Transcript Risk in Health_Feb.3.2015

Module Summary
Kashika M. Sahay MPH, PhD (c)
Date: February 3, 2015
Today
 Go over infographics and discuss the process/challenges of





creating the graphic (get into your groups and finish up graphic
right away)
Housekeeping—Reminder Office hours today from 2-3:30
Genomic Sciences Building 1334; Tomorrow from 11:30-1
Summary of this module: How do we calculate risk from a health
perspective?
Evaluations of module
Genre analysis and overview of the final project
Thursday: Guest lecturer Jennifer Hill
 Due on Thursday: 3 potential topics and a paragraph for each topic
Presenting infographics
 Email graphic to [email protected] if electronic,
provide drawing if hand-drawn
 In 1-2 minutes:
 What strategies did you use to decide which health statistics you
chose? Which statistics did you end up using
 Challenges in presenting the information in the infographic?
 What was the story you wanted
 What principles of design did you use/reject in designing your
infographic?
Summary of this module
Learning objectives (from the syllabus)
 By the end of this course, you should be able to:
 Exhibit numerical literacy, specifically in modes and methods of
calculating risk.
 Critically evaluate and interpret presentations of risk in the popular
media.
 Write clearly and concisely in a variety of disciplines and for multiple
audiences (e.g. academic, professional, lay)
 How did we do this in this module?
 Personal assessment of risk
 Scientific assessment of (health) risk
 Reconciling personal
Personal Assessment of Risk
Personal Assessment of Risk
A Risk is …
 A DECISION you MAKE that has different OUTCOMES
 A Level at which you MAKE a DECISION about an
OUTCOME
 How we EVALUATE your own PERCEPTIONS about a
BEHAVIOR or OUTCOME
What is Risk Perception?
 One’s estimate of the
likelihood that something
bad/undesirable will
happen in a given time
frame
 One’s feeling or opinion of
the existence or size of a
risk
What is an Acceptable Risk?
 The degree to which
someone or some society is
able to tolerate the
existence of something that
poses a danger
How do we decide which risks are
acceptable?
Consideration
Rationale
Example
Is the risk assumed voluntarily?
Less acceptable if
involuntary
Secondhand smoke v.
smoking
Is it natural or man-made?
Less acceptable if
manmade
Chemical spill v. tsunami
Likely to be used as intended?
Less likely if likely to be
misused
Ex. Guns v. microwaves
Are alternatives available?
Less acceptable
Unsafe toy v. driving
through traffic
Catastrophic or common?
Less acceptable if
potential for catastrophe
Airplane crash v.
automobile crash
Fair or unfair?
Less acceptable if risk is
borne by a single group
People living near chemical
plant v. those living away
Is evidence in the scientific
community unanimous?
Less acceptable if
Health risks of smoking v.
contradictory information vaccines and autism
from responsible sources
The Socioecological model
 Our assessment of risk
depends on:
 Our knowledge about the
risks of a behavior
 Our affective response
 Anxiety about a behavior
 Satisfaction with the outcome
 Certainty of our conclusions
 Assessment of the
information provided
 Decision making process
 Compliance with
recommendations
Scientific Assessment of Health
Risk
Epidemiology
 What is the distribution of




disease in a population?
What is the distribution of
exposure in a population?
Over what time frame?
Are the results from your
study generalizable to the
overall population of
interest?
How certain are you of
your estimates?
Some study designs
 E=exposure
 DZ=disease
Note that the timing of data
collection, timing of
exposure, and timing of
disease are factors that differ
across study design
How do you decide which study design
is best?
Risk as a mathematical quantity
 The probability that a health event will occur (usually
comparing probabilities in two groups)
Odds
Odds 
Probability of event occurring
Probability of event not occurring
Important point: What’s the difference between a proportion and an odd?
Proportions involve comparing parts to a whole and can only be
between 0 to 1.
Proportion: Orange squares/all squares= 3/5=0.6
Odds can range from positive infinity to negative infinity.
Odds: orange squares/brown squares= 3/2 or 1.5/1 ; The odds of
picking an orange square is 1.5 times the odds of picking a brown
square
Reconciling personal experiences
with scientific data
Some mistakes to avoid and tips to keep in mind;
An example of odds

