Audio Collective. Welcome

0:16

back to TedHealth. It's your host, Dr.

0:18

Shoshana Ungerleiter. Have

0:20

you ever scratched your head at that super

0:23

specific ad on your social media feed and

0:25

thought, how the heck did

0:27

they know I was after that mini

0:29

avocado peeler? Well, it's

0:31

true. Tech companies and their algorithms have

0:34

kind of figured us out. But

0:36

global mental health researcher Sanderson Oni

0:39

thinks this can go beyond sponsored

0:41

content about kitchen tools or the

0:43

latest workout fad. In

0:46

his 2022 TEDx Sydney talk,

0:48

he dives into how tech might just

0:51

step up in our darkest moments. Imagine

0:54

if those sneaky algorithms shooting ads

0:56

our way could actually toss you

0:58

a lifeline if you're feeling hopeless.

1:02

This talk invites us to imagine

1:04

new and critical uses for this

1:06

everyday tech. And it

1:08

may just make you think twice about that next pop

1:10

up. This

1:12

episode is brought to you by Progressive

1:15

Insurance. Fiscally

1:20

responsible, financial geniuses, monetary magicians. These are the things

1:22

people say about drivers who switch their car insurance

1:24

to progressive and save hundreds. Visit

1:28

progressive.com to see if you can save. Progressive

1:31

Casualty Insurance Company and Affiliates' potential savings will

1:33

vary not available in all states or situations.

1:42

Support for this podcast comes from Odoo. Imagine

1:45

relying on a dozen different software programs to

1:47

run your business, none of which are connected,

1:49

and each one more expensive and more complicated

1:51

than the last. It can be pretty stressful.

1:56

Now imagine Odoo. Odoo

1:58

has all the programs you'll ever need.

2:00

and they're all connected on one platform.

2:03

Doesn't Odoo sound amazing? Let Odoo

2:06

harmonize your business with simple, efficient

2:08

software that can handle everything for

2:10

a fraction of the price. Sign

2:12

up today at odoo.com. That's

2:15

odoo.com. Sometimes

2:21

I'll be scrolling through my social media on my phone

2:24

and I'll get an ad. But

2:26

how on earth do they know that

2:28

sometimes I snore so loudly my brother

2:30

needs to wear earplugs the next room

2:32

over? I

2:35

think we're all familiar with this not-so-secret habit

2:37

of the tech industry, where they use these

2:39

complex algorithms to get to know who we

2:41

are, what we like, what we don't like.

2:43

Also, they can tailor ads to us and

2:45

sell us things. But

2:47

what if we could use this technology

2:50

for good? When

2:53

I was diagnosed with depression in 2015, I

2:57

had multiple intense suicidal

2:59

episodes. At

3:01

one point, it actually got so bad that

3:03

I spent the whole day planning out how

3:05

I would die later that evening. I

3:08

was probably less than an hour away from

3:10

ending my life when something saved me. What

3:13

was it? Someone reached

3:16

out. I now know

3:18

that I'm not alone, where in Australia,

3:20

suicide is the leading cause of death

3:22

for individuals aged between 15 to 44

3:24

years old. And

3:28

for every single suicide death,

3:31

there are 25 to 30 times

3:34

the number of attempts. Our

3:37

recent research has even shown us that

3:39

less than half of people who attempt

3:41

suicide will actually reach out for help

3:43

prior, meaning that if we

3:45

are to reach them, we need to meet

3:48

them where they are. I'm

3:50

now a researcher who tries to understand, how can we

3:52

do this? And the research

3:54

tells us that people, most

3:57

people who attempt suicide, or

3:59

even think of suicide, don't actually

4:02

want to die. In

4:05

fact, they're looking for any

4:07

reason to continue on living.

4:10

But what that tells us is that

4:12

any sort of interruption, whether

4:14

it's a tap on the shoulder, a

4:16

text message, or even an internet pop-up,

4:19

could potentially save someone's

4:22

life. So for me,

4:24

the question then becomes, how do we do this in

4:26

a better way? How can we

4:28

reach out in a more effective,

4:30

rapid and far-reaching manner? Well,

4:33

we have great hotlines and emergency

4:35

services in Australia, but they still

4:37

require the individual to take the

4:39

first step. But another

4:41

thing is, how on earth are we supposed

4:43

to know what a person's thinking or feeling

4:45

if they're not speaking up? One

4:48

way is through their

4:51

search behaviors. Because

4:53

you and I both know, we've

4:55

searched things up on the internet that

4:57

we would never ask another living soul.

