Did News Reporters Get It Right?

Translation of the 2002 Hormone Study Findings

Mary K. Canales, PhD, RN, Erica S. Breslau, PhD, MPH, David E. Nelson, MD, MPH,
Rachel R. Ballard-Barbash, MD, MPH

Background: The news media play a critical role in communicating health information to the public.
The unexpected findings in July 2002 about increased health risks associated with
hormone therapy provided an opportunity to examine the process of translating scientific
findings to reporters through communication intermediaries and appraise subsequent
reporting in newspapers in the United States.
Methods: Using qualitative research software, a qualitative analysis was conducted in 2006 to consider
four types of messages: (1) hormone therapy health risks outweighed benefits (balance);
(2) adverse hormone therapy health outcomes (health risk); (3) positive hormone therapy
health outcomes (benefit); and (4) risk level (magnitude). The print materials analyzed
included the original 2002 Journal of American Medical Association (JAMA) article and
editorial; JAMA and National Institutes of Health (NIH) press releases; the NIH press
conference transcript; and 198 articles about hormone therapy in 22 U.S. newspapers
published from July to September 2002.
Results: The major study finding that hormone therapy risks outweighed benefits was reported
consistently and accurately. Analyses of language and numbers on risk magnitude, and its
interpretation revealed some variability, both within the translation materials and news
stories. When risk numbers were included in newspaper stories, absolute risk was used
more often than relative risk.
Conclusions: Despite much criticism of journalists’ coverage of health issues, U.S. newspaper reporting
about hormone therapy in 2002 was generally consistent. Several translational and
communication strategies used with hormone therapy may be applicable to other efforts
that involve working with reporters on major health stories or events. An important process
oversight was the absence of hormone therapy communication efforts and guidance
directed specifically to medical practitioners.
(Am J Prev Med 2008;34(1):61– 68) © 2008 American Journal of Preventive Medicine

Introduction and poor decisions about allocating public and private

funds.5 Therefore, it is important for scientists, jour-

T
he news media, because of their agenda-setting
function,1 can have an important influence on
healthcare decision making.2 There are many
examples, unfortunately, where news reporting on
health or environmental issues, particularly concerning
risk, has been less than ideal, such as, mammography
recommendations in the United States3 and mad cow
disease in Great Britain.4 Sensationalizing, emphasizing
controversy, and other reporting practices can lead to
risk amplification, uncalled-for screening or treatment,
From the Mashantucket Pequot Tribal Nation, Cancer Education
Grants Department (Canales), Bemidji, Minnesota; Behavioral
Research Program (Breslau), Applied Research Program (Ballard-
Barbash), Division of Cancer Control and Population Sciences,
National Cancer Institute, Bethesda, Maryland; Health Communica-
tions Branch, Office on Smoking and Health, Centers for Disease
Control and Prevention (Nelson), Atlanta, Georgia
Address correspondence and reprint requests to: Mary K. Canales,
PhD, 215 Pine Grove St. SW, Bemidji MN 56601-6603. E-mail:
mkcanales@charter.net.
nals, government agencies, and other institutions to
convey scientific information accurately to journalists
when an important health news story or event is likely.
A “scientific bombshell” occurred in July 2002 with
the Journal of the American Medical Association (JAMA)
publication of the Women’s Health Initiative (WHI)
study. Researchers’ reported that hormone therapy
increased certain health risks for healthy postmeno-
pausal women, especially cardiovascular disease and
breast cancer,6 contrary to two decades of observational
research suggesting beneficial effects, especially for
cardiovascular disease prevention.7 As a result, the
National Institutes of Health (NIH) halted the hor-
mone therapy intervention trial several years early.
At the time of these events, an estimated 14 million
postmenopausal U.S. women were using hormone ther-
apy.8 Not surprisingly, the JAMA study and the ending
of the WHI trial resulted in substantial news cover-

