From a new member Swapnil Rajpathak
Biren
I would certainly like to be a part of this wonderful effort. I am
trained physician from India and my clinical specialization is
Diabetes. I do not practice clinical medicine in the US but have been
trained in public health (Masters in Epidemiology from George
Washington University and Doctoral Degree in Epidemiology and Nutrition
from Harvard University).
Recently, I finished my doctoral degree and currently I am a junior
faculty member in the Dept. of Epidemiology and Population Health at
the Albert Einstein College of Medicine in NYC. My goal (as for every
epidemiologist) in the future is to start a cohort and specifically a
South Asian Cohort.
As you may know that diabetes and cardiovascular disease is a major
concern in our community and substantial evidence suggests that South
Asians are more susceptible to central obesity/insulin
resistance/metabolic syndrome. Many epidemiologists with interests in
South Asians may have individually thought of starting a cohort for
long-term work. There are many issues that come up before actually
establishing a cohort and this effort needs collaboration between
physicians, epidemiologists, nutritionists, community organizations and
of course members of the community. If SAHRI can bring together people
with this interest, this may be a possibility in the near future.
Given that health disparity is a top priority in the US, I assume that
the funding for research in minority population is adequately increased
in last few years. Please let me know your thoughts and suggestions.
Regards,
Swapnil
P.S. In the month of June (I think) 2005, there was a 1-2 day event on
South Asian Health at the UMDNJ. Is anyone aware of this? It seemed to
be many of the presenters were already working on issues realted to
South Asian Health
Biren
I was at the UMDNJ conference and had the opportunity to speak with
some of the organizers (including couple of people from SAPHA board,
Dr. Jayasankar, Dr. Arora) all of whom were very enthusiastic about
SAHRI (of course with their healthy skepticism).
I actually had a power point presentation that I gave a few people
there - it had four goals (review of literature, designing
EPIDEMIOLOGICAL study, provide a collaboration platform, and work with
community organization).
Unfortunately I don't have the training - so I looked around and
contacted a social science researcher Dr. Deborah Carr (who is on this
list-serv) at Rutgers University who gave me some good advise
including: trying to contact NHANES, CDC etc. to see if data that they
have already collected has "South Asians" as a category in demographics
(so that we could "oversample"). Unfortunately it was met with
failure.
So, She gave me a few pointers and a book on how to design studies.
But as you can imagine it is very hard to start something from scratch.
With your background, and as you are interested, I think we can take
this to the next level of planning stage -- Some of the questions that
I have had (and they may not be the best questions to ask in early
planninng phase, but I will ask them anyway).
- What would be the scope of the study? For example, when I started
thinking about it, I thought it should be a very thorough survey to
include all aspects of health (demographics data, habits, existing
health, diet, exercise, mental health, end of life preferences etc).
- How would you administer it? - there are few options - may be a
cohort of college students (follow them longitudinally) - benefit:
likely to use internet which may be advantageous for planning phase.
VS. cohort of physicians (get their information from parent
organization like AAPI, APPNI, and others) and randomly select them -
again benefit: they may value what we do and more likely to
participate.
- If it is internet based, it cuts down on costs and easily imported
into a database for analysis (no data entry).
IF we decide to do survey (not on internet) then the issues are: how do
you randomize, how do you recruit? How do you ensure completion? The
cost of administering the survey etc. (of course we can use handheld
computers to make sure that we avoid the cost of data entry). We can
also have the surveys in their native language. (of course, tracking
the cohort over years will be an issue as well).
I think that's it for now. I just wanted to jump start the discussion.
It may be slightly irratic for your liking, but take this opportunity
to lead the discussion and move it forward.
Look forward to hearing from you (or others intersted).
Biren
PS. I came across a that might of a use in future: there is a software
(naam pehchan) that someone in UK has developed which is able to
extract data from telephone directories and select likely south asian
names.
Swapnil Rajpathak
These are my thoughts and ideas based on my understanding.
