Public Health 250A - Lecture 3

Professor: Good Morning. Why don't we get started. First
of all I wanted to just say a brief word to follow up.
I got a little feed back that perhaps my discussion of
racial categories the other day may have been
indelicate. That people were put off by that
discussion. I apologize. The basic point was of course
in the United States we have enormous disparities in
health by race. Public health focuses on those
disparities to understand them and mitigate them.
Simply to point out we can't actually measure race
specific rates of anything without putting people into
boxes without classifying people into these different
groups. These are in fact partly social constructs and
they are increasingly confusing and difficult in
multiracial societies. That was really all that was
intended to do. I would point out two additional thing.
In this week's New England's journal of medicine.
There's an article on racial and ethnic disparities in
children. They compare whites and Hispanics and blacks.
And in these race categories they find that once they
control for socioeconomic statuses the differences
between whites and Hispanics disappears but the
differences between whites and blacks remain, but the
differences are mitigated by about 50 percent. You may
want to take a look at that article.
The other thing I wanted to point out walking in
the door and seeing this. Some of you may have noticed
this on campus. This occurred yesterday. What I didn't
point out the other day there is certainly a theory that
one reason there might be racial disparities or one of
the contributing factors might be racism. Perhaps the
experience of racism or how one responds to racism might
have a negative impact on one's health through stress.
One of our former doctoral students now on the faculty
of Harvard has been in the field of this research. If
there is evidence of that theory it might not matter
what box you put yourself into but what other people put
you into.
I want to come back to the issue of age
adjustment and why it's important.
>>>: I'm having a hard time hearing you.
Can you turn up the mic?
Professor: A couple of years ago I was asked -- some of
you may know the state prison system has been under
federal receivership because of overcrowding and poor
health care for prisoners in the California prison
system. And there have been many lawsuits filed against
California and previously against governor
Schwarzenegger. A couple years ago I was asked to help
with the lawsuit. Basically the prison system argued
there could not be any real problem with the health care
in California prisons because the mortality rates in
California prisoners were lower than mortality rates
among prisoners in other states and in fact lower than
Californians who were not in prison.
The question was could I help them with their
argument, the prisoners in defending against this.
The basic question was is the following. Is the
quality of the medical care in the California prison
system adequate. The prisons argued that because the
mortality rates were lower therefore the health care
must be adequate.
So, that is basically what the question is. And
so here are some of the data they provided to me. In
all U.S. prisons excluding California the mortality rate
was about 232 per hundred thousand. The Western states
excluding California, a little bit lower at 200 per
hundred thousand. The California prison system the
overall mortality rate was 172 per hundred thousand. It
is the case that mortality rates are lower in California
prisons. Does that prove health care is adequate in the
California prisons?
>>>: No.
Professor: Why not?
>>>: Age adjustment.
Professor: What are you saying in English that you
would explain to your grandmother. She wouldn't
understand what age adjustment is?
>>>: You need to see the if the California
system population is younger.
Professor: If our prisoners are younger than other
prison systems and if mortality rates vary by age this
might not be a fair comparison. Are these age adjusted
mortality rates? No, they are not. I suggested there
was perhaps confounding by age. It is a topic we'll
come back to later in the semester. Here are the
mortality rates among U.S. prisoners and primarily the
point of this is to show while the mortality rate
overall at that time was 250 per hundred thousand it
varies enormously with age. They were primarily in
these age groups in California. It wouldn't be
surprising California prisoners have a lower mortality
rate. It would not be informative in terms of the
health care. The fact they were not age adjusted. They
were crude mortality rate meant that was not a fair
comparison which you couldn't make. It's critically
important to always ask particularly mortality rates
have been age adjusted. What about this idea that
California prisoners have a lower mortality than
Californians not in prison. Does that demonstrate
health care is adequate in the California prison? I'll
show you the data. It is true California prisoners have
lower mortality than Californians not in prison
>>>: No.
Professor: Why not? Are those age adjusted? I'm going
to show you in fact when you stratify when you look at
age specific mortality rates it is still true.
People in prisons have lower mortality rates than
people outside prison. Even by age category. Yes.
