SCIENTIFIC

THINKING
By May Bakr
16th of November
 MAJOR ASSIGNMENT 2
Intervention group
took less time to fall Intervention group slept
asleep for longer duration

Intervention group had

lower PSQI score, which
Thirty-eight participants were randomized to either an reflects better sleep
intervention group (n = 19), where members were instructed quality
to avoid using their mobile phone 30 minutes before bedtime,
or a control group (n = 19), where the participants were given
no such instructions.

[Link]
QUANTITATIVE REASONING
D A TA A N A LY S I S , V I S U A L I Z AT I O N , &
M A N I P U L AT I O N
 D ATA A N A LY S I S

• Data in mathematical and scientific speak is a group of information collected

often used to prove or disprove a hypothesis.
• Once data is collected, the next step is to get insights from it.
• Researchers rely heavily on data because they have a story to tell or a
problem to solve.
 D ATA A N A LY S I S

• The more we enrich our data with

meaning and context, the more
knowledge and insights we get out of it.
• At the top of the pyramid, we have turned
the knowledge and insights into a learning
experience that guides our actions.
 T Y P E S O F D ATA
• Two different data groups: Categorical and numerical/Quantitative data.
• Numerical data: information of something that is measurable, always collected
in number form (Ex: A person’s height, weight, IQ or blood pressure).
Numerical data is very popular among researchers due to its compatibility with
most statistical techniques. It helps ease the research process.
• Categorical data: data which can be divided into groups (Ex: color, gender,
race).
• Ask yourself: Can I average it?
Yes – Quantitative
No – Categorical
 T Y P E S O F D ATA
• Ask yourself: Can I average it?
Yes – Quantitative Zip codes
No – Categorical Calendar Months
Phone numbers
Locker combos
ID numbers

This is Categorical !
D ATA V I S UA L I Z AT I O N
H O W TO P R E S E N T I N F O R M AT I O N

• Provides a quick and effective

way to communicate information
in a universal manner.

• Your brain prefers visual

information!
 WHO CAN
ARGUE WITH A Graphs can help us
GRAPH? understand complex data,
but they can also mislead
and outright manipulate.
MISLEADING GRAPHS
AND THINGS TO LOOK
FOR
1 . V I E W C O R R E L AT I O N S W I T H S K E P T I C I S M

• When two variables X and Y are correlated:

A. X causes Y
B. Y causes X
C. A 3rd variable, Z, affects both X and Y
D. X and Y are completely unrelated
1 . V I E W C O R R E L AT I O N S W I T H S K E P T I C I S M
1 . V I E W C O R R E L AT I O N S W I T H S K E P T I C I S M
1 . V I E W C O R R E L AT I O N S W I T H S K E P T I C I S M

• What Does It Mean To Say, "Correlation Does Not Imply Causation"?

a) Two variables can only be strongly correlated if there existed a cause-

and-effect relationship between the variables.
b) The fact that two variables are strongly correlated does not imply a
cause-effect relationship between the variables.
c) The fact that two variables are strongly correlated implies a cause-
and-effect relationship between the variables.
d) Two variables that have a cause-and-effect relationship are never
correlated.
1 . V I E W C O R R E L AT I O N S W I T H S K E P T I C I S M

• What Does It Mean To Say, "Correlation Does Not Imply Causation"?

a) Two variables can only be strongly correlated if there existed a cause-

and-effect relationship between the variables.
b) The fact that two variables are strongly correlated does not imply a
cause-effect relationship between the variables.
c) The fact that two variables are strongly correlated implies a cause-
and-effect relationship between the variables.
d) Two variables that have a cause-and-effect relationship are never
correlated.
2 . R E L AT I O N S H I P S D O N ’ T L A S T F O R E V E R

How much money do you need to be happy?

3 . A LWAY S C H E C K T H E A X I S O F T H E G R A P H
3 . A LWAY S C H E C K T H E A X I S O F T H E G R A P H

Bar charts must have zero baseline.

4 . S TAT I S T I C S O N S M A L L S A M P L E S
A certain town is served by two hospitals. In the larger hospital, about 45
babies are born each day, and in the smaller hospital, about 15 babies
are born each day. As you know, about 50% of all babies are boys.
However, the exact percentage varies from day to day. Sometimes it may
be higher than 50%, sometimes lower.

For a period of 1 year, each hospital recorded the days on which more
than 60% of the babies born were boys. Which hospital do you think
recorded more such days?
1. The larger hospital
2. The smaller hospital
3. About the same (that is, within 5% of each other)
4 . S TAT I S T I C S O N S M A L L S A M P L E S

With smaller sample sizes, we get more extreme results!

