Journals into stories

1
Journals into stories

• Maryn McKenna
• 16 February 2007

2
Communication across cultures

• Constantly I felt I was moving among two groups
  - comparable in intelligence, identical in race,
  not grossly different in social origin, earning
  about the same incomes - who had almost ceased to
  communicate at all, who in intellectual, moral
  and psychological climate had so little in common
  that ... one might have crossed an ocean.
• - C.P. Snow, Rede Lecture, Cambridge, 1959

3
Why so difficult? Have different qualities

• Science
• Quantitative
• Weights findings equally
• Disdains hype

• Journalism
• Qualitative
• Needs new developments
• Needs significance

4
And use different vocabularyScientific
language is precise

• The influenza branch at CDC has developed
  primers and probes that can be used in real-time
  reverse transcription polymerase chain reaction
  (RT-PCR) assays with respiratory specimens and
  viral cultures to presumptively identify the
  presence of influenza A/H5 (Asian lineage)
  viruses. This will be the first FDA cleared test
  of this kind. It will be sent and limited to LRN
  designated labs.

5
while media language seeks to be simple and
engaging

• New Test Quickly Spots Bird Flu in People
• Health Day, 3 Feb 2006
• U.S. officials announced Friday the approval of
  a new rapid lab test to detect bird flu in
  humans. The test works by detecting viral
  genetic material, which, in turn, is used to
  demonstrate the presence of bird-flu infection.

6
though media language can be sloppy.

• WashPost, 13 Jan 1988
• Airline Accident Rate Is Highest in 13 Years

• WashPost 14 Jan 1988
• CORRECTION
• The accident rate, expressed as the number of
  accidents per 100,000 departures, has been
  declining since 1983 although it rose slightly in
  1987 over 1986. The 31 airline accidents in 1987
  were the highest number in 13 years.

7
Science and news employ different written
structures.

• A scientific paper is a narrative It proceeds
  forward in time from the initial question to the
  research results.
• MMWRs are similar though more condensed.

• A media report may be a narrative (rarely) but is
  more commonly an inverted pyramid with the most
  important facts at the top or a donut that
  departs from/returns to the same point.

8
Each structure has a skeleton

• JOURNAL PAPER
• Introduction
• Methods
• Results
• Discussion

• NEWS ARTICLE
• Lede (key new fact)
• Additional facts
• Quotes
• Background/context
• What comes next

9
on which are hung different questions.

• Scientific report
• Why did you do it?
• How did you do it?
• What did you find?
• Why does this matter?

10
News report

• What is new about this? What is not new?
• What is the significance of this and why?
• Who will disagree?
• What evidence is this finding is based on?
• Why were you interested in this?
• Who funded this research?
• What related research exists?
• What happens next?

11
How do we distinguish worthwhile stories?

• 1. Type of study
• Case report
• Case series
• Retrospective
• Prospective observational
• Randomized and blinded

12
More characteristics of worthwhile stories

• 2. Size of study (statistical power)
• 3. Length of study
• 4. Study subjects (lab model, mice, people)
• 5. Real-world applicability (basic science or
  clinical usefulness?)
• 6. Prior studies (confirms/extends earlier
  results?)

13
Try not to write this story

• The protocol resulted in the death of 33 percent
  of the mice and the complete recovery of 33
  percent.
• The third mouse escaped.

14
Consider the publication

• Big 6 (JAMA, NEJM, BMJ, Lancet, Science, Nature)
  implies better peer review.
• BUT hardly infallible
• Lancet willing to be provocative (Iraqi deaths)
• Science fooled by South Korean stem-cell work
• Major journals refused Barry Marshalls original
  H. pylori studies overturning ulcer hypothesis
  (eventual Nobel)

15
The ultimate confounder

• Is
• MONEY

16
How this plays out in the real world of journalism

• Gets play
• Positive results
• Significance
• Context
• More than one voice
• More than one study
• Local relevance
• Names and faces

• Does not get play
• Null results
• Increments
• Single-source stories
• Numbers without interpretation
• Refusal to draw larger significance

17
given that we are launching stories into a data
storm.

• Pharma direct-to-consumer advertising
• Creates expectations of miracle cures
• 24-hour news cycle echo chamber
• Perceived amplifier effect causes viewers to
  tune out
• Rise of the blogosphere
• Much opinion, little evidence, wide readership

18
What saves us is that people want to read these
stories

• Shannon Brownlee, Washington Post
• Theres an unwritten rule in mass-market
  magazines and television
• If your ratings are low, run a medical story.
• If your ratings are really low, run two.