>>>: Because we don't have an adequate
system outside the prisons.
Professor: Well, that may be true.
>>>: There's nothing to compare it to.
Professor: Okay. I guess that might be a useful
argument. Any other suggestions?
>>>: Cause of the death.
Professor: What's your concern?
>>>: If there's a high rate of mortality
from car accidents is not going to happen in prisons.
Professor: What's the age group primarily in prison?
They are not children. For the most part they are not
elderly. For the most part they are between the ages of
20 to 65. What are the most common causes of death in
your age group in the general community?
>>>: Smoking.
Professor: Smoking as you get older is a contributing
factor.
>>>: Trauma.
Professor: Things like automobile accidents. Do any
people in prisons die of automobile accidents? There
aren't a lot of trauma deaths in prisons. There might
be an occasional fight or murder. Trauma related
accidents are very uncommon in the prison system. They
are an extremely important cause of death outside of
prison. You would need to remove all those trauma
deaths. I'm going to show you once you do that it's
still in the case people inside prison have lower
mortality rates than people outside of prison.
Here are the data for prisoners. Here are the
data for U.S. residents 15 to 64 years of age. The age
group of most prisoners. Here you can see in these
different age groups. Prisoners, first of all you
exclude transportation deaths. In other words,
automobile trauma. Yes. And that brings it down a
little bit from 308 to 289. Once you do that you can
see different age groups prisoners have lower
mortalities, substantially lower mortality than people
in the general population. Except when you get to the
very eldest category. Right?
Any explanations for why this might be the case?
The hypothesis --
>>>: Cause of deaths as well. If you are
looking at specific things, for example, heart disease
or complicates from non treated prison. People in
prison aren't suffering from gun deaths. If you are
ruling out accidental things it might be different.
Professor: Even once you get rid of all the trauma
related deaths there's still a difference.
>>>: Less stress. They don't have to go to
jobs.
Professor: Less stress. Stress and diet. What else?
>>>: Maybe they are receiving regular
physicals.
Professor: They get good medical care you are saying
>>>: It's not up to them independently to
seek out medical care. Maybe they are unlikely to have
untreated illnesses.
>>>: Maybe it's hard to rob a bank and
murder someone if you are unhealthy to begin with.
Professor: Exactly. That's the theory I put forward.
I told you when I showed you data for farmers in the
United Kingdom. We have a healthy worker effect.
Workers on average are healthier, especially in a
strenuous job. Comparing mortality rates workers to the
general population. Workers are on average healthier.
I argue a healthy prisoner effect. You might term this
a healthy criminal effect. If you have severe diabetes
or terminal cancer you have a hard time committing
crimes that get you into prison. I think people with
many severe illnesses are underrepresented among the
people committing crimes and going to prison. It's a
variant of the healthy worker effect or healthy warrior
effect. People in the military on average are
healthier. The military won't take you if you have a
variety of underlying illnesses
>>>: Is there also possibly something to do
with lots of exercise?
Professor: It's obviously possible. I think the
primary problem making this comparison is what you said.
It's a very skewed comparison. You basically have to be
relatively healthy to commit crimes that get you into
prison. It's really not a fair comparison.
>>>: Inaudible.
Professor: I don't know the relationship to that in
relationship to drug. The point is prisoners won the
case against the governor and prison system. A, it's
critically important you age adjust. B, it's also
important you think about whether the comparisons are
fair in other ways in terms of who the underlying groups
are. That's really the point I wanted to make. I don't
want to spend a lot more time on it.
There are a number of things. We may not get to
it all today. We may push it off to Friday. Many of
theses are metrics you will see and should be familiar
with. They are really useful metrics of the healthy
population in general and health of the community.
Things such as infant mortality rates. Disability
adjusted life years, years of potential life lost.
These are important metrics for people working public
health. We will come back to issues about race and risk
and number of times. I want to reiterate the cumulative
incidence on the risk is approximately equal to the
incidence rate times time.
