Experimental Studies!!!

Experimental Studies

• Need For Such Studies:
  

The observational studies are useful to study the events, which cannot be controlled by the researchers. For example, it will not be possible to make a group of people chew tobacco to study whether they develop oral cancer or not. This will be ethically incorrect. Experimental studies are extremely rewarding to assess the efficacy of an ethically permissible and plausible intervention. The following illustrates and consolidate this point.

· A pharmaceutical scientist develops a new compound and claims that it is good for curing oral cancer. But the drug cannot be given to patients of oral cancer prior to testing the claim in a scientific manner.

· Similarly, health planners and policy makers of the country may like to start a new education campaign for preventing people from indulging in high risk sexual behavior. But the policy makers would like to know how efficacious and effective that particular campaign would be, prior to committing a substantial amount of available funds for this campaign.

The common thread running through both the scenarios described above is a decision- making process, about use of a drug and use of education campaign. In the first case, it is the lives of people that are at risk and in the second case, it is people’s money and some person’s jobs that are stake. And to save both, everyone would like to have enough evidence in favour of the decision that they are going to make.



To be continued……………

B. Randomized Control TrailIn the above given example, the effect of drug in the first case and the education campaign in the second case, may be examined on an experimental basis on a specified group of people, in such a scientific and statistically valid and ethically correct manner that its results help the decision makers to be as confident as possible about the success of the drug and the education campaign. The answer to such riddles lies in experimental studies. The most ideal study design in such situation would be randomized control trail (RCT), the initial two steps in a RCT are forming a hypothesis and selecting the sample.
(i) Forming a Hypothesis. In experimental Epidemiology, a hypothesis (a supposition or conjecture put forward to account for known facts) is propounded, the hypothesis is to be proved right or wrong, with experimental epidemiological study.
(ii) Selecting the sample
Studies are generally carried out in a defined group of people (experimental population) and the results can be applied to a larger group (reference population). It is therefore imperative to select a sample population which is fairly reprehensive of the reference population.
C. The Two Groups.The one which receives the intervention is called the experimental group and the other which does not receive the intervention is called the control group. The important question which the researcher must address is how to assign individuals to either the experimental group or the control group in such a manner that the results of the study are not erroneous on account of selection (selection bias). This is achieved by randomization.

To Be Continued............

D. Randomization
It means that an individual is assigned to the experiment or control group in a random manner i.e. by chance. In other words, the researcher cannot exercise any kind of particularity or bias in assigning and individual to an experimental or control group. There are several methods of randomization. Random number table, sealed envelops systematic randomization, etc. Usually in placebo-controlled trails, individual patients are randomized to the treatment group or placebo group.
(i) Group Randomization. Sometimes, especially in the community set up, simple randomization may not work. Suppose that we want to know the effect of an educational campaign programme in school students. The students from a class are randomly assigned to treatment or control group. After the intervention, the student may mingle and exchange information about the activities of the intervention. In such condition, the response becomes contaminated. To avoid this, instead of individual randomization, group randomization is carried out. Thus students of one school are considered as treatment group and students of anther school are assigned to control group.
(ii) Similarly, clusters of population or villages are randomly assigned to treatment or control group, which is known as cluster randomization. To minimize potential biases, sometimes clusters are chosen in matched pairs and each of pair is assigned to intervention or control group.
E. Blinding In a controlled trail, as the name suggests, the interventions are controlled by researcher, therefore such research runs the potential risk of being knowingly or unknowingly manipulated by people involved in the research study. If the patient knows that he is in put in the control group and not in the treatment group, he may under-report the improvement. Similarly, if the researcher is biased towards a particular treatment modality, he may over- report the positive out comes. Even the data analyst may manipulate the data in the computer to give biased results .Such biases have been reported in many previous studies. The method by which these biases can be minimized is known as blinding or masking.
(i) Single Blind Trail The experimenter knows about the assignment of the subjects to the study or control group. However, the subject on whom the experiment or assessment is being carried out is unaware of his status, whether he belongs to the control or experimental group.
(ii) Double Blind Trail In these studies, both the researcher as well as the study subjects are unaware of the group assignment.
(iii) Triple Blind Trail In addition to research and subjects, persons involved in the data analysis are also blind to the allocation of the group assignment. After the completion of analysis, the codes are broken and the results are announced.
Blinding is of greatest importance when the outcome is being subjectively determined. In placebo- controlled trails, the placebo and the experimental drugs are made to look identical; hence the patient and the drug administrator cannot differentiate between the drug and the placebo. Blinding reduces biases considerably but we cannot assume that blinding eliminate biases altogether. Another important aspect of RCT is allocation concealment.
To Be Continued...........

F. Intention to treat.In prospective studies of relatively long duration, patients may drop out of the study prior to its completion, or are may lost to follow up due to migration, death, serious illness, etc. Sometimes the treatment modalities to which an individual is assigned at the beginning of the study may have to be changed due to progress of disease or personal choice of treatment by the subject. These subjects will continue to be included in the same group to which they were assigned at the beginning of the study, whether they completed the treatment or not. This reduces the possibility of biased loss to follow up. At the same time, it also requires to be demonstrated that patients lost to follow up were not different in their characteristics from those who remain to continue in the follow up This approach, where the analysis is carried out with the dropped/ lost to follow up subjects being considered as part of original assignment is called ‘intention to treat analysis’.

G. Ethical Issues Conduct of randomized control trails is being is increasingly being governed by ethical and moral codes of conduct. Questions are raised about the justification of withholding treatment to patients by putting them in the placebo group and thereby increasing the risk of detriment in illness. Such problems are circumvented by comparing the new treatment modality with the existing treatment modality and allowing the patients to make informed choices in opting for the new or interventional regime. Another solution to this problem is by using cross-over study designs. The same group of patents, who initially acted as controls, are put in the treatment group and vice versa. In other words, patients in a time bound manner, as their own control. Though this method has its own limitations, nevertheless it serves a better solution for the problem at hand.
H. Obtaining Results After intervention, the data is converted by the computer into spread sheets. Suitable statistical analysis is carried out to compare the outcome of the two treatments. Honesty, integrity and transparencies are the key to success in clinical trails. Vested interests like pharmaceutical companies and biased researchers have produced bogus results in the past, by manipulation. Such malpractices can be minimized by making it mandatory to have a study cleared by ethical committee and institutional research board. Similarly the researchers have to preserve the record and other record available to the scrutiny by a competent authority.

Keep it up Abuzar

Thank You Sir!