## Types of Studies

Case series is the most basic form of study. It looks at a cluster of people with a disease. Often it is the initial report of a new disease. It may ask questions and propose hypotheses, but it cannot be used to test them or pretend to provide any hard data on risk of exposure or consequence of therapy.

Cross-sectional studies, the next step, study the possibility of linking risk factors with disease states. They study a group of people at one particular point in time; therefore these studies only look at the prevalence of a disease against the backdrop of an exposure. Since they are temporally static (one group at one point in time), they shed no information on the relationship between a risk factor and a disease in terms of which came first. These studies are useful for developing hypotheses, including those raised by the initial case reports.

Retrospective studies, also called case-control studies, compare afflicted individuals ("cases") to those unafflicted ("controls"). These two groups are identified, interviewed, and data (especially exposure to known, or sometimes unknown, risk factors) is retrieved from the information the subjects provide. From this the odds ratio can be calculated, which, as stated above, approximates the relative risk (the relationship between disease and exposure). The advantages to this kind of study are that few subjects are needed, which makes these relatively quick and inexpensive to perform. This format lends itself well to rare diseases and those with a long latency period. The main disadvantage to this type of study is the reliance of the data on the subjects' memories. This can lead to a significant recall bias, whereby memory causes the results to deviate from the truth. Another disadvantage to these types of studies is the inability to calculate disease incidence from the data, because the diseased group and the healthy group are preselected.

Prospective studies, also known as cohort studies, begin with a group of individuals ("cohort"), who may, or may not, share a common exposure, but it characterizes members of the cohort based on different features (e.g., habits, geography, income) The study then follows them over time and evaluates their health as some are exposed to certain risks and others are not. A good example is The Framingham Heart Study. Advantages to prospective studies include being able to examine a multitude of exposures and diseases and the relative risk can be calculated from this type of study. It is important to have a method to calculate the relative risk because ultimately that is the goal: what is the cause/effect relationship between an exposure and a disease?, an intervention and a change in a patient's status? Also, obviously, recall information depends on the relationship between a risk factor and a disease in terms of which came first. These studies are useful for developing hypotheses, including those raised by the initial case reports. Also, obviously, recall bias is not an issue since the production of data paces its collection. The main disadvantage to prospective studies is the beginning cohort (the very premise of the study) is a group of otherwise healthy individuals who presumably have the same rates of exposures and disease as the general population. Therefore it is exceedingly difficult to study rare diseases or disorders, and a large number of subjects are needed to adequately perform the study. This then leads to two more drawbacks of cohort studies: they are more expensive and may take years to collect the necessary data. The gold standard of testing for the evaluation of a new intervention is the randomized, double-blinded, placebo-controlled clinical treatment trial, a type of prospective study. In these, subjects are arbitrarily assigned to one of two groups (randomization). One cohort then receives the treatment, while the other receives the standard of care or at least an "inert" treatment (placebo-controlled). Because neither the examiner nor the subject knows who receives which, the design is termed double-blinded. These four qualifiers reduce bias by the greatest margin.

A hypothesis, in its most basic form, is a statement that postulates a difference exists between two or more groups. For example, a new cholesterol-lowering drug X is compared to an old drug Y. The hypothesis for this comparison could state (blandly), "treatment cohort A has a lower total serum cholesterol than cohort B, because treatment cohort A was treated with drug X, while cohort B was treated with drug Y." This draws a distinction between cohorts A and B. Once a hypothesis is created, it must be tested and withstand scrutiny. Studies are conducted, but the hypothesis must still be statistically proven to be true to be accepted. To do this, statistics employs a special tool to test the validity (effectively the sensitivity and specificity) of a hypothesis. This tool is simply the opposite of the hypothesis. If the hypothesis claims a difference between two groups, the null hypothesis claims that no significant difference exists between the groups. Statistical analysis proceeds to either accept or reject the null hypothesis. By rejecting the null hypothesis (the antithesis of the hypothesis), the hypothesis is assumed to be valid, and a significant difference is assumed to exist between the groups. To the researcher the null hypothesis has but one function: rejection.

Probability refers to the number of times a result would recur, if an experiment were repeated indefinitely. Probability is quantified as the p value, and quantifies how strong a connection exists between an outcome and an intervention (hypothesis), while the null hypothesis states that a result and an intervention are random occurrences. If the p value is equal to or less than 0.05, it is likely that the null hypothesis is rejected, and the hypothesis is true. Two types of "errors" can be made when considering a hypothesis by making erroneous assumptions from data or if data conceals a bias (a force that deviates data from the true values). A type I error is committed, when the p value is equal or less than 0.05 and the null hypothesis is rejected, despite it being true. Because if the null hypothesis is true (drug X has no effect on serum cholesterol levels; it was random chance that cohort A has lower serum cholesterol levels), then it must be accepted, and therefore the hypothesis is false. It is an error to reject the null hypothesis under these circumstances. Therefore a p value less than 0.05 is considered "statistically significant", assuming no error is made. Likewise it is an error to accept the null hypothesis even if it is false (type II error). It requires reconsideration of the data or the study design to determine that the hypothesis is in fact true despite the null hypothesis being accepted.

It is important to remember that between the different kinds of studies, boxes, and validity tests, the subjects under consideration remain medicine and human disease. If some information does not match clinical experiences or training, then it likely warrants further investigation. Also just because a hypothesis is found to be true and a particular intervention is statistically significant, that does not convey clinical significance.

References

1. Fadem B. Behavioral science. Philadelphia: Harwal Publishing, 1994.

2. Feibusch K, Breaden R, Bader C et al. Prescription for the boards: USMLE step 2. Philadelphia: Lippincott-Raven Publishers, 1998.

3. Hill A. A short textbook of medical statistics. Philadelphia: JB Lippincott Company, 1977.

4. Mack S. Elementary statistics. New York: Henry Holt and Company, Inc., 1960.

5. Weinberg G, Schumaker J. Statistics, an intuitive approach. Belmont, CA: Wadsworth Publishing Company, Inc., 1962.

## Lower Your Cholesterol In Just 33 Days

Discover secrets, myths, truths, lies and strategies for dealing effectively with cholesterol, now and forever! Uncover techniques, remedies and alternative for lowering your cholesterol quickly and significantly in just ONE MONTH! Find insights into the screenings, meanings and numbers involved in lowering cholesterol and the implications, consideration it has for your lifestyle and future!

## Post a comment