0:24

So these are two different types of quantities.

Â Prevalence is a proportion,

Â its the easy one to calculate actually.

Â So we have 1,000 people in a population or a

Â village or you have a sample survey of 1,000 people.

Â And a 100 of them have an illness so you just make a numerator out

Â of 100 and the 1000 is the denominator,

Â and the prevalence of that disorder is 10%.

Â And this is important because it estimates the

Â burden of the disorder better then anything anyway.

Â If you were trying to figure

Â out about demand for services, If everyone in that...

Â Illness category needed services, then that would

Â help you estimate how many, what types of

Â services, how, how many services, the quantity

Â of services you would have to plan for.

Â So, prevalence is a proportion.

Â It's the proportion in the population with the illness.

Â That's the easy one. Incidence is a little trickier.

Â Incidence is the rate at which new cases form in the population.

Â So, it's sometimes referred to as a rate per

Â time, and it's not as good an estimate of burden.

Â But we think it estimates the force of morbidity.

Â So when I say the force of morbidity.

Â The force of morbidity is like a, a force, like gravity, pushing a ball downhill.

Â Except this force is pushing the disease

Â in the population it's creating the disease.

Â And so we will be trying to estimate where

Â the force morbidity is high and where it is low.

Â Where it is high,

Â we'll think gosh maybe we can locate the cause of that high force of morbidity.

Â And so incidence is used for ideologic studies, and trying to

Â understand the causes of disorders so we have to define the population.

Â lets take a thousand people again, and you have to define a cohort.

Â So in those thousand people, it could be

Â that 100 people have already had the disorder.

Â Well, they're not at risk for new onset

Â of the disorder. And this is oriented toward new cases.

Â So now, we have 900 people who have never had the disorder.

Â They are the so called risk set.

Â And now we're going to follow those people, we might follow them for one year.

Â So we would then say we have 900 person years.

Â And that would be our denominator, and then we estimate the number of new cases.

Â If we had, for example, I'm making this easy, 90 new cases, it would be 90

Â out of 900.

Â It would be 10%, and that would be if it occurred in one

Â year, it would be 10% incidence, and that would be an annual incidence.

Â But you can see, if we follow 900 of the people for

Â two years and that would be 1800 person years for the denominator.

Â And we would count the number of new cases might

Â be 90 and that case, the incidence would be 5%.

Â So, it's a little trickier but, and I'm going to show

Â you data for prevalence and incidence as we go along.

Â 3:03

Here's the prevalence of depression in percent.

Â It's just a simple percentage from selected surveys.

Â And there are 42 studies that we reviewed in the book, Public Mental Health.

Â And there are other studies.

Â Even as we speak, there are new studies being done.

Â And I've selected these studies to give you a feeling for the range.

Â I selected three from the United States, these are all of adults.

Â And, we have prevalence quantified

Â in two ways.

Â We have the one year prevalence, that

Â means, how many people in the population, what

Â percentage of the people in the population,

Â have had the disorder within the prior year.

Â So, for the, US, epidemiologic, catchment area study 3.5% over

Â the past year have met the criteria for major depressive disorder.

Â But we can also estimate what percentage

Â of the population have never met the criteria

Â for major depressive disorder over their entire life.

Â And we call that lifetime prevalence and that's

Â naturally higher its got more time to run.

Â And that turns out to be about, 6% in the, in the ECA study.

Â you can see that the prevalence rates range a fair amount.

Â I've selected only studies of adults.

Â And I selected the lowest study I could find,

Â which was in Taipei, Taiwan, the major city in Taiwan.

Â And that was

Â a well-conducted study, but it's only 0.6, that is less than one in 100 people.

Â Depressive disorder, according to the diagnostic criteria in one year.

Â And that contrasts with Urdmutria, Russia, which is

Â an area about 1000 kilometers east of Moscow.

Â A rural area and there are four rural village there.

Â And this again is a credible study conducted by this fellow,

Â [UNKNOWN].

Â But they found 22% even in one year and 32% lifetime prevalence.

Â So, you can see, depression varies around the world.

Â There's lots of variation, but it's big.

Â Prevalence rates are bigger than many other major illnesses.

Â 4:57

This gives you a feeling.

Â This is the one year prevalence, the 12

Â month prior from the World Mental Health Surveys.

Â And the disorders are a little varying depending on which country

Â you're looking at.

Â But you can see, we have about 10% in the United States.

Â The green is North America.

Â And then we have the European continent in brown, and African

Â and Middle East in black, and, and East Asian in red.

Â And again, we see a relatively low rate of depressive

Â disorder in East Asia, and a surprisingly low prevalence in Nigeria.

Â So I think it may be interesting, you will hear from Professor Tull

Â in a later lecture about the meaning

Â of depression, and the cultural variation in it.

