Content uploaded by Rebecca Blaylock
Author content
All content in this area was uploaded by Rebecca Blaylock on Jun 23, 2021
Content may be subject to copyright.
Available via license: CC BY
Content may be subject to copyright.
RESEARCH ARTICLE
Risk messages relating to fertility and pregnancy: a media
content analysis [version 1; peer review: 1 approved]
Olivia Marshall 1, Rebecca Blaylock 1, Clare Murphy 1, Julia Sanders 2
1Centre for Reproductive Research and Communication, British Pregnancy Advisory Service (BPAS), London, EC4A 1JQ, UK
2School of Healthcare Sciences, Cardiff University, Cardiff, CF14 4XN, UK
First published: 14 May 2021, 6:114
https://doi.org/10.12688/wellcomeopenres.16744.1
Latest published: 14 May 2021, 6:114
https://doi.org/10.12688/wellcomeopenres.16744.1
v1
Abstract
Background: The UK print and online media is an important channel
by which scientific research is communicated to the public. Media risk
messages relating to pregnancy or fertility contribute to the context of
reproductive decision making, but their fidelity to the underlying
science has been questioned.
Method: We measured the volume, distribution and content of
science-based risk headlines relating to pregnancy or fertility in the
UK media over four months. We grouped headlines into unique
stories and categorised them by exposure and outcome of interest.
We selected four unique stories for closer content analysis and
assessed their fidelity to the underlying science, with attention to the
role of press releases.
Results: We identified 171 headlines over four months (average 43
per month), comprising 56 unique stories. The unique stories most
commonly concerned maternal risk factors (n=46) and child health
outcomes (n=46). Maternal health outcomes were less frequently the
focus (n=20). The most common risk factors in the media coverage
were maternal food and drink (n=15), maternal medication and
medical interventions (n=9), and maternal health factors (n=6). Media
reports were largely faithful to press releases. Where substantive
deviations from the underlying scientific study were identified, these
could mostly be traced back to press releases or quotes from the
study’s authors. Press releases often omitted caveats which were
reinstated at the media reporting stage, alongside additional expert
criticism.
Conclusions: Frequent science-based risk messages in the UK media
frame mothers as vectors of potential harm to children, who are the
focus of health outcomes. Largely, the media does not introduce
misinformation, but reports press releases faithfully with additional
caveats and expert commentary. Press releases fulfil an interpretative
role, often omitting caveats and introducing new elements and advice
to women. Their role as a bridge between scientific and lay audiences
is discussed.
Open Peer Review
Reviewer Status
Invited Reviewers
1
version 1
14 May 2021 report
Norah MacKendrick, Rutgers University,
New Brunswick, USA
1.
Any reports and responses or comments on the
article can be found at the end of the article.
Page 1 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
Corresponding author: Olivia Marshall (olivia.marshall@bpas.org)
Author roles: Marshall O: Conceptualization, Formal Analysis, Investigation, Methodology, Visualization, Writing – Original Draft
Preparation; Blaylock R: Conceptualization, Formal Analysis, Investigation, Methodology, Writing – Review & Editing; Murphy C:
Conceptualization, Writing – Review & Editing; Sanders J: Conceptualization, Writing – Review & Editing
Competing interests: No competing interests were disclosed.
Grant information: This paper is part of the WRISK project, funded by the Wellcome Trust [212089/Z/18/Z].
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Copyright: © 2021 Marshall O et al. This is an open access article distributed under the terms of the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
How to cite this article: Marshall O, Blaylock R, Murphy C and Sanders J. Risk messages relating to fertility and pregnancy: a media
content analysis [version 1; peer review: 1 approved] Wellcome Open Research 2021, 6:114
https://doi.org/10.12688/wellcomeopenres.16744.1
First published: 14 May 2021, 6:114 https://doi.org/10.12688/wellcomeopenres.16744.1
Keywords
Risk; fertility; pregnancy; media; science communication; health
Page 2 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
Introduction
The UK media is an important channel through which new
health research is disseminated to the general public1–3. Health
stories are a particular staple of UK media reporting and are
frequently communicated in the form of ‘risk messages’, wherein
common exposures, habits, or lifestyle characteristics are posi-
tioned as risk factors for certain health outcomes2,3. Science-
based health messages in the media influence the public’s
understanding of health risks and their health-related decisions
and behaviours4–6.
Scientific studies are commonly communicated to journalists
via press releases (PRs), which bridge the academic and media
communities, enabling research institutions to generate cover-
age, and journalists to produce swift copy that conveys complex
findings to a lay readership1. PRs thus provide an important
step on the risk reporting pathway between scientific papers and
media reports (Figure 1).
Concerns have been raised that the media’s reporting of health
science is sensationalist, inaccurate, and undermines the intended
messaging of the scientific community1,3,7,8. A number of stud-
ies have examined the role of PRs in science communication,
and found they may be responsible, especially through the
removal of caveats or the framing of correlational associa-
tions as causative1,2,7–9. While a certain degree of interpretation is
necessary for PRs to fulfil their bridging function, such changes
in the framing of risk messages at the PR stage are likely to
later influence their public interpretation and reception, and
therefore risk a detrimental impact on public health2,4,10,11.
Pregnancy risk is an area of contemporary debate and atten-
tion. Previous research has found that pregnancy is increasingly
portrayed as a high-risk state, during which women must be
closely surveilled, not for the sake of their own health but to
protect the fetus from risks introduced by their behaviour12–16.
This narrative can be traced back to the thalidomide disaster
of the 1960s17. Its reinvigoration in recent years has partly been
fueled by research into the developmental origins of health
and disease (DOHaD), which has recently focused on the impact
of chemical exposures in utero on fetal health outcomes18–20.
This paper anlayses risk messages in the UK media relating to
pregnancy and fertility, to better understand the context that
informs women’s reproductive health decisions and behaviours,
and to assess the extent to which (if at all) pregnant women are
framed as vectors of potential harm (or benefit) to their fetus.
Secondly, the fidelity of media reports to the underlying science
is assessed through closer analysis of four illustrative examples.
For each example, the communication of risk messages along
the risk reporting pathway is examined, with particular attention
to the role of PRs.
Objectives
1. Understand the landscape: Describe the volume, dis-
tribution and content of science-based risk messages
relating to pregnancy or fertility, as reported in UK-
based print and online media outlets, over selected time
periods.
2. Assess reliability: For selected news stories based on
a scientific study, describe and map the risk reporting
Figure 1. The risk reporting pathway (Wrisk project, 2020)21.
Page 3 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
pathway(s) “from study to story”, by tracking key
themes and informational elements across the original
study manuscript, PR and media reports.
Methods
Objective 1: Understand the landscape
We used publicly available media reports. There were no
particular ethical considerations.
Selection of relevant time period(s)
We randomly selected one month from each of the most recent
four quarters (Q4, 2018 – Q3, 2019) for analysis, to accommodate
possible seasonal variations in media reporting. The months
selected were November 2018, and February, May and August
2019 (the second month of each quarter).
Search of media database
Using the LexisNexis news database, headlines published in
UK print or online media outlets during the selected months
were searched for relevant key words (“babies” or “unborn”
or “pregnant” or “pregnancy” or “fertility” or “miscarriage”
or “stillbirth” or “stillborn”).
Exclusion process
The inclusion and exclusion criteria were agreed by two
researchers (OM and RB). The full exclusion process is
summarised in Figure 2.
Figure 2. Exclusion ow diagram.
Page 4 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
Inclusion criteria: Headlines relating to a social, environ-
mental, or clinical risk to fertility, a pregnant woman or her
offspring, and based on a new scientific study (including
research conducted/published outside of the UK).
Exclusion criteria: Headlines from newswires, broadcast
media or media outlets with a predominant readership outside of
the UK; headlines with no risk element (including celebrity,
entertainment, or miscellaneous news stories); headlines about
the postnatal period or infant care; and headlines not based
on a new scientific study.
Analysis
Stories were deemed to be based on a scientific paper if
they mentioned a new scientific study or research, and the
corresponding paper was identified and noted based on its key
findings, authors and date of publication. We grouped media
reports based on the same scientific study into “unique
stories” and assigned a Unique Story ID (USID). For each
month we calculated the number of headlines and unique
stories. We calculated the number of headlines per unique
story, and the mean and median overall.
We coded each unique story according to the exposure and
outcome of interest. We categorised exposures by risk locus
(maternal; paternal; offspring; other) and by topic. We
categorised outcomes by outcome locus (maternal; paternal;
offspring; other). Two researchers (OM and RB) undertook this
process independently and established and a consensus. Some
unique stories were assigned to multiple categories as appro-
priate. For each category we calculated the number of relevant
unique stories. We used Microsoft Excel (v2103) for our
analysis.
