Basic Statistics

6 INITIAL STEPS

Casdcont rol

end beginning

Cross-sectional

Surveys

beginning Experiments

beginning Clinical trials end L

cause outcome

?cause

I beginning Field trials end-

WUSe outcome

outcome

Prospective studies

beginning

outcome

Past time Present time Future time

Figure 1.3 Time relationships in biomedical studies.

and field trials are special types of experiments. Finally, biomedical studies looking

backward in time are called casekontrol or caseheferent studies. These studies are

displayed in Figure 1.3. The arrows denote the direction of the measurement, and

the question marks are placed where there is uncertainty in the factor measured. The

word cause can be either a (1) treatment or preventive measure or (2) the occurrence

of a causal or risk factor. The effect is the outcome of the risk or causal factor or

a treatment.

1.3.1 Surveys

Surveys or cross-sectional studies are used in public health and biomedical studies.

In these studies, the participants are measured only once, so information on exposure

or risk factors, treatments, outcomes, or other factors are all obtained at the same

time. Surveys are used when both the disease (outcome factor) and the supposed

risk factor (causal factor) have a high prevalence. A survey of a prevalent condition

such as depression or osteoarthritis is appropriate. On the other hand, stomach cancer

is a disease with a fairly low prevalence, so if a survey were taken of the general

population, it would have to include a huge sample just to obtain a handful of cases.

Obviously, chronic conditions are easier to study with surveys than are acute diseases.

Sometimes, surveys are done using populations that are at high risk for a given disease,

such as the elderly. In the public health field, surveys have been used to study the

association of possible causal factors such as smoking and lifestyle on overall health

status and risk factors for infectious diseases.

One problem with surveys is that the information that can be obtained most ac-

curately is current information on risk factors or disease conditions. It is difficult to