What would be the scope of the study? For example, when I started
thinking about it, I thought it should be a very thorough survey to
include all aspects of health (demographics data, habits, existing
health, diet, exercise, mental health, end of life preferences etc).
Establishing a prospective cohort is the best option when studying a
specific population and trying to understand the risk factors and
common disease outcomes. The scope of the study can be anything
depending on what are the objectives.
For this South Asian cohort, the important diseases are cardiovascular
disease, diabetes and some cancers. There may be other outcomes of
interest but this can be easily incorporated anytime. At baseline and
during the follow-up participants typically report their diagnosed
medical illnesses on questionnaires that we send and so we will have
information on all common outcomes.
In terms of risk factors, we need information on following
- family history of medical illnesses
- Diet - the best option is a food frequency questionnaire (FFQ). This,
however, is culture specific and needs some literacy. Nutritionists
working with South Asians can easily compile list of commonly eaten
foods and we can test this is a group and then validate it with diet
records. We can discuss more details when time comes.
- Information on other lifestyle factors e.g physical activity
- Anthropometric data - weight, height, waist and hip circumference
- Blood samples to be collected at baseline to measure basic
biochemical parameters - glucose, lipids etc
- Also collect and store blood samples for future use
- Other issues - duration of stay in the US, diet and lifestyle when
in India
Eventually everything we collect needs to validated for its accuracy in
some way.
- How would you administer it? - there are few options - may be a
cohort of college students (follow them longitudinally) - benefit:
likely to use internet which may be advantageous for planning phase.
VS. cohort of physicians (get their information from parent
organization like AAPI, APPNI, and others) and randomly select them -
again benefit: they may value what we do and more likely to
participate.
The crucial step is selection of the population and to obtain a list of
potential participants. Having worked with the Nurse's Health Study
and Health Professional's Follow-up Study, I certainly think that
such professionals will be best source due to their literacy and
potentially higher commitment to such effort. Follow-up is long term,
hence it is good to have smaller sample size and higher follow-up rate
as compared to larger sample size and poor follow-up.
Physicians are perhaps a good option and we can get the list of names
and contacts from organizations like AAPI and also a list of physicians
from Pakistan, Bangladesh, Nepal and Sri Lanka.
- If it is internet based, it cuts down on costs and easily imported
into a database for analysis (no data entry).
IF we decide to do survey (not on internet) then the issues are: how do
you randomize, how do you recruit? How do you ensure completion? The
cost of administering the survey etc. (of course we can use handheld
computers to make sure that we avoid the cost of data entry). We can
also have the surveys in their native language. (of course, tracking
the cohort over years will be an issue as well).
Yes, internet based in good option but that will be for questionnaires
only.
This is an observational study hence there is no issue of
randomization. We select the people from a list and invite them to
participate and hopefully, we will have a minimum of 70% response rate!
Using physicians will be good so that we can conduct everything in
English and it can be easily standardized.
......
I am not sure where to start but it may be first important to
- Find investigators/faculty interested in this
- find organizations like SAHRI
- find lists of organizations like AAPI
- Literature review on South Asian health in the US
- Since this will be a multicenter study, identify centers/cities that
are like to have higher recruitment and also investigator/s in that
location
- Pilot data is most important and funding will be required to carry
this out. Pilot data will include selection of small group, testing a
FFQ, understanding common diseases and risk factors which will help us
understanding the feasibility.
Swapnil
Biren Saraiya
Attached within is a word file that has several discussions I had with Dr. Kothari and Dr. Carr about a study like this. It was a novice's stab at a large project with some advise from local experts. Please feel free to edit the file and go back and forth with the changes (please ensure that you "track changes" - again to understand who made what changes).
The file includes two things:
1. Discussion similar to what we are having here - but with a few other points.
2. Sample survey.
I have also created a wiki page for the discussion (as to keep track of tasks that Swapnil has pointed out. Check out http://academia.wikicities.com/wiki/SAHRI:Cohort_Study).
Biren
Biren Saraiya
Biren