Exposure and a Disease Using a Minimum of Resources Connection


Epidemiology can be defined as the study of the determinants and distribution of disease frequency in human beings in addition to the application of this study in controlling health problems (Gordis, 2009). The study of epidemiology incorporates several study designs that include case-control studies, observational studies, and experimental studies (Woodward, 2005). Each of these study designs symbolizes different ways of harvesting or collecting information. How they are implemented also depends on how a research question is framed.

The main objective of these epidemiological study designs is to investigate the connection between exposure and disease using a minimum of resources with precision and validity (Szklo & Nieto, 2007).


There are 6 commonly applied epidemiological study designs and these include experimental studies, observational studies, cohort studies, case-control studies, cross-sectional studies, and ecological studies (Rothman, 2002).

An experimental study or a trial is that study investigating the role of some agent in regards to prevention or treatment of a particular disease (Gordis, 2009). In the course of this study, the investigator appoints individuals to two or more teams that receive or do not receive the therapeutic or preventive agent (Szklo & Nieto, 2007). Since an experimental study is similar to a controlled lab investigation, the majority of epidemiologists tend to believe that this study produces more precise results as regards science than most observational studies (Woodward, 2005). Experimental studies are usually categorized as per their objectives whereby an investigator manipulates the team that is to receive the agent under this study.

An observational study is usually involved with the study of the effects concerning a variety of exposures as compared to experimental studies and incorporates treatments, possible causes of illness, and preventions (Rothman, 2002). It also gives information explaining causes of illness incidence and determinants of the progress of the illness, to manage disease by studying what can be done to not only prevent illness but also prolong life with illness and to predict the future health care needs of a particular population (Gordis, 2009).

Unlike in the case of experimental studies where an investigator manipulates the process, in an observational study, the investigator inactively observes as circumstances and situations are determined by nature.

A cohort is a group of individuals sharing similar characteristics or experiences and therefore cohort studies is a term used to describe and investigate following groups with similar characteristics (Woodward, 2005).

Cohort studies examine numerous health effects as caused by exposure. Subjects are also defined as per their exposure levels and follow-ups made for the occurrence of disease (Rothman, 2002). On the other hand, case-control studies have always been perceived as being an inferior form of a cohort study. Here, subjects are chosen whether they suffer from disease or not. This is distinct from cohort and experimental studies where the main comparison is disease incidence between the unexposed and exposed groups (Szklo & Nieto, 2007).

Cross-sectional studies determine the connection between illness and other variables of interest as per their existence in a defined population at a particular time (Gordis, 2009). As compared to case-control and cohort studies, cross-sectional studies population is chosen without any concern of status of exposure or disease. They are also mostly carried out for purposes of public health planning and etiologic research (Woodward, 2005). Lastly, ecological studies examine the rates of illness as related to a factor described on a population level (Rothman, 2002). They usually point out groups as determined by time, place, or both of these two factors.

In other words, an ecological study examines the connection between disease and exposure as per the population level rather than individual-level data (Szklo & Nieto, 2007).

Different Types of Biases

It has been observed that the different types of epidemiological study designs are subject to different types of biases. Bias can be defined as a systematic error resulting in an invalid estimate of the measure of association (Rothman, 2002). It can either mask a relationship where there is one or establish fake relation when there is none (Gordis, 2009). Biases are introduced by the investigator or individuals participating in the study and are found in the design as well as the conduct of a particular study.

There are two main types of biases and these are observation and selection bias. Selection bias is those results acquired from procedures used in selecting subjects into a study that lead to a result distinct from what was previously expected from the entire population targeted for the epidemiological study (Woodward, 2005). They are also likely to be observed in cohort or case-control studies since exposure and outcome have already taken place at the time of study selection. On the other hand, observation bias can be defined as that error resulting from data obtained as regards exposure or illness from the study groups (Rothman, 2002).

They also result in those participating in the study who are not appropriately classified as either being exposed to or not exposed to or diseased or not diseased (Szklo & Nieto, 2007).

How to Minimize the Entrance and Side Effects of these biases

Biases normally occur due to the method in which variables for the study are collected or measured, the manner in which participants to the study were chosen, and the attitude or preferences of an investigator assigned to the study (Gordis, 2009). To minimize the entrance and side effects of these biases, a few methods may be applied for instance randomization. A randomization is a powerful tool used in reducing bias and is applied in the comparison of a new therapeutic technique with a conventional one, where participants are assigned to different therapy groups (Rothman, 2002).

The other method that can be implemented is that of consecutive recruitment. It has been observed that non-consecutive recruitment usually introduces new characteristics that may tend not to be in proportion to those observed in the target population and this may interfere with the expected outcome (Woodward, 2005).

Blinding can also be used to minimize the entrance and side effects of these biases by making sure that data such as demographic data or disease data that may influence an investigator’s interpretation of the test or assessment of the results is not available (Rothman, 2002). By these applications, biases will be minimized and the overestimation of diagnostic accuracy greatly reduced.


Investigators normally use epidemiological study designs as a protocol for carrying out the study where they are allowed to translate the presumed hypothesis into one that is operational. These study designs are dependent on three factors in the description of a health condition, that is, the individual, time, and place.


Gordis, L. (2009, Dec). Epidemiology. 4th edition. Philadelphia: Elsevier Health Sciences.

Rothman, K. J. (2002). Biases in Study Design. In: Epidemiology: An Introduction. New York, NY: Oxford University Press. Pp. 94 – 112.

Szklo, M. & Nieto, F. J. (2007). Epidemiology: Beyond the Basics. Boston, MA: Jones & Bartlett Learning.

Woodward, M. (2005). Epidemiology: Study Design and Data Analysis. 2nd edition. Boca Raton, FL: CRC Press.