In 250 B you'll go into much more detail about
this. But this equation only holds true. This
approximation only holds true when the risk is
relatively low and relatively short periods of time. It
assumes the incidence density rate which is an
instantaneous concept remains constant over time. When
the rate is not constant over time as it frequently is
not you really need to use what's called a life table
approach. You should have a general understanding of
what life tables are and particularly how they are used
to generate life expectancy. All of you are familiar
with the concept of life expectancy at birth. How do we
generate those estimates? And basically you need a life
table in order to do that. Anybody who has worked in
the insurance industry will know insurance companies
decide how much to charge you on life insurance based on
your age and based on what they know to be the mortality
experience of different age groups. They are smart
enough to do the math so they don't make mistakes when
selling an insurance policy. This is an example taken
from the Rothman book if anybody is interested in doing
more reading. Pointing out why it is if you take
something like the risk and multiply it times time, over
time that will in fact diverge from what you would get
from the incidence density. Okay.
That is, if you took that formula and said, well,
the risk is 11 deaths per thousand persons. And I do
that over 20 years over 20 years 20 times 11 is 220.
There would be in a cohort of a thousand people there
would be 220 deaths. Okay? Is that true? And the
answer is over an extended period of time it is less and
less true. And that equation that approximation is not
true and the reason is because that rate of 11 per
thousand persons is being applied to a shrinking
population.
As people die there are fewer people and that
rate of 11 per thousand is being applied to a smaller
and smaller number of people. The actual number of
deaths is not 11 per year. In fact even in the first
year it's not really 11. It's only 10.9. Once you
subtract that from the thousand people you are starting
with, the population living at the beginning of the
second year is not a thousand. It's only 989. You now
apply that rate to that and in fact there are slightly
fewer deaths and this continues over the course of time.
And simply to show that over 20 years you don't end up
with 220 deaths. You only end up with 197 deaths. This
is simply to point out that this approximation between
rates and risks is only true over relatively short
period of times and for relatively low rates.
Okay. And this is shown on a graph in the
Rothman book you simply can't take the rate and multiply
it by times and come up with the risk. You'd have
what's here referred to as exponential decay. It's
simply to point out if you want to start looking at
experience over protracted periods of time the death
rate, things like that, you need to really take what's
called a life table approach. We're not going to spend
a lot of time in this course on life tables. I'm going
to skip this for the sake of time.
Life tables calculate the probability of
surviving through each successive time interval during
the period of interest. The overall survival
probability over some extended period of time, say over
decades or your entire life. The overall survival
probability equals the cumulative product of the
probabilities of surviving through each successive
interval. These life tables assume that the age
specific rates remain unchanged in the future.
So, in other words, particularly when we are
estimating life expectancy. What's a life expectancy of
a baby born in the United States in 2012? A male baby?
What's the life expectancy at birth approximately?
>>>: 80.
Professor: Somewhere in the range of 80 years. More so
for a girl than a boy. How do we know it's 80 years?
What we do is take the current age specific mortality
rates and we basically take a life table and we project
on average how long, how many people will die in each
time period. And therefore what's the probability of
surviving for a certain period of time.
Okay. Now, when we do this, this typically
assumes no competing risks in the case of all cause
mortality there are no competing risks. That's all
causes of death. I'll show you if you are interested in
something like motor vehicle accidents or risk of dying
of breast cancer there are competing causes of
mortality. There are things that may kill you before
you can die of a motor vehicle accident or breast
cancer.
This would be a typical example of a life table.
Here you can see this is for people in the first year of
life. What's their probability of dying in the first
year of life? It's actually quite high. Babies in the
first few weeks of life have a high probability of
dying. That probability of dying shrinking over time.
So in your age category it's generally quite low. And
it stays quite low until the time you get into your 40s
and 50s and by the time you get to my age the
probability of dying in any one year or 10-year time
period starts to go up and up and up. All of you
understand this intuitively. Your risk of dying varies
with your age.
If you were to take a cohort of 100000 people
born alive. Some of them will die before their first
birthday. Those that survive will have a certain risk
of dying in the next five years. Those who survive that
five years will have a certain risk of dying in the next
five years. Etc. This is really simple arithmetic.