5:02

Our struggles about suicide are no

5:04

exception. And the

5:06

research actually supports this. For example, in

5:08

one study, the investigators went up to

5:10

these individuals who had recently attempted suicide

5:12

and asked to see their anonymized search

5:14

history. And in fact,

5:16

these individuals had searched up keywords

5:19

pertaining to suicide means and

5:21

methodology, as well as surrounding hopelessness

5:23

and loneliness. Another finding

5:25

that supports this and has been

5:28

found in many different countries, for

5:30

example, in Japan, US and Taiwan,

5:33

is that the volume of internet

5:35

searches for suicide-related terms actually correlates

5:38

with the national suicide rate.

5:42

In a completely different field, they found

5:44

that us Australians, we spend about 40

5:46

hours a week on the internet. And

5:48

93% of the time,

5:50

it starts on a search page.

5:53

So could it be that the search page

5:55

be an ideal place to intervene? That's

5:58

why in many countries, including Australia, If

6:00

you search up suicide-related terms on the

6:02

internet, you'll see a hotline banner, which

6:04

is great, but the research

6:07

shows that not everybody on the verge of

6:09

taking their life wants to call a hotline.

6:11

It's not as easy as picking up a phone

6:14

and calling. Internationally, many

6:16

countries still don't have a

6:18

sustainable hotline, and even

6:20

in Australia, if you put in a

6:22

combination of keywords in the search bar,

6:24

you're not going to be shown the

6:26

hotline banner, but you're going to

6:28

get information that might actually aid a suicide

6:31

attempt. So

6:33

what if we could use the

6:35

same algorithms that these tech companies use to

6:37

tailor ads to us to be able to

6:40

identify and reach out to individuals thinking of

6:42

suicide, and we use information,

6:44

their age, gender, language, and geography,

6:47

to do so in a

6:49

deeply personal and engaging way? So

6:52

what does this look like? Imagine Amy,

6:54

a 22-year-old girl living in Sydney.

6:57

She lives with her housemates, and despite

6:59

her cheerful demeanor, she's recently been having

7:02

thoughts of suicide. She's

7:04

tried talking to her family members, but

7:06

they haven't been particularly helpful, and

7:08

today she turned to the search page to find a

7:11

way to end it all. And

7:13

as he's searching, the algorithm triggers, and on

7:15

the very first line, she sees an ad

7:17

that leads her to a landing page. And

7:20

that landing page is designed to lead

7:22

her away from a suicide attempt. Now

7:26

on that page, she's not met with

7:28

overly fluffy language or with directives, you

7:31

have to do this, you have to

7:33

do that, but rather she's met with

7:35

stories of other people who've

7:38

been where she is, and after

7:40

hearing their stories, she finally decides

7:42

to seek help through one of

7:44

the many options available. We

7:47

work together with a group of people that have

7:49

thought of suicide, that have

7:51

attempted suicide, as well as lost

7:53

someone to suicide, where together we come up with

7:56

a list of keywords that people are likely to

7:58

search, we come up with what that is. the

8:00

ad is going to say, and we come up

8:02

with the look and feel of the landing page.