Am J Prev Med 2008;34(1) 0749-3797/08/$–see front matter 61

© 2008 American Journal of Preventive Medicine • Published by Elsevier Inc. doi:10.1016/j.amepre.2007.09.023
age.9 –13 Subsequent research has shown that the exten- Table 2. Newspapers, circulation, and number of hormone
sive hormone therapy media coverage was associated therapy–related articles22
with changes in attitudes and decreased hormone No. of
therapy use.10 –11,13–16 hormone
To the authors’ knowledge, there has been little Daily therapy–related
research on the entire translation process by which Newspaper circulation22,a articles
findings from a scientific study are conveyed to report- USA Today (VA) 2,253,000 14
ers through information products or presentations Wall Street Journal (NY) 2,107,000 12
prepared by press officers or other intermediaries, and New York Times (NY) 1,153,000 25
Los Angeles Times (CA) 973,000 13
then subsequently reported to the public through news Washington Post (DC) 759,000 29
stories. At issue is how much consonance exists among Rocky Mountain News (CO) 606,000 5
published results in scientific journals, translated scien- The Denver Post (CO) 575,000 4
tific information to the press, and the language and San Francisco Chronicle 519,000 5
numbers used by reporters in health news stories, (CA)
Boston Globe (MA) 459,000 10
especially for controversial issues?17 Seattle Times (WA) 385,000 12
The 2002 hormone therapy–related events provided Seattle Post-Intelligencer 385,000 9
an excellent opportunity to examine the accuracy of (WA)
the entire translation process, given that extensive San Jose Mercury News (CA) 274,000 12
efforts were made by leaders at JAMA and the NIH to Charlotte Observer (NC) 231,000 7
The News & Observer (NC) 168,000 5
convey fairly complex and controversial scientific find- Albuquerque Journal (NM) 108,000 2
ings to journalists. The research questions for this study Greensboro News & Record 94,000 4
were twofold: (1) Was the study’s key message, that (NC)
overall health risks outweighed benefits from hormone Winston-Salem Journal (NC) 84,000 8
therapy use, translated accurately by the scientific com- Oakland Tribune (CA) 65,000 7
Union Leader (NH) 57,000 3
munity and reported in newspapers; and (2) What type of Morning Star Wilmington 53,000 3
language and numbers were used to convey the relation- (NC)
ship between hormone therapy use and disease risk. Burlington Free Press (VT) 50,000 1
Santa Fe New Mexican (NM) 25,000 8
a
Based on paid circulation.
Methods
Data Collection
Data for this study were based on a 2006 analysis of text from The newspapers and articles selected for this study (Table 2)
several sources (Table 1): (1) the JAMA article itself6; (2) the were part of an existing data set created for a prior news
JAMA editorial18; (3) the JAMA press release issued in con- media research study, based on a search using LexisNexis
junction with the publication of the study and editorial19; (Dayton OH).23 The data set included local, regional, and
(4) the NIH press release about the study and decision to end nationally syndicated newspapers from the seven geographi-
the study early20; (5) the transcript from the NIH press cally dispersed sites of the National Cancer Institute’s Breast
conference21; and (6) hormone therapy news stories and Cancer Surveillance Consortium. The terms hormone, hor-
editorials appearing in 22 newspapers. mone therapy, hormone replacement therapy, estrogen, pro-
gesterone, WHI, menopause, and menopausal were searched
to identify relevant articles.23
From this existing data set, the authors identified hormone
Table 1. Documents analyzed
therapy stories and editorials published from July 8, 2002 (the
Source Author Publication date date the WHI story was first reported) through September 5,
JAMA article 6
Writing Group for the July 8, 2002 2002. The choice of the September cutoff date was consistent
WHI Investigators with previous research, which suggested that the majority of
JAMA Fletcher and Colditz July 8, 2002 articles about the WHI study were published within the first
editorial18 month of its release, with a gradual decline in stories over the
JAMA press JAMA press officer July 8, 2002 next month.23 Electronic versions of full-text articles were
release19 available for analysis. Duplicate publications were identified
NIH press NIH press officer July 8, 2002 manually and deleted. Included in the final analysis were 198
release20 newspaper articles: 183 news stories and 15 editorials.
NIH press NIH (federal) July 9, 2002
conference21 scientists Coding and Analysis
Newspaper Staff journalists, July 8–September
articles editors, and 5, 2002 The qualitative research method used to code and analyze the
syndicated text documents built on a prior descriptive analysis of news-
columnists paper articles conducted by Haas et al.23 NVivo software
JAMA, Journal of the American Medical Association; NIH, National (previously called NUD*IST; Version 1.3, 1999) was used to
Institutes of Health; WHI, Women’s Health Initiative. organize and manage the text for analysis. Initial analysis

