Questions Pts Content Criteria for pts
Sample How do you select cases and
controls for your study?
4 Cases - children aged 0-19 diagnosed with Type II diabetes from
Brno Children's Hospital clinic; cases can also come from the general
population and be screened using Hba1c, fasting glucose or similar
Need to provide full
definition
Controls - children without diabetes from the same age group and
geographical area (from pediatricians); cases can also come from the
general population and be screened.
Need to provide full
definition
Methods How do you measure the
exposure?
2 For example: FFQ or other dietary questionnaire; food diary;
interview
Can be just one or
two
What other variables do you
need to consider?
2 For example: family history of diabetes, gender, age, physical
activity, SES, BMI, hormonal disease, pubertal status
At least 3-4
variables
Bias How do you address
confounding
at the study
design stage
3 Case matching, exclusion criteria Should mention
matching at least
one
at the analysis
stage
2 including control variables (multiple regression analysis); stratifying
by subgroups; analyses employing matching; adjustment for
confounders/influential variables
At least one option
How do you deal with bias:
selection bias 2 multiple sites of recruitment (multiple pediatricians), random
sampling, stratified sampling; good definition of cases & controls;
controls coming from same underlying population as cases
At least one option
recall bias 2 providing guidelines, food diaries that need to be frequently updated,
reference portion sizes, minimize recall period, proxy reporting
At least two options
reporting bias 2 anonymity of reporting, neutral and non-judgment environment,
behavioral observation
At least one option
misclassificati
on of exposure
3 using validated instruments that represent the actual diet of the
population; confirming intake by observation; avoiding
categorization, for example into low/high
At least one option
Generali
zability
Who is the study generalizable
to?
2 generalizable to children aged 0-19 years in Brno, Czech Republic;
depends on how cases & controls selected -from clinical population
or from general population.
Should mention at
least that
Analysis What measure of association
would you use?
2 Odds ratio Need to mention
OR
Why? 2 OR is suitable for case-control study Mention that
How would you check for
effect modification by age?
2 stratified analysis; including interaction term in the model At least one option