 5 . L O O K AT A L L T H E N U M B E R S T H AT D E S C R I B E
A D ATA S E T
6. LOOK FOR BIAS IN SAMPLE SELECTION

Diverse viewpoints will give you a better overall picture of events.

We can’t always get to the complete truth of a matter, but we can at least see it
from multiple sides.
 • Research and statistics can be
TH E deceptive and used to
manipulate decision making.
MI S US E O F • Consider the facts behind a

S C I EN C E scientific statement more

critically.
THE MISUSE OF SCIENCE – TOBACCO CONSUMPTION

• The Tobacco Institute and Council for Tobacco Research in the

1950s released data which showed that only 1 in 100 smokers
develop lung cancer and concluded that “ Cigarettes do NOT cause
lung cancer ”.
• This data was responsible for nicotine addiction and unnecessary
cancer deaths over the past century.
• What percentage of people with lung cancer were once smokers?
90%
• What percentage of smokers die from smoking related illness? 67%
THE MISUSE OF SCIENCE – CANCER AND HYGIENE

“A UCLA (University of California at Los Angeles) study was

reported in October 2011 about the role bacteria play in
causing pancreatic cancer, a dreaded disease in which only 5
percent of people survive after five years. The report shows
that Streptococcus mitis, a type of oral bacteria, decline in
numbers in the oral flora of people with pancreatic cancer.
Streptococcus is a genus of bacteria that contributes to tooth
decay and looks like chains of circles under a microscope.
The report states that tooth decay is thus linked to pancreatic
cancer and concludes that “good hygiene can avoid the
terrors of getting the disease.”
THE MISUSE OF SCIENCE – CANCER AND HYGIENE

• People with pancreatic cancer have fewer S. mitis cells in their mouths, so
effective tooth brushing would have the opposite effect—increasing
pancreatic cancer susceptibility. So why does the report make such a
claim? It is probably because such a finding sounds good.
• By making people feel empowered because they could somehow avoid
pancreatic cancer by brushing, the report is positive, interesting, “headline
grabbing,” and makes sense to the public. However, it is an example of bad
science in every way.
A N A N A LY S I S O F D ATA M A N I P U L AT I O N
& ETHICS IN PHARMACEUTICAL
ADVERTISING: THE VIOXX CASE STUDY
Q1) What is the advantage of Vioxx, in
terms of side effects, over the naproxen
Leslie was previously taking?
Q2) Define and describe the purpose of a
placebo and a control group in an
experimental design.
Q3) What types of medical conditions are
included as a cardiovascular
thromboembolic adverse (CTA) event?
Q4) According to the table of data, how
does the incidence of CTA events with
Vioxx compare to the other drugs and the
placebo?
Q5) What do you know about the source of
these data?

Q6) Based on these data, what should Jeff

tell his wife when he goes home?
Q1) What information is on the y-axis? Express
this in your own words, being sure to define
“Cumulative Incidence.”
Q2) What information is on the x-axis? Express
this in your own words.
Q3) What do you think is a “gastrointestinal
event?”
Q4) Suggest why these gastrointestinal event
data would make Dr. Sara pleased with Vioxx.
Q1) Reexamine the table of data on the Cardiovascular Card. Can you determine the incidence of myocardial infarctions (heart attacks) with Vioxx or
any of the treatments? Explain your answer.

Q2) The Cardiovascular Card was printed several months before the VIGOR study was published but after the study was complete. What are some
possible explanations for why Merck did not directly refer to the VIGOR study on the card?

Q3) In the VIGOR study, with a total of about 8000 participants, 0.1% (4 out of 4 ~ 4000 subjects) on naproxen experienced myocardial infarctions
compared to 0.4% (17 out of ~ 4000 subjects) on rofecoxib or Vioxx. Explain whether or not you think this is a meaningful difference.

Q4) It has been estimated that 80 million people worldwide took Vioxx before it was removed from the market. Predict how many of these Vioxx patients
could have experienced myocardial infarctions. Does this calculation change your interpretation of whether or not this is a meaningful difference?