62 American Journal of Preventive Medicine, Volume 34, Number 1 www.ajpm-online.net

included in-depth review and coding of each document.
Nodes, or specific themes, were identified and used to
describe the text content of each document. Codes were
categorized (as nodes) and themes extracted and compared
across text categories. Node development progressed, begin-
ning with the coding of the original JAMA article and
editorial, then the press-related materials, and concluded
with the news stories.
Analyses were based on four thematic nodes: (1) hormone
therapy health risks outweighed benefits (balance); (2) spe-
cific adverse health outcomes associated with hormone ther-
apy (health risk); (3) positive health outcomes associated with
hormone therapy (benefit); and (4) risk level
(magnitude), that is, numeric relative or absolute
risk estimates or text descriptors of risks (e.g.,
“minimal risk”) that interpret the quantitative
level of risk (Table 3). related
Commentary
Results by Katz in this
Hormone Therapy Risks, Benefits, and issue.
Balancing the Two
The primary objective of the WHI study was “to assess
the major health benefits and risks of the most com-
monly used combined hormone preparation.”6 This
objective was reiterated in the JAMA editorial, with an
emphasis on the statistical results: “The data and safety
monitoring board recommended stopping the trial
because women receiving the active drug had an in-
creased risk of invasive breast cancer (HR⫽1.26; 95%
CI⫽1.00 –1.59), and an overall measure suggested that
the treatment was causing more harm than good
(global index⫽1.15; 95% CI⫽1.03–1.28).”18
The message about risks outweighing benefits, the
major study finding, was included in both the JAMA
and NIH press releases and highlighted during the NIH
press conference:
There were some benefits to the use of estrogen
plus progestin–women on these hormones had
fewer hip fractures and colon cancers– but the
magnitude of these benefits was not sufficient to
overcome the accumulated risks observed for
other diseases.21
The terminology journalists used concerning balanc-
ing risks and benefits was often in the format that
hormone therapy “caused more harm than good”:
Researchers have halted a major clinical trial of
the most commonly used hormone-replacement
therapy used by postmenopausal women because
the risks of taking the hormones outweigh the
benefits. (Seattle Post-Intelligencer [WA], July 9,
2002)
It was cut short because it became clear that the
therapy, whatever benefit it provided against the
specific effects of menopause, caused more harm
than good in the long term: It increased the risk
January 2008
of breast cancer and cardiovascular problems.
(Boston Globe [MA], July 12, 2002)
For the directors of the study—and for a good
portion of the estimated 6 million women who
take hormone therapy—the risks of treatment
now outweigh the benefits. (Denver Post [CO],
July 16, 2002)
Reporters sometimes addressed the decision women
and their physicians needed to make for balancing the
increased specific health risks against hormone therapy
benefits:
For instance, a woman with a strong
See family history of breast cancer might
opt for alternative measures over hor-
mone replacement therapy. But a
woman with no breast cancer in the
family but with lots of family osteoporo-
sis might choose to stay with the estro-
gen–progestin hormone therapy, which
the study said did offer some protection from hip
fractures. (Seattle Times [WA], July 15, 2002)
But Lanka herself, who at age 62 has been on
hormone replacement therapy for 7 years, said
she isn’t planning to stop. “I’m telling patients
they have to make their own decisions,” Lanka
said. “For me, they’re worth it. You have to look at
your family history, and my family history is osteo-
porosis and colon cancer.” (San Francisco Chron-
icle [CA], July 10, 2002)
In an effort to translate the “balance” message,
reporters often countered more-dramatic risk language
by using more positive or moderating language about
benefits, or the weighing of risks against benefits. The
majority of reporters chose the latter approach and
focused on balancing risks and benefits, with 71% of all
articles containing such language. Overall, health risks
were reported in 80%, and benefits in 67%, of articles.
In summary, the primary message from the study—
balancing the risks and benefits of hormone therapy—
was generally translated accurately and consistently
from researchers to press officers to journalists to the
public.
Translating Risk to the Press
The JAMA article6 contained both relative and absolute
risk estimates for seven health outcomes; the JAMA
editorial18 repeated the relative and absolute risk data
from the article. Although both press releases and the
press conference consistently cited relative risk (e.g., a
27% increase, two times the risk) rather than absolute
risk estimates (seven cases for every 10,000 women),
these materials also relied on text descriptors of risk
magnitude (e.g., increase, high). Table 4 provides a
comparison of how absolute risk data were used across
Am J Prev Med 2008;34(1) 63
 64
American Journal of Preventive Medicine, Volume 34, Number 1

