[ Applause ]

>> Linda M. Collins: It's been 50 years

since the first Surgeon General's report on smoking.

I was a little kid at the time, when it came out,

but I remember it really well even though I was small.

Because I remember my father's reaction to the report.

Like many smokers at the time, he was stunned to learn

that the habit he enjoyed could be lethal,

and could eventually lead to lung cancer.

He quit smoking as a result of reading that report,

and so did many other people.

Things are different now, of course,

many fewer people smoke.

It's no longer permissible to smoke in public places

like restaurants and airplanes,

but smoking remains the number one preventable cause

of death.

Not only worldwide, but just here in the United States.

In the United States alone,

the equivalent of the entire population of the City

of Atlanta dies every year as a result of cigarette smoking.

And perhaps even more disturbing,

the equivalent of the undergraduate population here

at University Park, dies every year as a result

of breathing in passive smoke.

These are non-smokers dying as a result

of other people smoking.

And think about that, it's not just the freshman,

not just the sophomores, and not just one year,

but every year, the equivalent

of the entire undergraduate population here.

Smoking is a really serious public health hazard,

and it's everyone's problem, not just a problem for smokers.

A lot of research has shown that most smokers,

the majority of smokers, would like to quit.

Nicotine is a very addictive substance,

and it's really difficult to quit smoking.

Fortunately, there are behavioral interventions

that have been developed to help people quit smoking.

A behavior intervention is an evidence-based program aimed

at helping people change a particular focused

health behavior.

Now human behavior is complex,

smoking is a complex behavior,

and so behavioral interventions

for smoking cessation are complex too,

they're usually made up of a number of different components.

So, for example, a behavioral intervention

for smoking cessation might include pre-cessation

counseling, to help an individual get prepared to quit.

Supervised practice quit attempts,

so that the person will have a sense of what it's

like when they quit smoking for real.

Counseling offered across the entire quitting process,

in person and possibly over the phone

to help the person stay on track.

And a pharmaceutical, such as nicotine replacement therapy,

and many behavioral interventions include pharmaceuticals.

In the year 2000, the nation set goals

for the year 2010, health goals.

And the goal they set for adult smoking was

to reduce the prevalence of adult smoking

from the level it was at that time, in 2000,

which was about 24 percent to 12 percent.

When those goals were reviewed in 2010,

we found that we had not met that goal.

At that time, in 2010,

the prevalence of adult smoking had dropped only

about 3 percentage points, to 21 percent.

So, when new goals were set for the year 2020,

the same goal was set, to reduce the prevalence

of adult smoking to 12 percent.

Now let's think about that for a second.

We didn't not -- we did not make that goal in the year,

for the year 2010, what are our chances

of making it for the year 2020?

I'm optimistic that we can make that goal,

but in order to make it,

we need more effective behavioral interventions

for smoking cessation.

And I think the only way we can get those more,

those much more effective behavioral interventions

for smoking cessation is to develop behavioral interventions

in a radically different way

from the way we've been doing it up until now.

Let's take another goal that was set by the American people,

one that we met much more successfully

than the smoking cessation goal.

In 1961, President John F. Kennedy challenged the nation

to put humans on the moon and bring them back safely

by the end of the decade.

We met that goal resoundingly.

In 1969, two humans walked on the moon and came home

to tell us about it.

Let's take a look at how NASA approached that goal,

and compare it with the way behavioral science is going

about developing smoking cessation interventions.

NASA knew that it had to engineer a spacecraft

that could carry people to the moon and bring them back,

and what do I mean by engineer a spacecraft?

They knew that they had to work programmatically

and systematically toward this goal of landing humans

on the moon and bringing them back successfully.

They started with a set of components that were candidates

for inclusion in a prototype spacecraft.

They examined, empirically, the performance of each

of those components, and also looked at how they performed

with other components that they had to work with.

Any component that wasn't working properly,

or didn't work well with other components,

was either revised or removed or replaced.

Only when they had a set of components

where they understood the operation

of the components really well,

they were confident that each component was making an

important contribution doing what it was supposed to do,

only then would they build a prototype rocket and launch it.

Did any of those prototype spacecraft fail?

Yes, a number of them failed.