Q5) One piece of information the VIGOR study does not provide is the heart attack rate of a control group in a population similar to the one that was in
this study (50+ years of age, not morbidly obese, no history of cancer, strokes or heart attacks within past 5, 2, or 1 years respectively, and no history of
gastrointestinal surgery or inflammatory bowel disease, to name a few). How does this make it challenging for Jeff to interpret the meaning of the
numbers?
Q1) Why do you think Jeff, as a “field personnel with
responsibility for Vioxx,” is prohibited from discussing the
results of the VIGOR study, the same results published in the
New England Journal of Medicine, or the new data provided
at the bottom of the memo?
Q2) How is the percentage of heart attacks cited in this
memo different from the published VIGOR study in the New
England Journal of Medicine that Dr. Sara showed him? Give
a possible explanation for the percent change in heart
attacks between the memo and the published paper.
Consider what Dr. Sara told Jeff .
Q3) Assuming 80 million people have taken Vioxx, calculate
how many more people would suffer from heart attacks
while on Vioxx with the percentage presented in the bulletin.
Do you think this number represents a real risk to the public?
Explain your answer.
Q4) What do you think of the data cited in the very last
sentence? How could this be possible, considering the rates
of nonfatal heart attacks? Is the risk of Vioxx that serious if
the rate of cardiovascular mortality is no greater while on
Vioxx than while on naproxen?
END OF CASE
FA ST
FORWARD TO
W H AT
A C T UA L LY
HAPPENED
 When defending its decision to continue
the study, the safety panel said it
couldn't tell if Vioxx was causing the
heart problems or if naproxen, acting
like low-dose aspirin, protected people
from them, making Vioxx just look risky
by comparison.

Merck Previously Knew of Vioxx Heart

Risks : A cardiologist testifies that drug
maker Merck had scientific evidence of
Vioxx's cardiac hazards as far back as 1999
-- well before it pulled the drug from market
in September 2004.

Five years after Vioxx's launch, Merck withdrew the

drug from the market. By that time, Merck had sold
billions of dollars of the drug worldwide. Research
later published in the medical
journal Lancet estimates that 88,000 Americans
had heart attacks from taking Vioxx, and 38,000 of
them died.
• “These reports reveal just how far a drug maker might go to market its product
and try to bury information that might hurt sales even when that information
directly affected the health and safety of the people taking their medicine,”

• “Revealing this kind of activity is very important in building pressure on the

Food and Drug Administration to regulate, not accommodate drug makers.”
“Data manipulation may result in distorted perception by shifting data
around, which could lead to billions of dollars in financial loss or even
potential loss of life, depending on the system in question, and the
type of data being altered.”
 “However, despite Merck's knowledge that rofecoxib might increase thrombus formation, none of the
intervention studies that constituted its new drug application to the Food and Drug Administration in
1998 were designed to evaluate cardiovascular risk. The nine studies were generally small, had short
treatment periods, enrolled patients at low risk of cardiovascular disease, and did not have a
standardized procedure to collect and adjudicate cardiovascular [Link], Merck seemingly
pooled data from these studies and others for analysis of cardiovascular risks, despite FDA concern,and
disseminated the results to promote the drug's cardiovascular safety to doctors in its “cardiovascular
card,”

Krumholz, H. M., Ross, J. S., Presler, A. H., & Egilman, D. S. (2007). What have we learnt from Vioxx?. BMJ
(Clinical research ed.), 334(7585), 120–123. [Link]
 PAT I E N T - T I M E O F E X P O S U R E
• Incidence rates are used in many epidemiological studies and statistical assessments of risk, which allows
researchers to reveal trends and communicate levels of risk. Studies can express their safety findings using incidence
rates based on 100 patient-years. (Events per 100 patient-years).
• For example, in a study with 100 patients, 25 of whom participated for 2 years, and the other 75 who participated
for 6 years, the denominator of the incidence rate as the total number of patient-years exposed is 500 patient-years
[(25 patients x 2 years) + (75 patients x 6 years) = 50 + 450 patient-years].
• If among those 100 patients, 5 occurrences of accidental bone fracture were observed, then the risk (proportion)
would have been 5/100 patients or 5%. However, expressing the risk in this way assumes that the risk during the
remaining 4 years for the 25 patients with 2-year participation would be no different from their first 2 years or from
the 6 years of participation of the other 75 patients.
• If instead we calculate the incidence rate, we see that it is 5 per 500 patient-years, or 1 per 100 patient-years. This
means that, on average, there was 1 occurrence of accidental bone fracture per 100 patients over the course of each
1 year of treatment.
 C U M U L AT I V E I N C I D E N C E

• Cumulative Incidence is calculated as the number of new events or cases of disease

divided by the total number of individuals in the population at risk for a specific time
interval.