Table 3. Comparison of text used to describe findings from the 2002 hormone therapy study
JAMA article6 JAMA press release19 NIH press release20 NIH press conference21 Newspaper accounts
Hormone therapy health balance (health risks outweigh benefits)
In addition, the “In addition, the The reduction in colorectal Women who are taking A major trial to measure the effects of hormone
substantial risks for substantial risks for cancer risk in the WHI the therapy to prevent therapy was stopped Tuesday when it became
cardiovascular disease cardiovascular disease [Women’s Health osteoporosis should talk clear there are more minuses than pluses.
and breast cancer and breast cancer must Initiative] is intriguing, with their doctor and (Boston Globe [MA], July 12, 2002).
must be weighed be weighed against the but the balance of harm carefully weigh any The study found that risks clearly outweighed
against the benefit for benefit for fracture in versus benefit does not benefit against their benefits – namely fewer fractures and less
fracture in selecting selecting from the justify any woman personal risks for colorectal cancer—after five years of therapy
from the available available agents to beginning or continuing cardiovascular disease (USA Today, July 11, 2002).
agents to prevent prevent osteoporosis,” to take estrogen plus and breast cancer. Since it was reported that the National
osteoporosis. the authors conclude. progestin for this Institutes of Health halted part of a major
purpose. study evaluating hormone-replacement
therapy because its risks outweighed its
benefits, doctors’ phones have buzzed. (Rocky
Mountain News [CO], July 11, 2002).
Hormone therapy health risks (adverse hormone therapy health outcomes)
The trial was stopped Researchers have stopped The Data Safety It [the Women’s Health A government study involving 16,000 women
early based on health the estrogen plus Monitoring Board’s May Initiative] has been was halted in July when it was found that
risks that exceeded progestin portion of 31 recommendation to stopped early due to an hormone replacement therapy can increase a
health benefits over the Women’s Health stop the trial was based increased risk of invasive woman’s risk of heart attack and stroke –
an average follow-up Initiative, a clinical trial on the finding of breast cancer, and even women with no history of the heart
of 5.2 years. designed to assess the increased breast cancer evidence that overall disease (Burlington Free Press [VT], September
major health benefits risk, supported by the health risks exceed any 4, 2002).
and risks of the most evidence of overall health benefits. The Women’s Health Initiative, a large-scale
commonly used risks exceeding any study on hormone therapy intended to last
hormone preparation benefits. until 2005 and backed by the National
in the United States on Institutes of Health, was abruptly ended last
healthy menopausal week because participants showed increased
women, after overall incidence of breast cancer, blood clots, stroke
health risks were found and heart attack (Denver Post [CO], July 16,
to exceed the health 2002).
benefits. Long-term use of the pill increases the risk of
coronary heart disease, invasive breast cancer,
strokes and blood clots in the lungs, the
researchers said (Wall Street Journal [NY], July
www.ajpm-online.net

10, 2002).

(continued on next page)

January 2008

Table 3. (continued)
JAMA article6 JAMA press release19 NIH press release20 NIH press conference21 Newspaper accounts
Hormone therapy health benefits (positive hormone therapy health outcomes)
This cohort experienced The benefits included a A 37 percent reduction in Women taking the therapy The researchers also found in their study that
low hip fracture rates. 37 percent reduction in cases of colorectal had a 34 percent hormone replacement therapy led to
Estrogen plus colorectal cancer rates, cancer. A one-third reduction in hip reductions in risks for colorectal cancer and
progestin reduced the hip fracture rates reduction in hip fracture fractures and 24 percent hip fractures (Los Angeles Times [CA], July 10,
observed hip and reduced by one-third rates. A 24 percent reduction for total 2002).
clinical vertebral and total fractures reduction in total fractures. The estrogen Those risks outweighed the drugs’ benefits – a
fracture rates by one reduced by 24 percent. fractures. plus progestin therapy small decrease in hip fractures and a decrease
third compared with also produced a 37 in colorectal cancer (New York Times [NY],
placebo, both percent reduction in the July 10, 2002).
nominally and risk of colorectal cancer. . . . the hormones did have some benefits—
significantly. reducing the frequency of hip fractures and
colon cancer . . . (Oakland Tribune [CA], July
10, 2002).
On the plus side, they [hormone therapy] also
had lower levels of colorectal cancer and hip
fractures (Rocky Mountain News [CO], July 11,
2002).
Relative risk magnitude: breast cancer (hormone therapy risk level)
The invasive breast The invasive breast The invasive breast cancer The invasive breast cancer Overall, researchers found a 26 percent
cancer rates in the cancer rates in the rates in the placebo rates in the placebo increased risk for breast cancer among
placebo group were placebo group were group were consistent group were consistent women who took the combination therapy
consistent with design consistent with design with design expectations. with design compared to those who got a placebo
expectations. The 26% expectations. The 26% The 26% increase (38 vs. expectations. The 26% (Charlotte Observer [NC], July 9, 2002).
increase (38 vs. 30 per increase (38 vs. 30 per 30 per 10,000 person- increase (38 vs. 30 per The study found that a daily regimen of
10,000 person- 10,000 person- years) . . . almost reached 10,000 person- estrogen and progestin led to a 26 percent
years) . . . almost years) . . . almost nominal statistical years) . . . almost increase in breast cancer (The News &
reached nominal reached nominal significance. reached nominal Observer [Raleigh, NC], July 9, 2002).
Am J Prev Med 2008;34(1)

