We are hoping that you may be able to help us with a particular data
request.  We have reviewed the National Cancer Institute's SEER
database to the best of our abilities, but we are having difficulty
coming up with a bona fide answer to a question.  Our question is:

   How many adult Americans diagnosed with cancer (one or more times)
in the past 6 years are still alive today?

   This information is probably as complicated to obtain as it is
important to us.  In obtaining a reliable answer, one would have to
account for multiple diagnoses of different cancer sites in the same
person, mortality from all malignant cancer causes, mortality from
other non-related causes, etc.  Ideally, we would like the results
broken out at least by Gender and Race, and Age if also available.  We
realize that this question will likely require some advanced
"treatment" of the available data sets, so we do have a budget of $50
to support any labor that may be required on your part to produce the
answer.  We do not even need an "exact" figure -- we simply need a
credible estimate that can be supported logically IN DETAIL by the
SEER data or other reliable survey data sets.  We cannot accept,
however, mere "guesstimates" that choose arbitrary factors to adjust
for cancer incidence and mortality reports over time.

---

Request for Question Clarification by missy-ga on 07 Oct 2002 14:49 PDT

Hi there!

I wanted to let you know that your query is being investigated. 
Currently, I'm on the phone with Shellie at the American Cancer
Society trying to make heads or tails of the survivability statistics
published by the ACS.

Shellie tells me that there is a difficulty in getting the information
you seek for the immediate past 6 years - in order to assemble an
accurate survivability study, patients must have been diagnosed at
least 5 years ago, and statistics may take up to 5 years to compile
and interpret.  The most recent available statistics are for the
period encompassing 1992 - 1997.  Shellie tells me that the ACS gets
its data from the National Cancer Institute's SEER database.

Will this time frame suit your purposes?

I have uncovered a document for this time frame detailing
survivability rates broken down by site (that is, where in the body
the cancer was diagnosed) and race.  It does not deal in actual
numbers, however.  Rather, it deals in percentages of those diagnosed.

As I am not yet authorized to view the SEER database (I have
registered, and have been informed that it will take 2 days), I cannot
yet compare the ACS document with the SEER datasets.

In the meantime, I did want to let you know that your question is
being worked on, and wanted to check with you on the acceptable time
frame for the statistics.  I expect it will take several days to
investigate your request fully.

--Missy

---

Clarification of Question by factsoncall-ga on 07 Oct 2002 20:15 PDT

I am comfortable with you taking a couple of days to investigate. 
However, I hope that this does not "lock out" someone else from
tackling the problem in the interim, should they be able to handle the
request more promptly.  The information will be of very little use to
us if it comes later than October 11th.

I do not have a problem with you taking earlier data, as long as it is
in the 1990's and it still adheres to the "consecutive six year"
window that I emphasized in the question.  Also, I want to point out
that mortality can be from any cause, not just cancer mortality. 
There are many (especially older) people who are diagnosed with
cancer, but succumb to things other than cancer, such as heart attack,
accidents, diabetes, etc.  This isn't really "new" information for the
question, but I just want to warn you that knowing cancer
survivability rates will not necessarily tell the whole story (unless
they take into account non-cancer causes of death).  I am not sure
whether they do or not.

I wish you the best of luck!

---

Request for Question Clarification by missy-ga on 07 Oct 2002 20:55 PDT

Hello again,

As long as the question is not locked, it is available to any
researcher to investigate.

The statistics I duscussed with the representative at the ACS focused
on cancer survivors - those diagnosed, but still living after 5 years.
 THe NCI doesn't deal with non-cancer related issues in their surveys.

I'll be able to look into this in greater detail when my authorization
to use SEER comes through (or if another Researcher is interested in
the question and has registered with SEER, s/he can do so).  I do want
to stress that we can't guarantee that we'll find what you need, but
we will certainly give it our best shot!

If you don't mind my asking, what are you plannng to use the data for?
 This may help me in narrowing the search, and may also open up ideas
for related data that may be unconvered along the way.

--Missy

---

Clarification of Question by factsoncall-ga on 08 Oct 2002 05:57 PDT

The population estimate will be used to project (weight) a sample of
500 cancer patients and up-to-6-year survivors.  This way, the data
from our independent survey can be "projected" to the "4.35 million
adult cancer patients and near-term survivors" (for example).

I should re-state that we already have access to the SEER database
(SEER*Stat) and have used it rigorously over the past few months.  We
have been unable to gain a reliable answer to our question using this
database alone.  But, you are more than welcome to give it a try on
your own!

I'm not sure I understand (okay, I'm pretty sure I DON'T understand!)
what you are going to be doing with the data. But I don't think that
statistics as you have requested them will be very useful for any
purpose.