One year Odds:
 Numerator= # of people in America who were alive
 Denominator= # people who died from that disease
Ex. Drowning : 275306000 people alive /3248 died of drowning
79,065 people alive/1 died of drowning
Odds of dying of drowning 1 to 79065 in the year 2000 in the US population
overall
Correlation does not equal causation
 Does ice cream cause murder?
 Based on this graph, the number of murders increases with the # of ice creams
sold. Does that mean that ice creams cause murder?
 No. Always remember, correlation (sometimes called association)
DOES NOT EQUAL Causation.
How did we determine that
concussions lead to brain injury?
Q: How do you establish causation?
A: Bradford Hill Criteria
 Temporality: Exposure comes






before the disease.
Plausibility: biological pathway
from exposure to disease is
established
Consistency: results are
reproducible in scientific study
Dose-response relationship: if you
increase exposure incrementally,
disease also increases
Specificity: single cause leads
directly to an effect
Reversibility: in the absence of
exposure, disease does not occur
Coherance: compatible with
existing knowledge
What can we do to better process
information about risk?
How do science and perception come together?
Great, you did this valid study…How sure
are you that you found the “right” answer?
Bias
Bias in scientific study
Bias in our perception
 “All of us show bias when
Ask yourself:
Are the study participants generalizable to
the overall population?
Remember, we rarely know the true
population value we are trying to estimate.
it comes to what
information we take in. We
typically focus on anything
that agrees with the
outcome we want.”
Noreena Hertz,
professor of Decision
Sciences
Numbers and Information can be
presented in misleading ways
 Even though we are
skeptical, we should always
critically evaluate
information
 Journalists, scientists,
others, may not be
intentionally misleading us,
they all have their own
reasons for presenting
information
What can we do to better process
information about risk?
 Evaluate:
 Emotion: Is this statement trying to elicit an emotional response
from me? Is the framing of the statement in terms of gain or loss?
 Numeracy: Are numbers presented as proportions, rates, or
odds?
 Source: What is the source of this information? Is it credible?
 Absolute v. Relative Risk: If totals are presented, are they
related to any other quantity?
Guidelines for Evaluating Risk
 Context Matters
 Ex. If you want to tell me how many chickens were killed because
they were suspected of avian flu, tell me what proportion of chickens
in all of England were killed
 Practice good “number hygiene” Consider your audience’s
numeracy and computational skills
 Make big numbers and small numbers relatable
 Avoid nonsensical comparisons
 Be wary of surveys and who collects data
 Ex. A survey on Taylor Swift’s fan page may not be nationally
representative
 Consider conflicts of interest
 A skin care company may have a vested interest in showing you it
reduces acne
Guidelines for Reporting Numbers
 Mean or Median which to report
 Mean- when you believe your data is normally distributed (evenly spaced
throughout the population but centered around an average value, ex. Men in the
United States have an average height of 6 feet)
 Median-when your data may be skewed to have values at extreme ends of the
spectrum (ex. Income, most of us are not billionaires, but
 Uncertainty: Confidence intervals should be renamed uncertainty intervals
 Note that we never know the true population value; We are trying to
approximate it
 P-values—debates with the values, statistical significance
 Statistical significance does not mean actual significance. (Just because there are
statistical difference between groups, does not mean that that statistical
difference has any practical value.
 Trends
 Be wary of trends; Remember that you can always cherry pick information over
the last 5, 10, 15, 100 years to prove a point.
Guidelines for Risk Perception
 Random v. likely events (winning the lottery v. freeze
warning)
 Realize that we tend to think of common events as less risky
(see section on risk acceptance)
 Relative Risk v. Absolute Risk
 Be wary of people telling you that a drug reduces your chances
of death by x% (a relative risk); Always ask what your absolute
risk of death is
 Ex) Remember the class example where the risk of death from Miracle
drug is reduced by 20%. However
 Another example of this:Your risk of blood clots when using birth control
(the pill) is doubled. However your absolute risk of blood clots goes from
2 in 10,000 to 4 in 10,000.
I hope this module inspires more
questions than answers!
Some questions to
think about..
 How do we store numerical information in
our minds?
 How do we quantify and avoid bias in our
perception and in scientific study?
 How can you design a study to be
generalizable to an entire population?
 Why are some studies believed more readily
than others?
Thank YOU!!
 Questions/comments
 Please complete
evaluations!
GENRE ANALYSIS
Rubric for Analysis
 Identify the communication purpose(s)
 Identify the audiences (internal, external, peers, up or down the
hierarchy.)
 Identify the organizational pattern
 What opening information is there – headers, titles, communication
chains, dates, etc.
 Organization
 What purpose do the sections of the document serve for the
audience?
 What is the purpose of the conclusion (to summarize, provide
information, provide highlights, to give contact information)?
Rubric for Analysis Continued
 Final questions
 What are the style expectations (formal, consultative,
informal)?
 What is the medium for delivery?
 Where does there seem to be flexibility in the genre
expectations?
Your assignment for Thursday
 In order to begin thinking about directions for your
introduction, theses, and arguments you might use, you will
generate a list of at least three possible topics for your
project.
 Each topic should be accompanied by a paragraph that explains
your interest in the possible topic, suggests the places in
science, politics, and culture you might look to for source
material, and point to possible arguments you might find
useful in the larger project.
Journal Prompt #4
 Considering that analyzing numbers and statistics is a
complicated process, think about how you would design a
study to learn more about [a disease of your choice]. What
kind of study design would you use? Who would be in your
population? How would you collect information and evaluate
risk of the disease? What biases might your study still have?
How would you present your study findings to general
audiences in a cohesive way?