You take the age specific risk of dying and calculate of
these hundred thousand will survive to their first
birthday, fifth birthday, twentieth birthday and
generate life expectancy. This is something you should
be moderately familiar with. What's the life
expectation? In 1970 the life expectancy at birth was
more like 72 years. Now it's closer to 80. What's life
expectancy if you make it to 25 to 30? Another
49 years. This is something you should be familiar
with. I don't want to go into the details. This is a
more current abridged life table for the total
population of United States in 2005. Somebody over the
age of a hundred will have a life expectancy. Right?
So, you can calculate the life expectancy at any
given age predicated on the assumption that age specific
death rates will continue to be the same as they are
today.
>>>: Can you explain why the probability of
dying per hundred and over is a hundred percent?
Professor: Because everybody dies. So everybody who
survives to a hundred is going to die eventually.
>>>: Right. Not everyone who is a hundred
has a probability of dying at a hundred.
Professor: Not at a hundred, but at some point. At
some point after reaching a hundred they will die. Bad
news, everybody dies. The probability of dying after
your hundred birthday is a hundred percent. But you
still have an average life expectancy. Right? That
make sense?
>>>: How do they calculate the average life
expectancy for people who have passed the average?
Professor: That's a good question. I don't know. How
did they come up with this 2.6 years, I'll get you an
answer
>>>: Do I get extra credit for stumping
you?
Professor: No, if you make it to a hundred you get
extra credit. On your hundredth birthday, call me.
What is more common, another thing that's common
is to look at what is the probability of dying of a
particular cause. So in this case it's motor vehicle
injury. What are your chances of dying of a motor
vehicle injury over the course of an entire lifetime?
And you'll particularly see this for breast
cancer. You'll frequently see the figure cited a woman
has an X percent chance of developing or dying of breast
cancer. Anyone know what that figure commonly cited is?
>>>: 12. One in eight.
Professor: I think it's higher. I think it's one in
six or something like that. Does that mean one in six
of the women in this room are going to develop breast
cancer? Why not? Because I'm sorry to tell you some of
you are going to die of heart disease or something else
before you can get breast cancer. It does not mean that
one in six of women 20 years of age will develop breast
cancer. It's predicated on the assumption that nothing
else kills you in the meantime. Right?
Simply keep that in mind. And so this would be
an example. If nothing else kills you then you can
calculate the lifetime risk of dying of a motor vehicle
injury by taking the age specific motor vehicle
mortality rate for people in this age category, this age
category, etc. And basically multiplying them out and
summing them. And if you do that you'll see you have
about a 1.6 percent chance of dying of a motor vehicle
injury if nothing else kills you before. But many of
you will die of something else before you could be
killed by a motor vehicle.
>>>: But the mortality rate you are using
is coming from a population that also had the risks of
dying from something else before they got into the car
accident. Breast cancer, you are getting that
information from people that had competing interests.
Professor: Competing mortality. These age specific
mortality rates are coming from populations where people
are dying of other things
>>>: Yes.
Professor: That's true. That doesn't change this
calculation. This says, for example, in your first
15 years of life this is the approximate likelihood of
dying of a motor vehicle accident. If you make it to
age 15, this is the probability of dying in a motor
vehicle accident per year over the next ten years.
And so on all the way through. But this doesn't
mean, this figure that you come up with at the end of
1.6 percent doesn't mean 1.6 percent of five-year olds
alive today will die of a motor vehicle injury. Many of
them will die of other things first. So do you find
this a helpful calculation? I'm not sure. You will see
it frequently given, particularly in the case of breast
cancer. I think it frequently requires this additional
explanation.
>>>: On this page all of the denominators
are 100000.
Professor: This is a completely artificial example
taking a cohort starting out at size 100000
>>>: They wouldn't shrink in the same way
they shrink in the table?
Professor: They shrink from competing mortality of
other causes. Yes.
>>>: That was my question.
Professor: Understand when you do these calculations
what's the risk of dying over a lifetime they assume no
competing mortality, when in fact there's invariably
competing mortality except if you are talking about all
cause mortality, in which case there is no competition.
One of the things we frequently have information
about is what's called the crude death rate. The crude
death rate is a number of deaths during the given year
divided by the average midyear population times a
thousand. Clearly the crude death rate reflects both
the age structure of the population and the age specific
death rates. Which is why we need to look at age
adjusted death rates. Here you can see for the United
States between 1960 and 2005 what's happened to crude
death rates and what's happened to age adjusted death
rates. So the crude death rate has gone down slightly,
despite the fact we are a much older population.
Right? And once you take that aging of the
population into effect, once you adjust the death rates
for age, you can see there's been a marked reduction in
the death rate. Make sense? On the average a little
under one percent of the population dies every year in a
rich country.
And you can also look at age specific death
rates. So these are age specific death rates for men
and women between 1955 and 2005. You'll see in pretty
much every category death rates are going down.
Basically we are a healthier society. Even taking age
into account. Right? So death rates are helpful. Age
adjusted death rates are helpful. Life expectancy is
another common metric for looking at the health of the
community.
The average number of years an individual is
expected to life if current mortality trends continue to
apply. It's a hypothetical measure based on the
assumption that current age specific death rates will
remain constant throughout the life of the individual.
It's thought to be an indicator of current health and
mortality conditions. As you all know life expectancy
is going up. Here's life expectancy in the United
States between 1970 and 2005. You can see all races men
and women the black line here you can see going up. 70
at 1970 to about 76 to 77 in 2005. You see these
enormous differences by gender and by race. These are
black males down here with the lowest life expectancy at
birth. These are white females with the highest life
expectancy at birth. One indicator of racial
disparities is life expectancy at birth. This gives you
a better sense. These are white black life expectancy
at birth. Here they are for blacks and whites. You can
see in 1900 a 14-year gap. By 2000 that had shrunk to
five years. There's copious information on the web.
You can see general life expectancy at birth has gone up
dramatically over the last hundred years. We still do
have a black/white difference.
I'm going to skip over this. It basically says
the same thing.
Basically this is the black-white life expectancy
gap for males and females. You can see while that gap
actually expanded in the 1990s it's now gone back to
approximately what it was in the 1970s.
Okay. Now if you ask the question how much of
this is due to socioeconomic status, how much of this
would disappear if you control for SES, the answer is I
don't know. Presumably much of this is attributable to
poverty.
I found this an interesting little comment a
number of years ago. This has to do with this question
of making it harder to vote.
And whether or not there would be a
disproportionate impact of making it harder to vote on
minority groups. Here's a politician who basically said
we don't need to worry about blacks being disadvantaged
by voter identification laws because everybody knows
blacks die sooner than whites. There aren't as many
elderly blacks as elderly whites. This will have less
of an affect on minorities because blacks don't live as
long. Interesting justification for this law. He later
apologized. He is correct that life expectancy differs
by race in the United States. This is pretty much the
same thing. I'm going to skip over that.
An interesting question to ask what are the
causes of this increase in life expectancy. What is
this attributable to? Now going back a hundred years
when life expectancy was more like 40 or 45 at birth to
know when it's more like 80, what have been the major
contributors to an increased life expectancy? Well this
basically says that about 70 percent, this is between
1960 and 2000, much of the increase in life expectancy,
70 percent is reduction in the rate of death from
cardiovascular disease. About 20 percent is due to a
reduction in the rate of death in infancy. Some minor
contributions. The other contributions from what are
called external causes, this is primarily trauma. From
pneumonia and influenza and to a certain extent death
from cancer. Much of it due to reduction in death from
cardiovascular disease.
Now before 1960, between 1900 and 1960 most of
this extended life expectancy was due to reductions in
risk of dying from infectious diseases, much of it due
to sanitation and public health measures. Not a lot to
do with antibiotics or vaccines. Vaccines certainly
contributed.
This is a little out of date. I don't know the
2012 data. This is the world word highest and lowest
life expectancy at birth. Japan has for many years had
the highest life expectancy at birth. Followed by
Sweden and Australia. The United States usually ranked
between 15 and 20. We are nowhere near the top and
ranked substantially below many other wealthy countries.
The contrast with the poorest countries in the world,
sub-Saharan Africa, it was less than half the life
expectancy at birth than the wealthier countries. Life
expectancy in many of these countries was going down,
not going up, primarily due to HIV and AIDS.