Â And that, we'll connect a little bit to

Â what these state would show lots of variation.

Â 5:44

Now, lets estimate incidence.

Â So we have to have a longitudinal study.

Â And I'm going to show you in the rest of this

Â lecture, several slides about the

Â Baltimore Epidemiologic Catchment Area followup.

Â And this particular aspect of the followup is

Â from 1981, when the first interviews were done,

Â until 1996, when a followup interview was done.

Â Just to give you a feeling for the population aspect of it.

Â We defined an area in Eastern Baltimore of 175,211

Â adults, then we designed a sampling regimen, which would

Â sample each individual into a sample with a known probability,

Â that's probability sampling.

Â And we ended up with designating 4,238 people for the sample.

Â But we only were able to persuade

Â 3,481 to, to actually participate in the interview.

Â So you can see the challenges in this kind of work.

Â And in the time between 1981 and the followup, 848 of these had died.

Â And we had trouble locating another 415. That's 16% of the survivors;

Â they just had moved, and we could not find them.

Â We tried to find them so we could interview

Â the same person again, and could not find them.

Â We located 2,218 and of those, 298 people

Â said, you know I don't really have time, or

Â I'm not interested, or I didn't like the interview

Â last time, or I don't want to do it.

Â They refuse.

Â So, this is attrition which happens in a cohort study.

Â These are the major sources of attrition

Â in cohort studies in epidemiology death, location

Â and refusal, but we had 1,920 that were interviewed.

Â So we're going to construct a cohort to estimate incidence with those 1,920.

Â And this spreads out the cohort according to sex and age.

Â And if you look at the first column.

Â The at risk column, you can see instead of 1,920, we have 1,665 at risk.

Â That's at

Â the bottom of the second column.

Â Well, that's because, some of the 1,920 already

Â had depressive disorder, so they're not at risk.

Â Okay so within each sex by age group it list the at risk

Â population and the number of new cases during that 13 year follow up.

Â So for 18 to 29 year old males, they were 18 to 29

Â in 1981, and they were 237 of them and 9 of them over

Â that period develop an episode of major depressive disorder.

Â And the right hand column is

Â the cumulative incidence, that's the proportion.

Â So, 9 divided by 237 is 3.8%. That's

Â 8:29

And you can see these 71 new cases and the number is

Â red at the, in the center-left there, that we have 71 new cases.

Â And during that time there were 23,698 person years of exposure,

Â among those 1665 people.

Â We divide 71 by 23,698 and we get 3.0, actually, per thousand per year.

Â That's the incidence of depressive disorder.

Â Now one of the things this slide shows is, it shows

Â the way the incidence rate evolves over the course of life.

Â We have females in the brown dotted line and males

Â in the black line.

Â You can easily see that the annual

Â incidence per thousand there is averaging about three,

Â but for females that are age 30 it's nearly eight, it's seven and a half.

Â That's the peak incidence.

Â It's interesting that the peak incidence for females is earlier than for males.

Â You can see the incidence rate for females is about twice as high as for males.

Â That's standard in the literature.

Â And now that we understand

Â incidence as the force of morbidity, we're trying to think, why is it.

Â That, the rate of the depressive disorder peaks at the age of 30.

Â And so, that's a clue to the ideology.

Â That's one of our jobs as epidemiologists, is to develop clues for etiology.

Â And we're going to wonder, what is happening in the life course

Â at that period of time that raises risk for depressive disorder?

Â And we can think back.

Â It's not inheritance,

Â because inheritance doesn't change over the course of life.

Â But stress and loss does change over the course of life.

Â So we're looking for that possibility.

Â Now one of the other things is, if you look on the right-hand side there, there's

Â a trough and the rate of incidents at about age 50, both for males and females.

Â And then it raises, and I can tell you that when we first

Â did this analysis We saw that second bump, and we thought, that must be

Â a mistake in the data.

Â Maybe we can massage the data somehow and

Â get rid of that bump, so it'll look smoother.

Â But we couldn't make the data obey, and it showed the bump.

Â And this bump now, we're trying to figure out what

Â is it that's low and rises at the age of 50?

Â So for example, we're led to think about loss.

Â After the age of 50, there is more bereavement.

Â For example, spouses are dying, male spouses

Â are dying.

Â But it's also true that we have more,

Â physical illnesses growing in force at that time.

Â And I'm going to show you data about that later in the lecture.

Â That concludes our study of the

Â prevalence and incidence of major depressive disorder.

Â If you had to remember something about this, you might remember that

Â depressive disorder has a prevalence of about 10%, in the Western populations.

Â And an incidence per year of about three per thousand,

Â which is less than 1% per year.

Â And we'll turn next to the natural history of depressive disorder.

Â