Public Mental Health is the application of the principles of medicine and social science to prevent the occurrence of mental and behavioral disorders and to promote mental health of the population. This course illustrates the principles of public health applied to depressive disorder, including principles of epidemiology, transcultural psychiatry, health services research, and prevention. It is predicted that by 2020 depressive disorder will be the most important cause of disease burden in the entire world! Every human being suffers from feeling depressed at some point or other, but only about one fifth of the population will experience an episode of depressive disorder over the course of their lives. This course illuminates the public health approach to disease, and the particular complexities of applying this approach to mental disorders, using depression as the exemplar.
Lecture 1D: Major Depressive Disorder and Medical Conditions
Loading...
来自 Johns Hopkins University 的课程
重性抑郁人群：公共卫生解决方法
110 个评分
从本节课中
What is Major Depression and Why Is It Important? The Burden of Depressive Disorder in the Population
This week, we will be focusing on the definition of major depression and the methods that we use for measuring its effect on populations.
与讲师见面
William Eaton, PhDProfessor, Mental Health
Bloomberg School of Public Health
Wietse A. Tol, PhDAli & Rose Kawi Assistant Professor, Mental Health
Bloomberg School of Public Health
Ramin Mojtabai, MDAssociate Professor, Mental Health
Bloomberg School of Public Health
Folks welcome back this is section D.
We're sort of expanding our discussion
of the natural history of depressive disorder.
What I showed you earlier what might be called the symptomatic course.
But now we're going to consider other possible
ramifications or elaborations or consequences of depressive disorder.
Especially medical conditions.
Here's non-insulin dependent diabetes onset in the Baltimore ECA Follow-Up.
This is the same follow-up I've been telling you about.
And when we did this study in 1996 there were about 25 or 30 even 35
studies of the prevalence of depression and depressive disorder among diabetics.
And it was well known that depressive disorder was
much more prevalent among diabetics than among the general population.
And the presumption was that the diabetes was producing the depression.
That it's sad to be diabetic, you discover you're diabetic.
You have to change your food, you can't
eat good stuff, you're looking forward to twenty
times the risk of having your foot amputated,
you know, impotence and stroke, and so forth.
I'm diabetic myself, when I told my father I was diabetic
and the doctor asked me, you know, to ask whether there
had been diabetes in the family, my father said well your
grandfather had diabetes but don't worry he died of a stroke.
Well, stroke is a consequence of diabetes so it wasn't very comforting.
But, so the point is, this was the first prospective study, and again this are
cohort approach, except now we're looking at the incidence of diabetes.
So when we start in our baseline survey, we're taking
all the people who already have diabetes and setting them aside.
We're only study people at risk
for the development of diabetes.
And we take that group and say which
of those people has had an episode of depression.
And we have other categories of mental disorder for example.
Who among the people who are at
risk for developing diabetes have had panic disorder.
Have had alcohol disorder, obsessive compulsive disorder.
So we can line up little pieces of psychopathology there and
see whether they predict to diabetes that only depressive order does
with an odds of 2.2.
So you can see it isn't being sad that's the dysphoric episode.
That has no effect on its risk for diabetes in this study.
And you can see the depression syndrome which is a collection of
symptoms but not meeting the criteria
for depression has relatively little influence.
Panic disorder, phobic disorder, none of those it's just major depressive episode.
Now one of the problems with this study is that
we aren't really diagnosing diabetes except by self report.
So maybe the folks with depression are reporting more diabetes you could say.
You could worry about that people would criticize that.
So this person Norito Kawakami a couple of
years later did a replication, a similar type study.
It was an eight year follow-up. In Japan.
It was males. They were at risk for
diabetes onset and he did not use the same measure.
He had the Zung Depression Scale.
But turned out people who had a severe rating on the Zung Depression Scale had
2.3 times the risk for onset of diabetes as people who did not have that rating.
And if you remember the study that I just showed you had 2.2, 2.2 and 2.3.
And now he's clarified about the advantage of this
study, the big advantage of this study, is that Kawakami used oral
glucose tolerance tests, which are the
benchmark, the standard way of defining diabetes.
Had no influence in self report, it was all oral glucose tolerance tests.
That's how he defined diabetes, so this was a very fortunate replication.
So this is an illustration of a so-called forest plot, its a horizontal forest plot.
This is
one of the tools you'll see in
epidemiology and public health, and medicine in
general a lot to summarize a number of publications, all on the same topic.
And I suppose you have to very imaginative to see that if you tilted
this 90 degrees it might look like a forest, but I think it's a stretch.
But nevertheless, it has the same form.
And you can see, I summarized the study in 1996 that we had done, that's the blue
study there. And the Kawakami study in 1999.
And what the forest plot works is it shows the authors on the left and the dates,
and then they X axis there is the odds
ratios associated with depression as a predictor of diabetes.
You can see that the square there, the small square and a third of the way
across from, from the in the blue line there on top is a little square that's
at 2.2.
And then a little just below that, Kawakami's odds ratios 2.3.