statistical significance. statistical significance. statistical significance. . . . they [risks] were dramatic for the study
population as a whole – the hormone therapy
was linked to a 41 percent increase in strokes
and a 26 percent increase in breast cancer
(San Jose Mercury News [CA], August 5, 2002).
However, women on the hormone treatment
had a 26 percent higher incidence of breast
cancer (Washington Post [DC], July 10, 2002).
65
Table 4. Absolute risk compared across documents
Health JAMA article6 and NHLBI press
outcome editorial18 conference21 Accurate newspaper accounts
Invasive breast 8 more invasive It [individual breast For example, (WHI) study
cancer: breast cancers cancer risk] was results show that during
Individual per 10,000 less than a tenth one year, for every 10,000
risk. person years. of 1 percent per women, eight more will
year. have breast cancer if they
are taking estrogen and
progestin than if they are
not (Charlotte Observer,
[NC] July 9, 2002).
For any individual woman,
however, the risk was
extremely slight - less than
a 10th of a percentage
point a year for breast
cancer (Boston Globe, [MA]
July 12, 2002).
According to the data, if
10,000 women took the
drugs for a year and 10,000
did not, women in the first
group would have 8 more
cases of invasive breast
cancer (New York Times,
[NY] July 15, 2002).
Inaccurate newspaper accounts
. . . it still translates to a
relatively small increase in
risk to an individual woman.
Breast cancer hit 1.5 percent
of women on the placebo,
and 2 percent on the
hormone combination
(Albuquerque Journal [NM],
July 29, 2002).
But the chance that an
individual woman on
hormone therapy will get
breast cancer is 38 in 10,000
or less than four-tenths of a
percent (Seattle Times [WA],
July 14, 2002).
. . . the increased risk of cancer
and heart disease is small –
less than one-half of 1
percent per year . . . (Santa
Fe New Mexican [NM],
August 18, 2002).

the documents. Despite the overall consistency in re-
porting risk numbers, there were variations.
In the JAMA press release,19 only relative risk data
from the hormone therapy study were included. Al-
though the NIH press release20 included both relative
and absolute risk data, absolute risk estimates were
mentioned within one paragraph while relative risk
estimates were placed in a separate table. During the
NIH press conference,21 hormone therapy benefits
were presented only in relative risk terms and not
heavily emphasized; relative risk data were presented
for all of the major risk findings, with absolute risk
addressed only for breast cancer.
Translating Risk to the Public
A total of 76% of newspaper articles reported risk data
using text descriptors; 30% of articles included absolute
risk and 19% reported relative risk numbers (note:
percents total to more than 100% because some articles
used more than one method). This greater use of
absolute over relative risk numbers was unexpected,
especially given the press release and press conference
materials. There was evidence that although overall re-
porting of absolute risk data was accurate, some journal-
ists struggled to interpret correctly the meaning of the
numbers. Table 4 presents examples of accurate and
inaccurate reporting of absolute data for breast cancer.
Risk-level text descriptors focused primarily on ad-
verse health impacts; language ranged from modest
characterizations (e.g., “slightly increases the likeli-
hood”) to much stronger language (“significantly in-
creases”). A similar range of terminology was evident
for reporting health impacts, ranging from neutral
language (e.g., developed [a specific disease or condi-
tion]) to more dramatic language (e.g., suffered from
[a specific disease or condition]). This range of risk
text descriptors was seen even in the same newspaper in
articles published 2 days apart, highlighting the inter-
pretation, or “spin” used by individual reporters:
A type of hormone replacement therapy taken
daily by 7 million women nationwide significantly
increases the risk of breast cancer, strokes and
heart attacks, according to the National Institute
of Health. (Greensboro News & Record [NC],
July 10, 2002)
Gynecologists are reacting cautiously to the news
that a combination of estrogen and progestin,
often taken in the pill Prempro, slightly increases
the likelihood of strokes, blood clots, heart attacks
and breast cancer, according to the National
Institute of Health. (Greensboro News & Record
[NC], July 12, 2002)
In summary, although relative and absolute risk data
were included in the JAMA article and editorial, trans-
lational information provided to journalists more
strongly emphasized relative risk for all health condi-
tions except breast cancer. Although journalists who
included numeric risk-level estimates in their articles
were more likely to include absolute rather than rela-
tive risk numbers, text descriptors of risk level were
more commonly used than the numbers themselves.