But here's the important thing,

because of the way they worked, every spacecraft was better

than the one before it.

Every prototype was better than the one

that had preceded it.

And they worked carefully and incrementally

until eventually they had a spacecraft

that they were confident could bring astronauts to the moon.

Let's compare that with the way behavioral science is

operating today to develop smoking cessation interventions.

I said before that smoking cessation interventions are made

up of a number of different components,

so the starting point is a set of components

that are candidates for inclusion

in the behavioral intervention,

but instead of empirically examining the performance of each

of these components, weeding through the ones

that don't work, selecting the ones that do work,

replacing the ones that don't work,

behavioral science today, the state-of-the-art is

to go directly from that list of components,

put together a behavioral intervention and test it,

in a randomized clinical trial.

That clinical trial is the equivalent

of NASA's rocket launch.

What's the problem with that approach?

The problem is that if in the clinical trial it turns

out that the intervention worked, that's great,

but we don't know why it worked.

If it turns out that the intervention,

that the intervention failed, it didn't work,

we don't know why it didn't work.

And either way, we're left without knowing,

what are the steps we need to take

so that the next intervention we develop is better

than this one?

In the 50 years since the Surgeon General's report,

a lot of time, money,

and expertise has gone into development

of behavioral interventions for smoking cessation,

and there are a lot of different interventions out there.

And yet we have not been working programmatically toward

better, and better, and better behavioral interventions.

This, essentially, trial and error approach where a set

of components is immediately assembled

and tried out in a clinical trial is not getting us toward

better, and better, and better behavioral interventions.

Instead, what if we took a page from NASA's book?

What if we started using engineering approaches

to develop behavioral interventions?

I'm involved in a project to do exactly this.

This project is funded by the National Cancer Institute,

and it's a collaborative endeavor involving scientists

from the University of Wisconsin, the University of Illinois

at Chicago, and Penn State.

We've assembled a set of 15 components that are candidates

for inclusion in a smoking cessation intervention.

We're examining the performance of these 15 components

in a set of 3 randomized experiments.

These experiments are taking place

in ordinary healthcare settings,

where you might go if you needed a routine doctor visit,

but each of these healthcare settings uses the electronic

health record, so if an individual is a smoker,

that popped up at check-in,

and they were invited to take part in the experiments.

The experiments are over, we've collected the data.

We're in the process of analyzing it now.

The data will tell us which of the components is working

and which are not working.

And based on that, we'll be able

to engineer a behavioral intervention.

An example of the kind of decision we'll be making based

on the data is we'll be able to tell whether a longer

or shorter duration of nicotine replacement is better.

Our goal is to engineer a highly-effective behavioral

intervention that also costs a reasonable amount

to implement, because we believe it's important

that the intervention be not only highly effective,

but also practical to implement.

Our hope is that after we've developed this behavioral

intervention, the work of others

and possibly our future work will develop another new

intervention that is more effective,

or perhaps equally effective, but cost -- but less costly.

Or better in some other incremental way.

The idea is to just to keep getting better,

and better, and better.

Is it feasible to take this engineering approach

to developmental behavioral interventions?

Yes, it's completely feasible.

This engineering approach does not cost any more

than today's trial and error approach.

The field of engineering has developed a lot

of very efficient approaches to experimentation.

Of course, experimentation and engineering is different

from experimentation that involves human subjects,

but a number of scientists, myself included,

are working on ways of adapting approaches that are

in wide use in engineering for use

in the behavioral sciences.

Imagine if we engineered behavioral interventions.

The approach I'm talking about today can be used not just

for smoking cessation, but for behavioral interventions

for lots of other areas.

Treatment of alcohol abuse.

Treatment of drug abuse.

Treatment of obesity.

Helping people to comply better

with chronic disease regimens,

such as regimens for diabetes and HIV.

Or any other area of health behavior

that you can imagine this approach could be used for.

If we engineer behavioral interventions, over time,

they will systematically

and incrementally become not only more effective,

but more efficient, more cost effective,

and just more practical, so that they can reach more people.

Can we meet the goal of reducing the prevalence

of adult smoking to 12 percent

by the year 2020 or even go further?

I think we can if we take a page from NASA's book

and engineer behavioral interventions.

Thank you.

[ Applause ]