Objective 2: Assess reliability
Selection of illustrative examples
We selected the unique story that generated the most head-
lines in each month and assessed the reliability of associated
media report. Four unique stories selected were based on
the following studies:
November 2018: McQuire, C, Mukherjee, R, et al. (2019).
Screening prevalence of fetal alcohol spectrum disorders in
a region of the United Kingdom: A population-based birth-
cohort study. Preventive Medicine, 118, 344–351. doi: 10.1016/
j.ypmed.2018.10.013
February 2019: Golding, J, Gregory, S, et al. (2019). Mater-
nal prenatal external locus of control and reduced mathematical
and science abilities in their offspring: A longitudinal birth
cohort study. Frontiers in Psychology, 10(Feb) doi: 10.3389/
fpsyg.2019.00194
May 2019: Gignac, F, Romaguera, D, et al. (2019). Maternal
nut intake in pregnancy and child neuropsychological devel-
opment up to 8 years old: a population-based cohort study in
Spain. European Journal of Epidemiology, 34(7), 661–673.
doi: 10.1007/s10654-019-00521-6
August 2019: Zhang, T, Sidorchuk, A, et al. (2019). Associa-
tion of Cesarean Delivery With Risk of Neurodevelopmental and
Psychiatric Disorders in the Offspring: A Systematic Review
and Meta-analysis. JAMA network open, 2(8), e1910236. doi:
10.1001/jamanetworkopen.2019.10236
Collating of materials
We collated the following documents for each illustrative
example:
· Original study manuscript
· PR as published by the university or research institution
(where applicable)
· Communications about the study published by the
Science Media Centre (SMC) (where applicable)
· Print and online media coverage across UK media
outlets
To fulfil the objective of assessing media accuracy, we consid-
ered it appropriate to analyse all the media coverage of these
four unique stories, including any articles not captured by
our original search. This was especially important since
previous research has indicated that as an individual risk
story ages, the tone of the media discourse may evolve2. To
ensure completeness, we conducted additional searches using
the LexisNexis database, with wider date parameters and more
specific keywords (for example “FASD”, etc.). Once we had
collated all media reports, we removed duplicates (defined as
an identical article published by the same news outlet on the
same day).
Coding of media reports, PRs and studies
Following methodological precedent from Lee, Sutton and
Hartley (2016) and Reisch and Spiegelhalter (2011), for each
illustrative example we identified and coded key themes and
informational elements in the original study, PRs, and media
reports. Two researchers (OM and RB) extracted this infor-
mation independently and agreed a summary by a process of
consensus.
The information extracted was grouped under four broad
headings:
1. Description of study and findings (including study design,
exposure and outcome of interest, use of statistics, description
of association, use of language, etc.)
2. Caveats, criticism or study limitations (e.g. a specific
statement that causation was not established)
3. Associated discussion topics (framing)
4. Advice, warnings, or reassurance to the public (both direct
and indirect)
Mapping and descriptive analysis
For every document, we noted the presence or absence of each
theme or informational element, and calculated the number of
Page 5 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
media reports containing each theme or informational element.
This allowed us to compare the content of the original study,
PR and media reporting.
We undertook a descriptive analysis to trace the themes and
informational elements along the risk reporting pathway,
according to the four groupings outlined above.
Results
Objective 1: Understand the landscape
Volume and distribution of headlines and unique stories
171 headlines were identified that met the inclusion criteria,
which were mostly concentrated in November 2018 and May
2019. Analysis of the 171 articles found that they comprised
56 unique stories (based on 56 research studies). Again, the
unique stories were concentrated in November 2018 and May
2019 (Figure 3). Per month, the mean number of headlines was
43 and unique stories 14.
Unique stories generated varying numbers of headlines. The
median number of headlines per unique story was 2, but some
stories achieved significantly more headlines, as illustrated in
Figure 4.
Content of unique stories
Of the 56 unique stories, maternal risk factors (n=46) and
child health outcomes (n=46) were most commonly identified.
Maternal health outcomes were less frequently the focus
(n=20). The categorisation of exposures and outcomes by
subject is given in Figure 5.
The most common risk factors were maternal food and drink
(n=13), maternal medication and medical interventions
(n=9), and maternal health factors, including genetic factors or
underlying conditions (n=6). The categorisation of exposures
by topic is given in Figure 6.
Objective 2: Assess reliability
Illustrative example 1
Documents:
• Study: McQuire, C.,et al. (2019)
• Press release: University of Bristol (2019). First UK
prevalence estimate FASD
• Science Media Centre (2018). expert reaction to screen-
ing study on uk prevalence of fetal alcohol spectrum
disorders (FASD)
Figure 3. Unique stories and headlines per month.
