S******9
发帖数: 2837 | 1 differential follow up bias？
详细解释一下吧
谢谢 | b******a
发帖数: 704 | 2 Cohort studies are susceptible to bias by differential loss to follow-up.
For example, the patients received OCP prescription maybe much easier to
follow up or stay in HMO over a long period time than the other
group.
Lead time is the length of time between the detection of a disease (usually
based on new, experimental criteria), not relevant here.
Choices C-E sound much less likely in a cohort study. Just my 2 cents. | d****y
发帖数: 2180 | 3 Agree. Only 40% of the participants enrolled the same HMO 20 years later,
these participants are easier to follow up. Obviously you can get more
information from these participants than other participants that are not in
the HMO network.
Incidence of breast cancer reported in the question stem: 12/100 ( OCP use)
Vs 4/112 (no OCP). It seemed like OCP is a risk factor for breast cancer,
but noted that these info are only from 40% of the participants. The results
from other 60% of participants will have a significant impact to the
incidence of breast cancer in this group of women. You won't have a valid
assessment when 60% of the participants' info are missing.
The exposure in this study: OCP use
The outcome: breast cancer
Both OCP use and breast cancer(histologically confirmed) are objective
measures,
and these measures are taking prospectively, so the occurrence of
misclassifications is low. Misclassification is more common in a
retrospective study, such as a case-control study.
Recall bias is also more a problem for a retrospective study, not in a
prospective cohort study.
usually
【在 b******a 的大作中提到】
: Cohort studies are susceptible to bias by differential loss to follow-up.
: For example, the patients received OCP prescription maybe much easier to
: follow up or stay in HMO over a long period time than the other
: group.
: Lead time is the length of time between the detection of a disease (usually
: based on new, experimental criteria), not relevant here.
: Choices C-E sound much less likely in a cohort study. Just my 2 cents.
|
|