You can see the enormous differences in the
world's poorest countries in terms of life expectancy at
birth.
This is a little graph showing within the United
States as we've already said, this is now survival to
age 65 in U.S. whites, people living in Bangladesh and
residents of Harlem and basically showing in fact that
within the United States we have subpopulations whose
life expectancy is no better than the world's poorest
countries.
This was an article back of the time of Hurricane
Katrina. Pointing out the United States will come back
to this metric, infant mortality, ranked forty-third in
the world in terms of our rate. We'll come back to
infant mortality in a minute. This is data we teach.
This is mortality among people on the Titanic. These
are deaths per hundred. You can see these sharp
differences by gender, by age and particularly by social
class.
Whether you lived or died whether you were on the
titanic was based on your access to lifeboats. The old
business of women and children first. That was true to
a certain extent among the higher social classes on the
titanic. Not so much on the lower social classes. Here
you can see among the lowest social classes the death
rates for women, 50 percent compared to the highest
percent, five percent. For children, none of the higher
social class children died as opposed to 73 percent of
the lower social class children on the Titanic. Some
people would argue the United States is some ways is
like the Titanic. This is an article about Berkeley.
We have an affluent population. Average life expectancy
for Berkeley residents is 83. Five years longer than
the national average. Four years longer than other
Alameda County residents. They were going to yoga
classes, shopping at the Berkeley bowl and have
recreational space. Also pointing out that one of the
reasons we have longer life expectancy is the number of
African Americans who can afford to live in Berkeley is
shrinking. There have been changes in the demographic
make up of the population of Berkeley. People have
lower life expectancy are moving out of Berkeley and
that is contributing to the increasing and high life
expectancy.
I want to move from all cause mortality to cause
specific mortality. What do I mean by cause specific
mortality? Instead of what's the death rate of all
causes, what's the death rate due to lung cancer?
What's the death rate due to heart disease? What's the
death rate due to suicide? That's cause specific
mortality. Those of you who know the movie the wizard
of oz will remember this little ditty here about the
wicked witch and her being dead. What this is meant to
point out we in the United States with rare exception
are really good at figuring out who is dead and who
isn't. It's rarely a problem. But figuring out what
someone died of is necessary for cause specific death
rates. How do we figure out what someone died of?
Where does that information come from? Pardon?
>>>: Autopsy.
Professor: What proportion of deaths have an autopsy in
2012. It's way under ten percent. It's been shrinking
for decades. Autopsies are very expensive. They
require a certified pathologists. We don't do them
except in special circumstances. Rarely is an autopsy
done. When you see on CSI it's rarely done.
>>>: Isn't there a recording of statistics
on birth and death rates.
>>>: Some resident at 4 o'clock in the
morning is trying to figure out what to put down.
Professor: Are you a doctor?
>>>: Yes.
Professor: Cause of death comes from the death
certificate. Every state has a certificate of death
that has to be filled out when someone dies. There are
a number of lines that relate to cause of death.
Primary cause of death. Who is it filled out by?
Generally by someone who has no training in how to fill
out a certificate. The only doctor who might be in the
hospital who has never seen the patient and is rushed.
Death certificate are often relied upon but many studies
show they are not very good for determining the cause of
death. Nevertheless, that's the source of information
we have is the death certificate. Once you attribute to
heart disease or lung cancer or stroke you can calculate
cause specific death rate. The number of deaths from a
specific cause in a year divided by the average midyear
population. Of course the problem is both the numerator
can be off because of difficulties in assigning the
cause of death. In rare instances the denominators can
be off between census years or if we're doing it for
subpopulations.
If you take these data you can then begin to see
what are the important diseases when it comes to things
such as cause specific death rates. Again these are
somewhat out of date. I apologize. Back in 1999 and
still the case today the leading cause of death in the
United States is heart disease. Okay. The second
leading cause of death is cancer. The third is stroke.
The fourth is pneumonia. Accidents, diabetes, pneumonia
and influenza. Sorry. This is chronic obstructive
pulmonary disease. 1999 the ten leading causes of death
only two for infectious diseases and eight of them were
noninfectious diseases. As opposed to a hundred years
ago when infectious diseases were the leading cause of
death. Over the past 30 years or so the rate of death
from heart disease in all groups is going down
substantially. But they remain substantial race and
gender differences and the likelihood of dying of heart
disease. In terms of racial disparities they certainly
exist. Black males compared to white males.