8:05

Now these ads and landing pages need

8:07

to be as engaging as possible. Why

8:10

is that? Because we don't

8:12

know if we're only minutes away

8:15

from an attempt. That's

8:17

why my team and I are using

8:20

extensive targeting, for example, age, language, gender,

8:22

and geography, to be able to reach

8:24

out to these individuals in a deeply

8:27

personal way where we can use the

8:29

tailoring to show different ads and landing

8:31

pages to different people. So

8:34

in Australia, for example, what

8:36

we found is that individuals who

8:39

are in their teens, perhaps, will

8:41

resonate, and different imagery and

8:43

different wording will resonate with them than say

8:45

somebody in their 40s or 50s. For

8:48

example, what we found is that the

8:50

younger generation actually really liked the calming

8:53

and distracting resources. They really

8:55

liked the breathing exercises. Whereas

8:57

the older generation, they don't want to

8:59

see any of that, and they prefer

9:01

direct, no nonsense, straight to the point,

9:03

language. So in Amy's case,

9:05

that's what she'll see. But

9:07

imagine Robert, a 55-year-old man living

9:10

in rural Queensland. Due

9:12

to the recent economic downturn, he's

9:14

been under a tremendous amount of

9:16

stress. And because he's

9:19

the breadwinner and the man of the

9:21

family, he feels that he cannot talk

9:23

to anyone, something we know

9:25

that is not true. And

9:29

so he turned to the search page to find

9:31

out what would happen to his life insurance policy

9:33

in the event of a suicide because he's still

9:35

thinking about his family. And

9:37

as he goes there, because his

9:39

age group indicated they want direct,

9:41

no-nonsense, problem-solving language, that's what he'll

9:43

be shown. And

9:46

on every single landing page, there's a welcome video.

9:48

And the person welcoming you to that

9:50

welcome video is somebody from your age group.

9:54

This is so that for every single

9:56

person in that dark time, they'll

9:58

be shown an ad and landing

10:00

page. tailored as closely as possible

10:02

to them. Now, we've

10:04

run this trial in Australia, and what did we

10:07

find? In just 19 days,

10:09

our campaign reached over

10:11

120,000 people surging

10:16

up distress and suicide-related

10:19

keywords. Now, when they're

10:21

on the page, we want them to consume content

10:23

that is likely to lead them away from

10:25

a suicide attempt and that

10:27

is gonna promote help-seeking. Now,

10:30

the industry standard for engagement is approximately

10:32

3%. That is 3% of

10:35

individuals who land on the page will

10:37

engage with the page in a meaningful

10:39

way. We found an engagement

10:41

rate of over 22%. That's

10:46

over seven times the industry average.

10:50

Suggesting there's a real need for this type of

10:52

intervention. Furthermore, we're also

10:54

developing pages for Aboriginal and

10:57

Torres Strait Islanders, as well as Mandarin-speaking

10:59

populations in Australia. Now, we're

11:01

doing this in several different countries

11:03

in several different ways. For example,

11:05

in Indonesia where I'm from, mental

11:07

health stigma is so bad that

11:09

your own family might put you

11:11

in a makeshift cage for

11:14

having a mental disorder. And

11:16

there are over 21 unique

11:18

regional dialects. Now, we're

11:21

using geographic targeting to show different

11:23

dialect pages to different locations. For

11:26

example, if you live in Samarang in

11:28

Central Java, you'll see the Atlantic page

11:30

in Bahasa Java. If you live

11:32

in Palembang, you're gonna see the page in

11:34

Bahasa Palembang. Now, our

11:36

pages are already being used in Kidul Mountain,

11:39

which is an area with some of

11:41

the highest suicide rates in the entire

11:44

country. And we're getting reports that it's

11:46

resonating really well with people. Why

11:48

is that? Because when you speak

11:51

the national language, you speak

11:53

to their minds. But when

11:55

you speak the regional language, the language

11:57

they spoke growing up, you speak to

11:59

them. their hearts. I'd

12:03

like to leave you with this. We

12:06

have an entire industry with

12:08

the technology, the data, and the means to reach

12:10

out to people who need it, rather

12:13

than just tailor ads to them and sell them things.

12:17

And this cannot replace the

12:19

role of relationships and communities, but

12:21

given that most people still don't

12:24

seek help prior to a suicide

12:26

attempt, this could play an

12:28

important piece of the puzzle in

12:30

making sure that every single

12:32

person in their dark

12:35

time is met with an

12:37

outstretched hand the same way someone saved

12:39

me, whether it's a tap on the

12:41

shoulder, a text message,

12:44

or even an internet ad. Thank

12:46

you. That's

13:06

it for today's episode. Thank you

13:08

so much for listening. This

13:10

episode was produced by Dan O'Donnell

13:12

and fact-checked by Vanessa Garcia Woodworth.

13:16

Special thanks to Maria Lajes, Grace

13:18

Rubenstein, Farrah Dae Grunge,

13:21

Jimmy Gutierrez, Anna Phelan, Michelle

13:23

Quint, and Colin Helms. I'm

13:26

Dr. Shoshana Ungerleiter. I'll talk to you again

13:28

next week.