Bias in clinical research.
"The most difficult subjects can be explained to the most slow-witted man if he has not formed any idea of them already; but the simplest thing cannot be made clear to the most intelligent man if he is firmly persuaded that he knows already, without a shadow of doubt, what is laid before him." - Leo Tolstoy
10/31/2009 TOUFIQ. SHAIKH 1
• What is validity? • What are chance, bias & confounding? • Various types of biases. • Methods to prevent bias. • Disadvantages of bias. • Blinding & randomization. • Conclusion.
10/31/2009 TOUFIQ. SHAIKH 2
What is validity?
• Validity is the truthfulness of a clinical study.
Internal validity:External validity:-
• A study must be internally and externally valid.
What are chance, bias & confounding.
• Chance:- It is a random variation which may occur while taking measurements. • Bias:- It is a systemic error which leads to deviation of results from the true values. Bias may occur in the way we select our patients, measure our outcomes or analyze our data which will lead to results that are inaccurate. • Confounding:- It is the distortion of the effect of one risk factor by the presence of another.
Types of biases.
Group A:1) Selection bias. 2) Measurement bias. 3) Analysis bias.
Group B:1) Publication bias. 2) Funding bias. 3) Reader bias.
Group C:1) Patients bias. 2) Investigators bias. 3) Statisticians bias.
• Systemic error in the selection & allocation of subjects into different groups in a study due to which the outcome gets biased. • Also, if the chosen sample of population are not representatives of the population at risk.
• It can be prevented and reduced with the help of randomization techniques.
• Systemic error which occurs due to use of different instruments or when different observers of instruments are applied to the study groups. • Also, due to inaccuracy of measurement instruments ( such as diagnostic tests, etc.). • It can be prevented by use of the same instrument throughout the study.
• Participants who change groups, withdraw from the study or are lost to follow up may be systemically different from those who complete the study, thus leading to analysis bias. • Analysis bias can be reduced by maximizing follow up and carrying out an intention to treat analysis.
• It is a type of bias wherein the investigators, reviewers and editors publish only those studies showing positive outcomes and conceal if any negative ones from the public at large. • It can be prevented and reduced by registering clinical trials at the time of trial approval by the ethics committee:-
ICMJE:- International committee of medical journal editors.
Meta – Analysis.
• Meta- analysis is a set of statistical techniques for combining information from different studies to derive an overall estimate of a treatment effect. • For eg., the use of beta- blockers in sec. prevention after M.I is widely recommended, largely on the basis of a metaanalysis published in 1985. • Case study:- Meta– Analysis ---- Magnesium ---- ISIS 4 trial. • ISIS 4 – 4th International study of infarct survival.
• The bias in the design, outcome & reporting of industry sponsored research in order to show that a drug shows a favourable outcome. • Since, the investigator receives huge amounts of monetary compensation from the sponsor the results of a study by such an investigator could be favoured or biased.
• This type of bias can be prevented and reduced by development of vigilance organizations and guidelines such as CONSORT.
• CONSORT:- Consolidated standards of reporting trials.
10/31/2009 TOUFIQ. SHAIKH 11
• Systemic errors of interpretations made during inference by the reader of clinical data. • Such bias occurs due to prejudice, traditions & human nature.
• Being aware of one’s own prejudice while making an opinion or decision may help in reducing this bias.
Group – C.
• Patient’s bias:- Patient’s expectation may influence the observed effects of the drug if the patient is aware that he is receiving an experimental drug or a placebo. • Investigator’s bias:- Investigators may influence the selection of subjects, also they may be more vigorous/lenient in observing subjects receiving the experimental treatment. • Statistician’s bias:- Statisticians may ignore the level of significance and the effect of other biases under the pressure of the sponsor.
Disadvantages of bias.
1) Bias reduces the validity of a clinical research study. 2) Publication bias may adversely affect the health care decision making and also affect other studies based on the original biased study. 3) Drugs being approved as a result of biased clinical studies could cause adverse effect disasters in patients. 4) Treatment failures due to funding bias in clinical study could lead to loss of trust and faith of volunteers in the clinical research enterprise.
• Blinding is the corner stone in the evaluation of treatment procedures in medical science.
Blinding prevents placebo effect, patients bias, investigators bias & the statisticians bias.
Blinding techniques are of the following types;1) Single blinding:- Only patient. 2) Double blinding:- Both patient & investigator. 3) Triple blinding:- Patient, investigator & the statistician.
10/31/2009 TOUFIQ. SHAIKH 15
• Randomization is a process that assigns research participants by chance, assuming that each participant has an equal chance of being assigned to any group.
• Randomization techniques:1) Simple randomization. 2) Block randomization. 3) Stratified randomization. 4) Covariate adaptive randomization. • It prevents selection bias & is considered a gold standard in most clinical trials.
• Altruism & trust lie at the heart of clinical research, thus the CR enterprise has an obligation to conduct research ethically & scientifically without any bias. • Randomized, controlled, double-blinded clinical trials are the gold standard for proving any hypothesis.
• Meta- Analysis is a simple, easy & cost effective way of proving a hypothesis, its contamination by publication bias should be best avoided.