Basic Statistics

INITIAL STEPS IN DESIGNING A BIOMEDICAL STUDY 3

The first of these objectives typically leads to what are called experimental studies

and the second to observational studies. An experimental study is one where the

researcher can decide on the treatment for each person or animal, and then investigate

the effects of the assigned treatment. For example, in a clinical trial where one-half

of the patients are assigned a new treatment for high blood pressure, their outcome

could be compared with that of those patients assigned the current best treatment.

Here, the objective is to compare the new treatment for lowering high blood pressure

to the current treatment to see which treatment is more successful.

An observational study is an investigation in which the researcher has little or no

control over the treatment or events. The relationship between the causal or risk factors

and the outcome is studied without intervention by the researcher. For example, the

proportion of smokers who develop complications after reconstructive breast surgery

following mastectomy can be compared with the proportion of nonsmokers who

develop complications. The past history of smoking cannot be controlled by the

surgeon; thus the objective is to assess the effects of this potential risk factor in an

observational study.

1.2.2 Making a Conceptual Model of the Disease Process

The second step is to write or draw a conceptual model of the disease process under
study. If an experimental trial is being planned, first, the beneficial effects of the
proposed treatments should be postulated; second, the time until the effects become
noticeable should be estimated; and third, the duration of the effects should be esti-
mated. Possible side effects (undesirable effects) need to be anticipated, as well as
when they may occur. For a more complete discussion of both the critical aspects of
modeling the disease and the effects of the treatments that should be considered, see
Pocock [1983], Wooding [1994], or Piantadosi [2005].
If an observational study is undertaken to determine the effects of possible risk
factors on the occurrence of disease, a conceptual model of the disease process should
be made. A possible framework is provided in Figure 1.1. The purpose of this diagram
is to remind the investigator of the possible causal factors (often called risk factors)
that could affect the disease and when they are likely to occur. For example, some
causal events may occur genetically before birth, some at birth, and some later in life.
There may be multiple events. Some of these events may have one-time occurrences
(accident, major life event, exposure to a virus, etc.) and some may be chronic events
(smoking, diet, continual stress, exposure to a low-level toxic chemical, etc.).
The period between the occurrence of the event and the occurrence of the disease
is called the induction or incubation period. This period could be quite short or very
long. In general, the longer the induction period, the more difficult it is to determine
the effect of a risk factor on the occurrence of the disease. The period between the
occurrence of the disease and its detection is called the latent period. For a broken
leg, this period could be very short, but for some diseases, such as multiple sclerosis,
it could continue for years.
Note that to be a risk factor the exposure or causal event has to occur in the proper
time period. For example, if lung cancer was detected in a patient in October 2008