66 American Journal of Preventive Medicine, Volume 34, Number 1 www.ajpm-online.net

Discussion
The frame of a news story is the way the news media
tells people how to think about an issue.24,25 The type
of news sources used by reporters can influence fram-
ing, and journalists commonly use health experts,
including representatives of federal agencies, as
sources.26,27 Despite these common practices, there
has been much criticism of how journalists cover med-
ical issues, especially about reporting on health risks
and their magnitude.2,4,28,29 In this study, the primary
message that the risks of hormone therapy outweighed
the benefits was present in nearly two thirds of stories,
and probably contributed to the subsequent decline in
use.14,15 Although only about one third of stories
included numeric risk estimates, absolute rather than
relative risk numbers were used more often, which
contrasted with prior research.30 These findings sug-
gest that the message translation process for conveying
complex findings from the July 2002 JAMA hormone
therapy study to communication intermediaries, and to
newspaper journalists, was generally quite successful.
There was variation in the extent to which hormone
therapy benefits were mentioned in newspapers, as well
as the risk level associated with hormone therapy based
on the choice of text modifiers (language) by reporters.
Consequently, different interpretations by reporters
about the meaning of the hormone therapy findings
were evident. Language used to report risk-level esti-
mates was subjective and malleable, depending on the
particular angle of the journalist, which clearly affected
the tone of the article and the potential interpretation
by readers. However, the fact that absolute rather than
relative risk numbers were used more often by report-
ers suggested that they may have reviewed other sources
of information available to them, such as the original
JAMA article and editorial.
It is worth considering factors that likely contributed
to the successful hormone therapy translation. First,
there was a well-coordinated effort among researchers,
JAMA staff, and NIH employees to develop and accu-
rately communicate key study messages at each step
of the translation process. Second, the work of commu-
nication intermediaries was essential. Referred to as
science-information professionals, press officers, or
public affairs specialists, they are the individuals who
create written documents summarizing major points in
scientific reports, regularly communicate with report-
ers by phone or email, assist scientists to work more
effectively with reporters, and serve as organizational
spokespersons.31 NIH communication staff prepared
and provided extensive written background materials
to reporters about the hormone therapy findings and
early termination of the WHI trial; this may have
encouraged journalists to comprehensively review ma-
terials, including the original JAMA study and editorial,
and consider the impact of their stories on the public.
January 2008
Third, messages were communicated to news media
representatives through several communication chan-
nels. Research has demonstrated a positive relationship
between press releases and subsequent news media
reporting32; thus it is likely that the more avenues used
to reach out to reporters, especially from highly credi-
ble information sources such as JAMA and NIH, the
greater the attention of reporters to the study findings.
Finally, the fact that absolute risk estimates were used
more often than relative risk estimates suggests that
some reporters who cover health and science topics are
becoming more knowledgeable about scientific meth-
ods, terminology, and limitations.2
Although not part of this report, no separate process
was used by JAMA or NIH to communicate hormone
therapy findings and implications directly to physicians
or other healthcare providers. Furthermore, no recom-
mendations were made to healthcare providers as to
what they should specifically communicate to their
patients as a result of the new hormone therapy find-
ings. This meant that healthcare providers were in-
formed about the study at the same time and in the
same manner as their patients: through the news me-
dia. Consequently, they often lacked knowledge about
the study and were not prepared to deal with the many
phone calls and office visits from concerned pa-
tients.33,34 An important lesson from the hormone
therapy experience is that when it is evident that a
major health study or report with implications for
medical practice occurs, extensive communication ef-
