BIOSTATISTICS: The Perfect Deception.
Biostatistics: The Perfect Deception
Statistics is the science of analysis and interpretation of data.
Biostatistics on the other hand is the application of statistics to biology and, most commonly to medicine. Because research questions in biology and medicine are various, biostatistics has expanded its domain to include any quantitative, not just statistical models that may be used to answer such questions. Sampling and door to door surveys is yet another biostatistics tool.
While hard-core statistical decisions affect only the profit and loss of a certain company, or maybe the very existence of a production house, the impact of decisions in biostatistics is much more severe. When it comes to Biostatistics, one is no longer dealing with concrete, inanimate objects but with actual flesh and blood. In Biostatistics, either you take the right decision or PEOPLE DIE.
An editor of a leading daily, after having claimed that “his newspaper was the most widely read in the city” was asked to define and explain the ways and methods of Statistics. He replied nonchalantly, “Torture numbers, and they'll confess to anything.”
In the case of biostatistics, I can say, “Torture people with questions (surveys) and they’ll say just about anything to escape it all.” Interpreting such data and basing decisions on the same isn’t really a good idea.
A huge problem with biostatistics at the human being level is of the human psyche. Ask 100 people 10 questions and repeat this experiment after 10 days. The probability of getting the same answers can be observed and calculated, I agree. But if this very questionnaire is repeated yet again after 10 days, the probability of getting the same answers cannot be predicted even if we are ready to accept a 99% error margin.
This is largely due to fact that the human brain is a very dynamic entity. Ideas are flashed and then vanish in nanoseconds. The train of thought runs at variable speed in different brains. What is now, may not be in the next minute and may absolutely disappear in the next hour.
Collecting data from such an active and ever changing “data generator” is not difficult but interpreting and then basing decisions on the following data doesn’t really make any sense. For what conclusion is valid today might just look very absurd tomorrow. What opinion people have in their minds is often not revealed as well. And it can be concluded without hesitation, when Biostatistics meets the human psyche, it fails to perform. And it is this non-performance of biostatistics, at this very crucial level that questions the very concept of ‘human biostatistics’.
One of the first things taught in elementary biology classes is an ordinary looking sentence, “Biology is a science of exceptions”. But no matter how insignificant this statement may seem, it is actually a Law in itself. And going by this, it can well be concluded that the results of any biological sampling will vary to such an extent that no matter how much error one is willing to accept, data interpretation will still not be accurate.
What can also be deciphered from this statement is that no hypothesis, no generalization can actually be given in biostatistics with very narrow confidence limits. In order to be able to give a generalization, the confidence limits will have to be so broad that the hypothesis will seem like just another ‘comment’.
A very statistically amusing concept is used in assessing a patient with burnt injuries. In what is know as the “rule of 9”, it is used to estimate the surface area of the body affected with the burns.
9% for the head and each upper limb; 9% for the front of each lower limb; 9% for the back of each lower limb; 9% for the front of chest; 9% for the back of chest; 9% for the front of the abdomen; 9% for the back of the abdomen, is the 99% of the body. The remaining one percent is for the external genital.
This has been deduced after some careful statistical studies of burnt patients who died and patients who survived the impact of fire. And its concluded that 50% or above is fatal and the patient with this condition cannot survive.
Now for the amusing part. There can actually be a case when the patient has suffered just about 9% injuries and still died! Or has suffered 63% injuries but survived. The “rule of 9” thus only proves to be just another ‘comment’ even when this is followed in all major Forensic Labs throughout the world.
According to Biostatistics therefore, if you were standing with one foot in the oven and the other in an ice bucket, then you should be perfectly comfortable!
Luck, Co-incidence, Chance and Prediction are words familiar to every Statistician. In Biostatistician, there important only increases!
On a regular Wednesday afternoon, two young women were having their lunch in the verandah of their second storied house. Below them was the town street, hustling and bustling with people. The afternoon was getting monotonous when the younger of the two women leaned towards the street and said to the other, “I bet you $50 that in the next one hour, more men will pass through the road from under our balcony then women.”
This came as a surprise to the other lady since both of them knew their town had a sex ratio in favour of the fairer sex. And she ‘statistically’ agreed to the wager and began planning in her mind, of all the things she’ll purchase with those $50.
But after an hour of counting, it was the first women who had proved victorious!
No, the statistical sex ratio wasn’t incorrect. It was just that there was a worker’s rally that was to pass from that area and while the first lady knew about it, the second did not!
So much for biostatistics.
