Transcript When Presenting Risk

Evaluating Information
and Presenting Risk
Today’s Class
•
•
•
•
•
Fact Sheet Assignment Review
Evaluating Information
Presenting Risk
In-class Activity
This week’s homework
Fact Sheet Assignment
• Well done!
– Nice use of headers, bullets, etc to
organize content
– Good sources
– Appropriate use of plain language
Fact Sheets
• Be Mindful
– Avoid clinical language
• “sleep disturbances”
• “menstruation”
– Provide intro text before a bulleted list
• “Symptoms of iron deficiency anemia include the
following:”
– Make sure your bullets are consistent
• Formatting (bold, hyphens, etc.)
• Structure (all phrases or sentences, start with verb or
noun, etc.)
Evaluating Information
• Consider the source
• Determine if information is consistent
within the source and with other
sources?
• Consider the study design
HealthNewsReview.org
• When reporting about a treatment or test, be
sure to:
– Discuss cost
– Describe how big (or small) are the potential
benefits AND harms
– Use absolute (not just relative) risk/benefit data
– Compare the new idea with existing alternatives
– Seek out independent sources, with no conflicts of
interest
– Look beyond the news release
HealthNewsReview.org
(continued)
• When reporting about a treatment or test, be
sure to:
– Avoid disease mongering - exaggerating or
medicalizing conditions
– Remember (and remind your readers) that not all
studies are equal
– Note whether it’s really a new idea or just new
wrapping on an old idea
– Explain whether/how widespread it’s available (or
is it years away?)
Before we talk about risk, a
a bit about numbers
• Do the math, if you can
• Give a visual
– “size of a football field”
• Use comparable forms
– 3/4 (fraction) vs. 75% (percentage) vs. 75 (absolute)
• Use sparingly
– Only two sets of numbers per sentence
• Avoid excessive detail
• Report only statistically significant numbers (or
clarify)
Risk
• To the scientific community, risk is a
statistical concept of probability
– The likelihood that a given event will occur
under certain circumstances
• For public, it is infused with emotion
When Presenting Risk
• Clarify the outcome under consideration
– What are you talking about (getting disease vs.
dying)?
– Is the outcome a composite or intermediate
outcome?
– What is the time frame?
• Provide context
–
–
–
–
How dangerous (lethality)
Compare your risks to the average person
Compare this risk to other risks
Compare risks alongside benefits
When Presenting Risk
(continued)
• Consider framing risk in more than one
way
– 1 out of 100 had X
– 99 out of 100 did not have X
• State treatment alternatives
• Review measures for controlling risk
• Highligh study limitations
Use Absolute Risk
• Simplest way to represent results is by
providing absolute event rates for the various
exposure groups
– Number who experienced outcome/number in
group
– “The 10-year risk of death was 10% (10 in 100) in
patients receiving drug A and 20% (20 in 100) in
patients receiving placebo.”
• Give numerator and denominator if you can
• Translate percentages less than 1 to
frequencies (X in 1000) and use a constant
denominator
Interpreting Absolute Risk
• No interpretation required
– X out of 100 patients in the new drug group
got sick compared with Y out of 100
patients in the placebo group.
What about relative risk?
• RR indicates the likelihood of
developing a given condition.
• RR is the absolute risk in exposed /
absolute risk in unexposed
– Group A 10% (10 out of 100 got sick)
– Group B 20% (20 out of 100 got sick)
– (10/100) / (20/100) = 0.5 (50%)
Interpreting Relative Risk
• RR > 1.0
– exposure increases the chance of the outcome
• RR < 1.0
– exposure decreases the chance of the outcome
• RR = 1.0
– exposure is not related to the outcome
Interpreting Relative Risk
• Relative risk < 1 (RR reduction)
– 1- RR
• 1 - 0.6 = 0.40 x 100 = 40% lower
• 1 - .23 = 0.77 x 100 = 77% lower
• Relative risk > 1
– RR - 1
• 1.25 - 1 = 0.25 x 100 = or 25% higher
– For RR greater than 2, just use the RR with “times
the risk”
• RR=2.5 (2.5 times the risk)
Relative Risk Pitfall
• Well-described finding that RR appears more impressive
than corresponding absolute risk reduction
• Reporting relative risk reductions without specifying event
rates leads readers to overestimate the magnitude of
findings
Study
Group A
Event rate
Group B
Event rate
RR
Reduction
Clinical
Importance?
1
10%
20%
50%
Yes
2
0.0001%
0.0002%
50%
?
Large samples make small effects statistically significant
Increases or Decreases
Risk?
•
•
•
•
RR = 0.8
RR = 10.0
RR = 1.5
RR = 0.4
Increases or Decreases
Risk?
• RR=0.8 for heart disease following new
therapy vs. controls
– Patients taking new therapy have a __%
higher/lower risk for heart disease than controls
• RR=10.0 for lung cancer in smokers vs. nonsmokers
– Smokers have __ times the risk for lung cancer as
non-smokers
Increases or Decreases
Risk?
• RR=1.5 for cancer in patients taking
new drug versus those not taking it
• RR=0.4 for stroke in patients taking new
drug versus placebo
When to Use Relative Risk?
• RR is an efficient way to summarize two
numbers or make comparisons across
treatments or studies
– RR for death was 0.9 for drug X, 0.7 for
drug Y, and 0.3 for drug Z.
Unless absolute risks are given, RRs are
incomplete information about effect size
For All Risks
• Be clear about three things:
– Exactly what the outcome is (having a
heart attack)
– Over what time period the outcome
occurred (5 years)
– In whom (adults with diabetes)
Activity
• Read the abstract of the provided study
• Together complete the RCT study
worksheet
Homework
• Reading assignment
– Gastel ch. 7 (pgs. 99-111)
– “How do U.S. journalists cover treatments, tests,
and procedures.”
• Writing assignment
– Write a 250-word summary of a research study
– Summary is due by 5:30 pm, Feb. 24 by email
• No class next week, but homework still due