Cancer survival rates have a LOT more to do with both the specific
type of cancer, and the "stage" of the cancer when treatment is
received, than with the Age, Race, and Gender of the patient when
diagnosed. For example, in the case of breast cancers of the type
"invasive carcinomas":
"The five-stage TNM staging system is an improvement over general
staging. It categorizes noninvasive carcinomas, which are highly
curable, as stage 0. For staging invasive carcinomas, it places
emphasis on size of the primary tumor and the extent of nodal
metastasis. Excluding the incurable cases with distant metastasis
(stage IV), the remaining stages of invasive carcinoma (stages I, IIA,
IIB, IIIA, and IIIB) have an increasing likelihood of treatment
failure and death. Five-year survivals for stages I to IIIB are 90,
80, 65, 50, and 40 percent, respectively."
(http://www.danaise.com/breast.html)

In other words, if I was diagnosed with Stage IIIB Ductal Carcinoma In
Situ, it doesn't matter if the all the Caucasian 40-something women
who were diagnosed with any kind of cancer in the last 6 years had an
overall survival rate of 78%. For me, there is only a 40% probability
that I will still be alive in 5 years.

By your GA handle, I presume that this information will be used for
some kind of web or telephone information database, or possibly for
marketing. If that is the case, providing cancer survival statistics
based simply on Age, Race and Gender is probably a really good way to
get sued, and would certainly NOT give your company any credibility
with the medical community.

If you can be a little more specific about the intended use of the
statistics, it might help GA researchers to do a better job of getting
you the information you really need.

I hope my comments have been of assistance to you!

Dear AceResearcher:

As I have already clarified...
The population estimate will be used to project (weight) a sample of
500 cancer patients and up-to-6-year survivors.  This way, the data
from our independent survey can be "projected" to the "4.35 million
adult cancer patients and near-term survivors" (for example).

We surveyed 500 people by telephone who said either (A) that they were
currently undergoing treatment for cancer [which would have a high
probability that they were diagnosed with cancer in the past 6 years],
or (B) that they were in remission and had been so for 1-5 years
[which again would have a high probability that they were diagnosed
with cancer in the past 6 years].  People who had been in remission
for 6 years or more were terminated from the study.

So, we interviewed cancer patients and survivors of ALL types who
happened to be alive (obviously, they completed the telephone
interview).  All we are seeking to do is understand HOW MANY PEOPLE in
the United States fit into the same qualifications of those 500 who
completed our survey, so that (as is industry practice) instead of
having to say "140 of our 500 respondents were white females", we can
rather suggest that "our 140 white females surveyed likely represent
2.7 million white female cancer patients and near-term survivors in
the United States today".

We are asking for non-site-specific data (as your Stage IIIB Ductal
Carcinoma example suggests we should) because our survey was
completely random, and not site-specific.  With 500 completes, we are
comfortable that any sampling errors in terms of cancer sites would
have been normalized by the large sampling.  So, we want "total
cancers" statistics, because that is what we feel we surveyed.  There
is no reason to believe that we would have oversampled Stage IIIB
Ductal Carcinoma survivors, or any other particular type of cancer
patient.  SEER data is available based on "All Cancer Sites", so what
we were doing was certainly not preposterous.

We are a MARKETING RESEARCH company, and we are bound by CASRO
regulations to never turn our survey lists into "some kind of web or
telephone information database, or possibly for marketing", as you
presumed.  I won't take offense at the comment, because so many people
mistakenly correlate "telephone research" with "tele-marketing".  We
do the former, never the latter.  Our company is 19 years old and has
never been sued.

On this particular project, we are not seeking credibility with the
medical community as you suggest.  The project is for the purposes of
public relations and to raise insightful SUGGESTIONS about what is
important in the feelings and opinions of cancer patients, versus
oncology nurses, versus the general population.  The research was
intended to be only "directional" in nature, not scientifically
binding.

Sorry to sound so defensive, but your comment has a tone that we are
sort of wasting both our and other people's time, which I patently
feel is not true.

I apologize if my tone appeared to be rude. That is not at all what I
intended -- I was striving for an informational tone. I truly was just
trying to be helpful.

I was not wondering if you were going to use this information for
telemarketing, but rather if you were developing a general information
database that people could access on the web (or by an 800 number) for
simple answers to questions such as "I just found out I have cancer:
what are my odds of survival?", or developing marketing data you could
provide to companies who have services or products that would appeal
to cancer survivors.

I was not suggesting that you look for body site-specific information,
but rather disease-specific information: there are a huge number of
different types of cancers, and unlike breast cancer, many cancers
(such as melanoma) can appear in any one of a number of places in the
body. I was also not suggesting that you were wasting anyone's time,
but rather that it was possible a different approach might be more
effective depending on your needs.

Because your expanded explanation makes more clear your intended use
of the information and the fact that you are wanting to project
emotional -- not medical -- correlations to the 4.35 million adult
cancer patients and near-term survivors, I now understand why you are
looking for age, race and gender information.

Again, I apologize that my intent was not clear from the way I worded
my response. Unfortunately, that is a deficiency of electronic
communication -- we are unable to see each others' faces and hear the
tone in each others' voices.
:(

I regret that I could not be of more assistance to you. I hope that
SEER will be able to provide you with the information you need.

Best wishes!