Present:

• Hard copy of instrument
  Electronic copy – 10 subjects for comparison / validation of response

• Variable for fasting (Table 2 clarify)
  Code as dummy variable? Check by site
  Smoking status
  Physical activity
  Medications that influence cholesterol
            Hormonal/thyroid

• Data to recode for yes/no format if subject takes medication or not. 
  Can’t include dose

• Next Step
  Biochemical data will be needed for those who haven’t delivered to URI.

Smoking status
Energy output
Medications
Fasting status
Supplements – diet
          Used different collection forms.  Will merge information into simple matrix.

• Data to be recoded to address yes/no vitamins. 
  If used form, will pull from or else re-code data from sites to match form
  

• Concern:  Some supplements prescribed, so not addressed by subject response.
  Geoff G. will send supplement data form for coding.

Form A – pg 2: Other supplements
Ask if sites have other supplements – 20 items
Option to track supplement use over time

• Outliers: Who is in/out of data set

-          Guidelines to ID suspicious values: Kal>5000; <5000

-          F/V Servings

-          Fat

• Correction / measure errors

• Missingness – What to do with missing data

• Stat analysis – GLM Mixed Method R/F

• Bias

-          Social desirability

-          Gender/age

-          BMI

-          Race/ethnicity

-          Region/center

-          Smoking

Summary of edits for outliers
Exclusions of extreme values
Discussion of assumptions

• Conference calls to be set for once a month.  Tuesday – around 5 EST.  Date to be determined by NCI.

• Question – How to proceed on level of publications?

  Analysis Demands
  Plan for overall process
  Prototype proposed by Claudio as general BCC Guidelines

• Access to data/publication
  Priority measures about sharing

• Quality Assurance Issues for analysis

Sue Rossi / Melina coordinates nutrition data from URI

• Deadline for next steps -- by next call agenda
  Issues related to F/V

o        Demographics

o       Other areas ethnicity

o       Intervention effects

o       Factors associated with consumption

o       Bias in measures

o       Dietary intake/recall

§          value for instrument vs. recall

§          difference = theoretical bias of instrument – so BMI + gender as predictors of energy intake didn’t effect bias

Look at Diet, Exercise, and Smoking.  Is there interaction?

•  Synergy between diet, exercise and smoking behavior – potential interaction
  Mediators in social support, depression

•  Data request for 3^rd Round

o       Social desirability

o       Kristal questionnaire

3rd Round.  Need enough comparable data to be able to handle inquiry.
If comparable, then other measures can be considered also.
Data collection for baseline ends by end of summer.
Project ends in March for many sites

•  Question: How to Handle Missing Baseline Data / Follow Up

o       Victor/Fran as consultants in future conference call on topic.