Retrospective
A retrospective study is one which looks back at a group of patients
through chart review or other archived system. This has the value
of allowing researchers to gather their numbers and do their analysis
without having to wait for time to pass. As long as the criteria
are explicitly laid out and the comparison tests are truly standardized,
then these can be good studies. When selecting retrospective studies
after you know the data and making conclusions direct the parameters
of the study, you may have then entered data dredging.
Prospective
This is a good way to do a study. You set the parameters, decide
how many patients need to be included to give the study power
(or think of this as meaning), and decide how to randomize them,
and THEN you start randomizing and collecting data—as time
goes on. The aforementioned parameters are not inherently necessary,
because prospective merely refers to the fact that the study was
done in a fashion where we collect our patients and data as we
go forward. In general, prospective studies that are done well
(double blind, randomized study, appropriate statistics) are weighted
very high in the reliability of the conclusions.