forts and practical guidance to physicians and other
healthcare providers may be needed.
This study has several limitations. First, one individ-
ual conducted all analyses. With one analyst, consis-
tency in coding was weighed against a single interpre-
tation. Efforts were taken, however, to maintain
scientific rigor through the use of strict interpretive
evaluation criteria.35 These included the foundational
criteria of credibility, transferability, and dependability;
ongoing collaboration and review of coding with a
research team; using a tracking system for audit pur-
poses35; and comparing data heterogeneity for the-
matic patterns and reaching consensus.36
Second, analyses were limited to the 22 newspapers
in an existing data set, and thus may not be represen-
tative of hormone therapy news coverage throughout
the U.S. Finally, analyses were limited to newspapers
only. Analyzing hormone therapy stories from other
news media outlets (e.g., television37), might have
resulted in different study results and potentially influ-
enced lay audiences in alternate ways. There are differ-
ences in the ways in which newspapers, television,
radio, and Internet news outlets select, frame, and
utilize sources for news stories.38 – 40 A recent study
found differences in reporting about tobacco issues
between newspapers and television,41 although other
Am J Prev Med 2008;34(1) 67
research suggests similarities in newspaper and televi-
sion reporting of health stories.31
A consistent, accurate message, even if complex,
when communicated in multiple ways increases the
likelihood of successfully translating scientific informa-
tion to journalists and the public. We concur with
Steinbrook42 that medical journals and news media
need to better define and understand their respective
roles, respecting the important contributions that med-
ical journals and medical reporters can make toward
improving health. Further, it is particularly important
for scientists to become familiar with the role of the
news media in disseminating scientific information and
learn how to work with, not against them, given the key
role the press has in educating the public about health
risks.2
The authors’ acknowledge financial support for the study
from the National Cancer Institute, an institute of the Na-
tional Institutes of Health, Department of Health and Human
Services, United States Government under contract
263-MQ-515940.
No financial disclosures were reported by the authors of
this paper.
References
1. Dearing J, Rogers E. Agenda setting. Thousand Oaks CA: Sage, 1996.
2. Schwartz LM, Woloshin S. The media matter: A call for straightforward
medical reporting. Ann Int Med 2004;140:226 – 8.
3. Fletcher SW. Whither scientific deliberation in health policy recommenda-
tions? Alice in Wonderland in breast cancer screening. N Eng J Med
1997;336:1180 –3.
4. Gregory J, Miller S. Science in public: communication, culture, and
credibility. Cambridge MA: Basic Books, 1998.
5. Bennett P, Calman K, eds. Risk communication and public health. Oxford:
Oxford University Press, 1999.
6. Writing Group for the Women’s Health Initiative Investigators. Risks and
benefits of estrogen plus progestin in healthy postmenopausal women:
principal results from the Women’s Health Initiative randomized con-
trolled trial. JAMA 2002;288:321–33.
7. Barrett-Connor E, Grady D. Hormone replacement therapy, heart disease,
and other considerations. Annu Rev Public Health 1998;19:55–72.
8. Agency for Healthcare Research and Quality (AHRQ). Two new hormone
replacement therapy reviews support findings of recently halted clinical
trial. Press Release, August 20, 2002. Rockville MD.: AHRQ. Available
online at: www.ahrq.gov.news.press/pr2002/hrtevpr.htm.
9. Bromley SE, de Vries CS, Farmer RDT. Utilisation of hormone replacement
therapy in the United Kingdom. A descriptive study using the general
practice research database. Br J Obstet Gynecol 2004;111:369 –76.
10. Heitmann C, Greiser E, Doren M. The impact of the Women’s Health
Initiative randomized controlled trial 2002 on perceived risk communica-
tion and use of postmenopausal hormone therapy in Germany. Menopause
2005;12:405–11.
11. Mishra G, Kok H, Ecob R, Cooper R, Hardy R, Kuh D. Cessation of
hormone replacement therapy after reports of adverse findings from
randomized controlled trials: evidence from a British birth cohort. Am J
Public Health 2006;96:1219 –25.
12. Taylor AW, Maclennan AH, Avery JC. Postmenopausal hormone therapy:
who now takes it and do they differ from non-users? Aust N Z J Obstet
Gynecol 2006;46:128 –35.
13. Breslau ES, Davis WW, Doner L, et al. The hormone therapy dilemma:
