How do Screening Tests Become Standard Tests?
Using chance to assign people to groups means that the groups will probably be very much alike and that the trial results won't be affected by human choices or something else.
Nonrandomized controlled trials
In nonrandomized clinical trials, volunteers are not assigned randomly (by chance) to different groups. They choose which group they want to be in or the study leaders assign them. Evidence from this type of research is not as strong as evidence from randomized controlled trials.
A cohort study follows a large number of people over time. The people are divided into groups, called cohorts, based on whether or not they have had a certain treatment or been exposed to certain things. In cohort studies, the information is collected and studied after certain outcomes (such as cancer or death) have occurred. For example, a cohort study might follow a group of women who have regular Pap tests, and divide them into those who test positive for the human papillomavirus (HPV) and those who test negative for HPV. The cohort study would show how the cervical cancer rates are different for the two groups over time.
Case-control studies are like cohort studies but are done in a shorter time. They do not include many years of follow-up. Instead of looking forward in time, they look backward. In case-control studies, information is collected from cases (people who already have a certain disease) and compared with information collected from controls (people who do not have the disease). For example, a group of patients with melanoma and a group without melanoma might be asked about how they check their skin for abnormal growths and how often they check it. Based on the different answers from the two groups, the study may show that checking your skin is a useful screening test to decrease the number of melanoma cases and deaths from melanoma.
Evidence from case-control studies is not as strong as evidence from clinical trials or cohort studies.
Ecologic studies report information collected on entire groups of people, such as people in one city or county. Information is reported about the whole group, not about any single person in the group. These studies may give some evidence about whether a screening test is useful.