METHODS:  RESEARCH DESIGN

The design of your methods is the way you plan to answer the questions posed in your specific aims.  It is important to be clear about what you are doing, and how you are going to do it.  You also have to convince your reader/reviewer that:

1. You know what you are doing
2. You know where you will do it, and with whom.
3. You know how to do it technically
4. You know how many subjects/experiments you need to do (not too many and not too few—just right!).  Don’t collect data you will never analyze!
5. That the research you will do actually addresses the question in the specific aims
6. That the research will give you (at least some) answers to the question in the specific aims
7. That you will get data you can analyze successfully. 

Here is an example of study design.  This is to develop a cohort of young Type II diabetics in Brownsville.

We will approach several clinics and pediatricians in the Lower Rio Grande Valley to identify patients meeting our case definitions.  To do this we will search clinic databases for the past one year for ICD9 entries within the desired age group.  The most important of these clinics is the Brownsville Community Health Clinic, serving the poorest stratum of the population.  In 2001 this clinic served 17,040 patients, 51% who were below the age of 25.  Ninety-eight percent were Hispanic, and 77% had incomes below the poverty level, and 69% were uninsured (62% uninsured children and 75% uninsured adults).  We will also approach the Brownsville Independent Districts and seek all diabetic children known to the school clinics.  One of the pilot studies submitted with this grant:  “Screening of Hispanic adolescents for obesity, diabetes and associated risk factors”, will be screening a sample of adolescents for markers of diabetes and associated risk factors.  Patients identified in this study as diabetic will also be invited to join the cohort. We anticipate that these resources will lead to the identification of 50 individuals with type 2 diabetes whose disease onset was prior to age 30 years (preferably prior to age 20, but we conservatively allow for the expanded range).

When an individual fitting our criteria for admission to the cohort is identified, he or she will be approached and the study explained in the same fashion as the main cohort. The examination will be identical to that above. In addition to the participant, we will also seek to enroll their parents. In practice, it is only these young type 2 cases where it is feasible to obtain samples from the parents. These trios will be particularly valuable for family-based association tests such as the transmission disequilibrium test (TDT).