Those are the data that we just looked at.
And, and you see a blue line with an N that's horizontal.
Those are the 95% confidence intervals in that data.
And another reason that Kawakami's replication was so nice is
that that first study was not actually quite statistically significant.
That is the odds ratio, the 95% confidence
intervale for the odds ratio included the value of one which is no relationship.
So it could have been due to sampling error.
But with Kawakami the lower end of his confidence
interval there is just above 1.0 so that's more comforting.
But you can see that this has been done now a dozen or so times.
These are all population based studies, they are not
based in clinics, they're all longitudinal, that is the perspectives.
So we have an incidence
of diabetes, that's the force of morbidity for diabetes being studied.
And adding all these studies together we get more than 1,715,000 person
years of exposure And 6,440 new cases of diabetes.
And one of the things that's remarkable about
this is, all the estimates are nearly the same.
The squares are all between about 1.5 and 2.5.
It's remarkable.
When you look as forest
plots, in general, they will not lo, look this similar.
They will not be that close together.
But this one does. So, now.
We reinterpret the literature on depression and diabetes.
People with diabetes are depressed, not because they have diabetes, but because
the depression has led to, or been an antecedent to, the diabetes.
Now we have to wonder, what is it about depression that raises
risk for diabetes?
This is the same set of data but it just shows
you which of those studies were conducted with oral glucose tolerance tests.
And three of them were. So this again is comforting.
The relationship is not really due to
self report but it's actually a true relationship.
History of depression predicts onset of diabetes.
Now, I could show you in great detail other
medical conditions and forest plots for these other medical conditions.
I'm just going to show you a few to give you a feeling for
why people are more interested in depressive
disorder than they were twenty years ago.
So in our Baltimore ECA follow-up we had 89 new cases of type two diabetes.
That's the analysis that I just showed you, and there were 1,715 at risk.
And the relative risk was 2.2, and I told
you, that relative risk was not quite statistically significant.
That's why it's not in bold.
But we've done the same kind of study for heart attack, cancer,
stroke, arthritis and actually other conditions not shown here on the slide.
And for heart attack, people with a history of depressant disorder
have a raised risk of heart attack, and it's statistically significant.
That's why the 4.5 is bold.
And you can see in our cohort study
there were 1,551 people at risk for a heart
attack and 64 new cases and if you have
a history of depressive disorder you have a 4
and half times the chance of having a heart attack.
For stroke you have two and half times the risk
for having a stroke after an episode of depressive disorder.
Now if you look at the grow for arthritis, that's the bottom row.
That's the one we were certain that we were going to get a positive relationship.
We did not.
This is something called publication bias.
We could never get published the fact that we
did not find a relationship with depression predicting to arthritis.
We were able to publish about cancer. The relative risk for developing cancer.
All types of cancer put together into one category is 1.0.
That means there's no relationship of depression predicting cancer.
It is important to understand that the
different types of cancer have different idealogies and
they might actually relate in different ways to
depressive disorder, but we can't really show that here.
Now, on the right hand side of this table, we might
wonder, well, does type two diabetes, does it predict to depressive disorder?
And you can see that the 71 new cases or the 71 new cases I show
you earlier, that the new cases of depressive
disorder they are at risk for depressed disorder.
Does a history of diabetes predict to onset of depressive disorder?
Well the relative risk is 1.1, not very important even
people with a heart attack its relative risk is 1.7.
Of course, now this is a 12 year relative risk, so we're letting the, the
effects of the heart attack show up even after 12 years in diabetes and so forth.
It seems logical that people would have fright and
sad mood, a reaction, to having a heart attack.
But after
12 years it doesn't really raise the risk for
depressive disorder, and this may be important that the,
with using the diagnostic criteria of depressive disorder not
the sad mood that may happen after a heart attack.
You can see that the relationship of arthritis predicting to depression
is actually very weak, even though arthritis involves lots of pain.
The strongest finding for predicting depressive disorder is connected
to stroke.
And we know from a clinical literature and clinicians will
tell us that people with a stroke, especially if the
stroke was in the left hemisphere of the brain, they
are very likely to have a episode of depressive disorder.
So there's kind of a brainy
connection between stroke and having depressive disorder.
And we've shown this in our population based data in
which were actually asking people about history of stroke and asking
people about history of depressive disorder.
We don't have a measure a clinical measure other
than just the interview with them but we still
get this finding of 8.4 times the risk of
depressive disorder for people with a history of stroke.
So that concludes our brief report of
our effect of depressive disorder on medical conditions.
And this literature has emerged in the past about 20 years, and makes people
understand the importance of depression.
And in fact, we wonder if we
could prevent depressive disorder or treat depressive disorder.
Would we then lower the risk for these important medical conditions?
And there are control trials out, trying to figure that out.
Whether the treatment of depression, the primary disorder, would lower
the risk for a, something like heart attack, the secondary disorder.
In the next section, we'll talk about a
quantitative estimate for the burden of depressive disorder.
