Medical Statistics: The Truth Behind The Numbers

1. A chocoholic friend is horrified by a newspaper headline: "A bar of chocolate a day triples your risk of migraine". What's the best advice you can give them?

A) swear off the brown stuff forever

B) cut their consumption by two-thirds

C) find out what their initial risk of migraine is and then decide what to do

D) find out how many migraine sufferers eat chocolate and then decide what to do

Answer: (c). Whether it's true or not, the headline portrays a relative risk. To assess your own personal response, you need to know how your absolute risk of getting a migraine in the first place, chocolate or no chocolate

Explanation

Answer: (c). Whether it's true or not, the headline portrays a relative risk. To assess your own personal response, you need to know how your absolute risk of getting a migraine in the first place, chocolate or no chocolate

2. You are researching your risk of getting heart disease. In the course of your researches you come across three different statements of your risk. 1. You have a 1 in 7 risk of getting heart disease. 2. Your risk of getting heart disease is about 14 per cent 3. 570 people suffered from heart disease out of 4000 people similar to you in a recent study. Without a calculator, which one of the following statements do you think is true?

(a) Statement 3 represents a greater risk that statements 1 and 2

(b) Statement 2 represents a greater risk that statements 1 and 3

(c) Statement 1 represents a greater risk that statements 2 and 3

(d) All three statements represent approximately the same risk

Answer: (d), all three statements are roughly equivalent. Easy, you might think, but studies show that the higher figures of statement 3 often induce a "ratio bias", increasing perception of risk

Explanation

Answer: (d), all three statements are roughly equivalent. Easy, you might think, but studies show that the higher figures of statement 3 often induce a "ratio bias", increasing perception of risk

3. Your newspaper reports that last July the number of cycling accidents also shot up when sales of sunglasses increased by 40 per cent. What is the most rational conclusion?

A) Cyclists wear sunglasses more than non-cyclists do

B) There are more cyclists in hot weather when sales of sunglasses increase

C) Wearing sunglasses while cycling is dangerous: you are less likely to see obstacles on the road

Answer: (b). In the words of medical statistics pioneer Austin Bradford Hill "Is there any other way of explaining the set of facts before us, is there any other answer equally, or more, likely than cause and effect?” In this case, there is a more likely link between the two facts than sunglasses causing cycling accidents: sun causes both sunglasses and cycling

Explanation

Answer: (b). In the words of medical statistics pioneer Austin Bradford Hill "Is there any other way of explaining the set of facts before us, is there any other answer equally, or more, likely than cause and effect?” In this case, there is a more likely link between the two facts than sunglasses causing cycling accidents: sun causes both sunglasses and cycling

4. A study found that out of the 100 people diagnosed with navel cancer in country A in 1995, 20 were dead by 2000, 20 more were dead by 2005, and 20 more were dead by 2010. The 40 people that still alive were declared cured. A similar study in country B showed that of 100 people diagnosed in 1995, 60 were dead within 5 years. On the basis of these figures, which of the following statements can we make (click all that apply)?

A) You have a better chance of surviving navel cancer in country A

B) The 5-year survival rate in country B is half that in country A

C) The mortality in country B is higher than that in country A

D) The 15-year survival rate is higher in country B

Answer: Without further evidence, we can only be sure that (b) is true; five years after diagnosis, 80 people were still alive in country A against 40 in country B. But the low 5-year survival rate in country B could be a systematic effect caused by later diagnosis; without knowing what happened in country B after 2000, we can make no comparative statement about the ultimate likelihood of surviving navel cancer in the two countries

Explanation

Answer: Without further evidence, we can only be sure that (b) is true; five years after diagnosis, 80 people were still alive in country A against 40 in country B. But the low 5-year survival rate in country B could be a systematic effect caused by later diagnosis; without knowing what happened in country B after 2000, we can make no comparative statement about the ultimate likelihood of surviving navel cancer in the two countries

5. To test a drug expected to reduce blood pressure by 5mm Hg you need 130 people for a well-designed randomised controlled trial. Suppose you are working on a new drug that should reduce the blood pressure by double the amount, 10mm Hg. Do you need to test it on:

A) more people?

B) fewer people?

C) the same number of people?

D) the exact same people?

Answer: (b). If a bigger effect is expected, it should be come apparent with fewer people

Explanation

Answer: (b). If a bigger effect is expected, it should be come apparent with fewer people