women respond. J Am Med Women Assoc 2003;58:33– 43.
68 American Journal of Preventive Medicine, Volume 34, Number 1
14. Bastian LA, Breslau ES, Davis WW, Moser RP. Hormone therapy for
menopausal symptoms: Tracking the translation of the Women’s Health
Initiative findings. J Clin Outcome Manag 2005;12:140 – 6.
15. Hass JS, Kaplan CP, Gerstenberger EP, Kerlikowske K. Changes in the use
of postmenopausal hormone therapy after the publication of clinical trial
results. Ann Int Med 2004;140:184 – 8.
16. Lawton B, Rose S, McLeod D, Dowell A. Changes in use of hormone
replacement therapy after the report from the Women’s Health Initiative:
cross sectional survey of users. Br Med J 2003;327:845– 6.
17. Whiteman MK, Cui Y, Flaws JA, Langenberg P, Bush TL. Media coverage of
women’s health issues: is there a bias in the reporting of an association
between hormone replacement therapy and breast cancer? J Women
Health Gender Med 2001;10:571–7.
18. Fletcher SW, Colditz GA. Failure of estrogen plus progestin therapy for
prevention. JAMA 2002;288:366 – 8.
19. JAMA. JAMA media advisory: Health risks outweigh benefits for combined
estrogen plus progestin; clinical trial stopped early in major study. Chicago:
American Medical Association, 2002.
20. National Institutes of Health (NIH). News release: NHLBI stops trial of
estrogen plus progestin due to increased breast cancer risk, lack of overall
benefit. Bethesda MD: National Heart, Lung, and Blood Institute
(NHLBI), 2002. Available online at: http://www.nhlbi.gov/new/press.
21. NHLBI, NIH. Press conference remarks: release of the results of the estrogen
plus progestin trial of the Women’s Health Initiative. Bethesda MD: NHLBI,
NIH, 2002. Available online at: http://www.nhlbi.gov/new/press.
22. Audit Bureau of Circulations (ABC), March 2003 to March 2004. Available
online at: http://www.accessabc.com/.
23. Haas JS, Geller B, Miglioretti DL, et al. Changes in newspaper coverage
about hormone therapy with the release of new medical evidence. J Gen Int
Med 2006;2:304 –9.
24. Gans HJ. Deciding what’s news: a study of CBS Evening News, NBC Nightly
News, Newsweek, and Time. New York: Pantheon, 1979.
25. Reese SD, Gandy OH, Grant AE, eds. Framing public life: perspectives on
our understanding of the social world. Mahwah NJ: Erlbaum, 2001.
26. Nelkin D. How the press covers science and technology. New York: WH
Freeman, 1995.
27. Manning P. News and news sources: a critical introduction. London: Sage,
2001.
28. Woloshin S, Schwartz LM. Press releases: translating research into news.
JAMA 2002;28:2856 – 8.
29. Chew F, Mandelbaum-Schmid J, Gao SK. Can health journalists bridge the
state-of-the-science gap in mammography guidelines? Sci Comm 2006;27:
331–51.
30. Moynihan R, Bero L, Ross-Degnan D, et al. Coverage by the news media of
the benefits and risks of medications. New Engl J Med 2000;342:1645–50.
31. Weingold MF. Communicating science. Sci Comm 2001;23:164 –93.
32. de Semir V, Ribas C, Revuelta G. Press releases of science journal articles and
subsequent newspaper stories on the same topic. JAMA 1998;280:294 –5.
33. Connolly C. Doctors working to clear the fog of hormone study. The
Washington Post, July 28, 2002, p. A1.
34. Fontanarosa PH, DeAngelis CD. The importance of the journal embargo.
JAMA 2002;288:748 –50.
35. Guba EG, Lincoln YS. Fourth generation evaluation. Newbury Park CA:
Sage, 1989.
36. Schensul JJ, Lecompte MD. Ethnographer’s toolkit: using ethnographic
data, researcher roles and research partnerships, analyzing and interpret-
ing ethnographic data, mapping social networks, and spatial data. Walnut
Creek CA: Altamira Press, 1999.
37. Pribble JM, Goldstein KM, Fowler EF, Greenberg MJ, Noel SK, Howell JD.
Medical news for the public to use? What’s on local TV news? Am J Manag
Care 2006;12:170 – 6.
38. Kerbel MR. If it bleeds, it leads: an anatomy of television news. Boulder CO:
Westview Press, 2000.
39. Project for Excellence in Journalism. The state of the news media 2004.
Washington DC: Project for Excellence in Journalism, 2004. Available
online at: http://www.stateofthenewsmedia.org/2004/index.asp.
40. Project for Excellence in Journalism. The state of the news media 2007.
Washington DC: Project for Excellence in Journalism, 2007. Available
online at http://www.stateofthenewsmedia.com/2007/.
41. Nelson DE, Evans WD, Pederson LL, Babb S, London J, McKenna J. A
national surveillance system for tracking tobacco news stories. Am J Prev
Med 2007;32:79 – 85.
42. Steinbrook R. Medical journals and medical reporting. N Eng J Med
2000;342:1668 –71.
www.ajpm-online.net