Your worried about confounders. how a future study could be carried

  • Context: Graduate 
  • Thread starter Thread starter sosme
  • Start date Start date
  • Tags Tags
    Future Study
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
Sign up Log in
1 reply · 2K views
sosme
Messages
2
Reaction score
0
Your worried about confounders. how a future study could be carried out in which one would not have to worry about confounders. What options would you choose (ignore ethical and feasibility issues for this study - focus on theoretical) (Check all that apply)

1 We will take sisters and give one sister a late epidural, and one sister an early epidural
2 We will randomly choose mothers from a list of those that had late epidurals, and those that did not have late epidurals
3 You could randomize mothers to receive either early epidural, or late epidural
4 None of the above
5 We will match mothers with and without late epidurals based on age
6 We will take the next 100 mothers that come into the labour and delivery wardMy Attempt

I know that I would us Hill's Causal Criteria but I am stuck bwteen 1 and 3 since I think RTCs don't have confounds?. Which ones would be the way to go with reducing confounders

thank you
 
Last edited:
Physics news on Phys.org


sosme said:
1 We will take sisters and give one sister a late epidural, and one sister an early epidural
2 We will randomly choose mothers from a list of those that had late epidurals, and those that did not have late epidurals
3 You could randomize mothers to receive either early epidural, or late epidural
4 None of the above
5 We will match mothers with and without late epidurals based on age
6 We will take the next 100 mothers that come into the labour and delivery ward
thank you
Whenever you can practically and ethically randomize treatments, you should do it. You didn't specify the outcome measure, but as long as the treatment, or the lack of it, doesn't pose a danger to the subjects, randomization is the best choice.

Note while randomization is the best way to control for confounding factors, but it's not certain. You always need to check that the randomization worked so that potential confounders are evenly distributed within acceptable limits. You can also match by age within a randomization scheme and it would be a good idea to do so in this particular study assuming you are talking about obstetrical outcomes.
 
Last edited: