| |
A
prospective study is a longitudinal study that starts with persons
not affected by the condition of interest in order to study risk
factors. Alternatively, a cohort of persons in the early stages
of illness may be identified if the goal is to study risk factors
for chronicity. These cohorts are then followed over time to identify
onsets of the condition, or other endpoints of interest (e.g.
recurrence of the condition, chronicity, or disability).
A true prospective
design requires frequent surveillance of the subjects to detect
endpoints occurring over the follow up period, and reasonably
precise dating of the occurrence of the endpoint.
Stratifying
the cohorts by risk factor status at the outset, or as the study
progresses, allows for inference about the relationship of the
risk factor to the risk of subsequent occurrence of the endpoint.
A retrospective
(also referred to as non-concurrent or historical) cohort study
uses historical records of exposure to a risk factor and subsequent
disease experience to investigate the relationship of a risk factor
to disease occurrence. For example, records documenting the distance
survivors were from the hypocenter of the atom bomb explosion
in Hiroshima were used to study the effect of different levels
of radiation exposure on the incidence of leukemia and the period
from exposure to onset of disease (Cobb
et al., 1959).
Example
|
|---|
| Von
Korff et al. (1993) used a prospective design to
determine whether elevated levels of depressive symptoms
placed individuals at risk of first onset of TMD pain,
back pain, headache, abdominal pain or chest pain. They
followed up a sample of 803 persons initially interviewed
in a pain prevalence survey. At a three year follow-up,
they found that persons with elevated levels of depressive
symptoms at baseline were more likely to report first
onset of headache and chest pain at follow-up, but were
not at increased risk of first onset of TMD pain, back
pain, or abdominal pain. For chest pain and headache,
first onset rates were highest among the chronically
depressed. The risk of experiencing a first onset of
TMD was higher among persons who experienced other pain
conditions at baseline. As the number of pain conditions
at baseline rose, the probability of developing TMD
pain increased. Similar relationships were found for
the other pain conditions as well. |
|
|