And so mortality from cardiovascular disease is
declining. Why is it declining? That's the main
contributor to increasing life expectancy is reduced
cardiovascular disease. What's contributed to the
reduction in cardiovascular disease?
>>>: Medication.
Professor: Treatment is part of it. Skip over this
This is United States. These are other
countries. It's estimated 40 percent of that reduction
is due to improved treatment. 55 percent is changes in
risk factors. Reduced prevalence of risk factors. The
number one most important risk factor for cardiovascular
disease is cigarette smoking. Better control of
diabetes and hypertension. That's what they mean by
reduction in risk factors.
>>>: I have a question about environment.
Changes in the environment. Like the clean water act.
Taking lead out of a lot of products.
Professor: Cleaner air and taking the lead out are very
important. They probably don't have much to do with
cardiovascular disease. In my understanding lead has
nothing to do with the risk of cardiovascular disease.
Air pollution might. Heart attacks vary with air
pollution levels. Perhaps pollution reductions have had
some impact on this.
About half of this due to better treatment.
About half due to better risk factors and some of it
unexplained.
These are the data for lung cancer. Respiratory
cancer basically means lung cancer for all intensive
purposes. Higher rates in black men than white men.
Basically starts in 1990 the rate of death from lung
cancer has been going down among men in the United
States. That's the good news. The bad news is it's
going up in women. That due entirely to smoking. These
are the rates of breast cancer mortality. Racial
disparities with blacks having higher rates of
mortalities. Whites it's going down, in blacks it's
staying pretty much the same during this time period.
This would is an example of the age standardized
mortality rates for breast cancer showing the rates are
highest in wealthy countries and lowest in poor
countries.
This I think perhaps the more important graph.
So this is age adjusted death rates for women in the
United States between 1930 and 1997 for breath cancer
and lung cancer. You can see we have reached a
situation where lung cancer kills more women than breast
cancer. Reflecting the increased rates of women smoking
starting 30 or 40 years ago. Lung cancer rates are
declining in men. These are 98 percent entirely
preventable deaths.
>>>: Why are breast cancer rates so much
lower in developing countries yet when you look at the
United States 70 years ago you don't see any significant
change in the rates of breast cancer?
Professor: The question why are breast cancer rates low
in poor countries is a complicated answer. There may be
some under diagnosis. The rates may be lower than they
actually are in real life. There are competing causes.
There are many risk factors for breast cancer that are
different for poor countries. Number of children born,
breast-feeding practices and a variety of other things
that relate to hormone levels that are quite different
in poor countries than rich countries. I'm sure some
people argue there are environmental exposures. That's
a highly debatable issue.
A lot is attributable to fertility and hormone
levels related to that and competing mortality from
other causes. What was your second question?
>>>: No, it was part of the same question.
Professor: Okay. Obviously when we talk about
screening later in the semester, what impact has breast
cancer screening had on breast cancer mortality. The
purpose of breast cancer screening is to reduce breast
cancer mortality. I'll show you data that suggests
there had been reductions in breast cancer mortality
more recently. You can see the beginning of it here.
That trend has continued. There are other countries
such as Scandinavia report no decline in mortality
despite breast cancer screening. It's quite
interesting. These are 15 leading causes of death.
Now in the top ten suicide is in the top ten.
Only one infectious disease is in the top ten causes of
death. That's influenza and pneumonia. Cause specific
mortality is important to tell us the causes of death in
a country.
This is simply to again point out that these data
come from death certificates. When you look at how good
death certificates are for capturing various causes they
are good for capturing cancer and heart disease. They
are not good for capturing mental disorders and the
like. Death certificates which we depend on have many
problems for cause specific mortality. I'm going to
skip over this. We're going to come back and spend a
little time on Friday talking about these other measures
because they are very important. We'll take another 10
or 15 minutes on Friday to finish this discussion on
mortality. Any questions before we break? We'll pick up
on this on Friday.