Why medical students should understand statistics

1
Why medical students should understand statistics

• Mike Campbell
• Professor of Medical Statistics
• School of Health and Related Research
• University of Sheffield
• m.j.campbell_at_sheffield.ac.uk

ScHARR
2
Medicine and Society Phase 1

• Medical statistics (P9)
• Aim To take a population approach to effective
  practice
• Academic competence
• Introduction (P36)
• Ideas about variability

3
Objectives of lecture

• Not to put students off statistics!
• Review reasons why medical students should study
  statistics.

4
Review of lecture

• Examples of how the population approach has
  helped different areas of medicine
• Bristol Royal Infirmary Inquiry
• Harold Shipman case
• 3rd generation oral contraception scare

5
Why should medical students understand statistics?

• It will help in other courses
• It will help in the interpretation of medical
  papers
• Ultimately, patients will benefit

6
Why Medical Statistics?

• Medicine is a quantitative science but not exact
• Not like physics or chemistry
• Statistics is about handling and quantifying
  variation and uncertainty

7
Variation

• Variation characterises much of medicine
• Humans differ in response to exposure to adverse
  effects
• Example not every smoker dies of lung cancer
• some non-smokers die of lung cancer
• Humans differ in response to treatment
• Example penicillin does not cure all
  infections
• Humans differ in disease symptoms
• Example Sometimes cough and sometimes wheeze are
  presenting features for asthma

8
Probability

• Thus
• Diagnosis and treatment are probabilistically
  based
• We talk of the risks due to exposure, the chances
  of cure, the probability that a patient has a
  disease

9
Memorable quotes

• 50 of what you learn about therapy in the next 5
  years is wrong.
• (The trouble is we dont know which 50)
• (Anon)
• Nothing in life is certain except death and taxes
  (Woody Allen)
• 86 of all statistics are invented on the spot
  (Huff How to Lie with Statistics)

10
Changes to therapy in last 30 years

• Old therapies discredited
• Bed rest after operations
• Tourniquets to stop bleeding
• Lying new born babies prone
• Laetrile for cancer
• Bland diets for peptic ulcers
• Stretching before vigorous exercise
• Arnica for bruising

11
New therapies in last 10 years

• Inhaled anti-inflammatory drugs for asthma
• Folate for pregnant mothers
• Anti-coagulants before surgery
• Knee replacements
• Statins for heart disease
• MMR vaccine

12
Clinical Trials

• Statisticians have been responsible for the
  development of the randomised controlled trial by
  which all new therapies should be evaluated
• Therapies should be evidence based

13
Usefulness of statistics

• People in large groups are very predictable
• 600,000 people will die next year in UK
• At least 110 of the medical students will have
  had asthma as a child
• It is likely that someone will win lottery this
  week
• For an average person, the chances of dying
  before the draw are higher than the chance of
  winning, if the ticket is bought any time up
  until 20 minutes before the draw.

14
Usefulness of statistics (2)

• Statistics can tell us whether events could have
  happened by chance
• Statistics can tell us the likelihood of rare
  events
• Statistics can tell us what we should expect

15
Recent examples from the news

• Recently a surgeon had a mortality of 60 for a
  heart operation in children less than one year
  old. Nationally the mortality rate for this
  operation in this age group was 16/123 13
• Would you sack him?
• No!
• He may be unlucky.
• He may be operating on more severe cases
• He may only have done a few operations.
• For example 2/3 is 66.

16
Surgeon Ctd

• In fact no evidence babies were more ill.
• The surgeon operated on 15 babies of whom 9 died.
• A lower limit for the 95 confidence interval (to
  be explained later) for this rate is 36.
• An upper limit for the 95 confidence interval
  for the national rate is 19.
• Would you now sack him?
• Yes!

17
General practice example

• A certain GP certified 39 deaths in 1995- should
  we be worried?
• No!
• Need to know how many deaths to expect.
• Solution look at contemporary records of
  comparable GPs

18
GP example 2

• Based on data from comparable GPs we would expect
  7.4 deaths
• Can use a Poisson distribution to show that 39
  deaths most unlikely to have arisen by chance

19
GP - Age sex distribution of deaths 1990-1998
Women Obs Expted Men Obs Expted
0-50 3 7.2 9 7
51-64 26 26.5 18 28.6
65-74 74 58 41 48
75-84 168 130 58 52
85 96 145 28 18
Plt0.001 P0.06
20
P-value

• Tells us whether the results could have arisen by
  chance
• Results for men possibly chance (6 in 100)
• Results for women only occur 1 in 1000

21
Distribution by time of day
22
Result

• Clear evidence of systematic murder

23
Results of Inquiries

• Surgeons will be statistically monitored
• GPs are likely to be monitored
• Both need to be aware of statistical variation

24
Definitions

• Relative Risk (RR)
• Ratio of probability of event for exposed person
  to probability of event for unexposed person
  (Pe/Pu)
• Absolute Risk Reduction (ARR)
• Difference in probability of event for exposed
  person and unexposed person
• (Pe-Pu)

25
Contraceptive Pill Scare (October 1995)

• Committee on Safety of Medicine issued warning of
  association between 3rd generation contraceptive
  pill and deep vein thrombosis
• Relative Risk of 2 widely reported
• Women advised to discuss risks with GP at time of
  next prescription
• Story leaked to the press before doctors could be
  properly informed

26
Contraceptive Pill Scare

• CSM warning based on the following results
• 30 cases DVT per 100,000 users per year 3rd gen.
  pill
• 15 cases per 100,000 users per year 2nd gen. pill
• Thus Relative Risk 3rd gen. to 2nd gen. 2
• However Absolute risk reduction 15 per 100,000
  women per year
• Thus number needed to get one extra DVT on 3rd
  gen. 6700 women years.
• Note also
• 5 cases per 100,000 per year non-users
• 80 cases per 100,000 per year for pregnant women
• Mortality from DVT 1-2

27
Contraceptive Pill Scare (October 1995)

• Advice to consult GP misunderstood
• Many women stopped taking the pill immediately or
  did not finish course
• Increase in unwanted pregnancies
• Increase in abortion

28
Legal abortions performed per quarter, 1991-1996
29
Conclusions

• Due to a misunderstanding of risk, many women
  came off pill
• Many subsequently got pregnant and had an
  abortion
• Risks due to pregnancy and abortion far exceed
  those due to 3rd generation contraceptive pill

30
Summary of lecture

• You should know
• Why Medical Statistics is important
• You should know about
• Relative risks, absolute risk reduction

31
Review of rest of course

• Later this term
• Logic behind statistical inference sampling,
  confidence intervals and p-values.
• Diagnostic tests sensitivity and specificity.

32
Books

• Campbell MJ and Machin D (1999) Medical
  Statistics A Commonsense Approach. 3rd Ed. Wiley
• Bland M (2000) An Introduction to Medical
  Statistics 3rd Ed. Oxford Medical
• Coggon D(1995) Statistics in Clinical Practice
  BMJ.
• Swinscow TDV and Campbell MJ(2002) Statistics at
  Square One 10th Ed. BMJ
• http//www.bmj.com/