Page 6 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
Figure 4. Headlines per unique story.
Page 7 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
Figure 5. Exposure and outcome of interest, categorised by subject.
Figure 6. Exposure of interest, categorised by topic.
• Science Media Centre (2018). uk prevalence of fetal
alcohol spectrum disorders
• Media reports (n=16)
Study summary:
Novel screening algorithms were applied to data from the
ALSPAC longitudinal birth cohort study to estimate the screening
prevalence of fetal alcohol spectrum disorder (FASD). Differ-
ent missing data strategies were evaluated to yield prevalence
estimates ranging from 6–17%.
Comparison of study, PR and media reports:
Key changes at each stage of the risk reporting pathway
are summarised in Figure 7.
Description of the research and its findings: The study’s
design, exposure and outcome of interest were accurately
described in the PR and in 15 of the 16 media reports. The one
remaining report included a passing reference to one of
the study’s secondary findings. The highest prevalence esti-
mate of 17% was quoted accurately in the PR and the majority
Page 8 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
Figure 7. Key changes along the risk reporting pathway (illustrative example 1).
of media reports (n=13), but eight reports also approximated this
to “one in six”, and two to “one in five”, in their headlines. The
PR used appropriate language to convey uncertainty (e.g. “fea-
tures consistent with FASD”; “children could have symptoms”;
“up to 17%”) and presented the figure of 17% as an estimate
throughout, including in the headline. Three quarters of media
reports (n=12) made clear somewhere in the report that the
figure was an estimate, but seven also presented the figure as
certain at least once, in statements such as “17% of babies are
harmed by mothers’ drinking”. The PR did not state that the
prevalence estimate had varied by screening method or include
the lower estimates (which had been obtained via different
imputation methods for missing data). Only one media report
(in The Guardian) reinstated this information.
Caveats, criticism, and study limitations: The caveat that a
positive screen for FASD is not the same as a formal diagnosis
was present in the study, the PR and half of the reports (n=8).
The caveat that a causative association had not been estab-
lished (between prenatal alcohol exposure and the developmen-
tal outcomes recorded) was omitted in the PR, but reinstated in
the majority of media reports (n=10). Ten media reports included
additional expert criticism sourced from either the Science
Media Centre (n=5) or a press comment issued by the British
Pregnancy Advisory Service (BPAS) (n=10), including a
criticism that the study was “causing needless alarm”, which
was cited in eight reports. Overall, the majority of media reports
(n=10) included at least one caveat or critical statement, with
most including three or more (n=9). Of the six media reports that
included no criticism or caveats, two were very short (50
words or less) and two were passing references to the study in
articles published several months later.
Associated discussion topics: The PR developed discussion
topics from the study, including UK rates of prenatal alcohol
exposure (PAE); the prevalence of binge drinking among preg-
nant women; conceptualising FASD as an under-diagnosed or
“hidden” condition; and framing FASD as a significant public
health concern. The PR also introduced one new topic, low
public awareness of FASD, which was mentioned in half of the
media reports (n=8). The discussion topics in the media cover-
age closely mirrored those in the PR, especially the placement
of FASD as a public health concern and discussion of PAE
rates in the UK.
Advice, warnings and reassurance: The PR introduced
direct advice to women in a quote by the study’s lead author,
who referred to guidance that the safest approach is to abstain
from alcohol. Two paragraphs at the end of the PR also
summarised advice from the UK’s Chief Medical Officer (CMO),
which affirms this but also offers reassurance by stating that
those who have drunk before realising they are pregnant are
unlikely to have caused harm. The majority of media reports
(n=11) included direct or indirect advice to women, but few
included the CMO’s reassurance (n=5).
Summary: The PR described the study accurately with
appropriate language to convey uncertainty, but failed to include
alternative estimates for FASD prevalence and some key
caveats. The PR introduced both advice and reassurance to
women through quotes from the main author and others, includ-
ing the CMO. The media coverage also communicated the study’s
findings effectively, but often approximated the prevalence
estimate of 17% to “one in six” or “one in five”, and occasion-
ally presented it as certain, implying a level of confidence that
the study authors had not claimed. The majority of media
reports addressed the study’s limitations through the inclusion of
caveats and criticisms, including caveats that had been omit-
ted in the PR, and additional criticism sourced from the SMC
and BPAS.
Page 9 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
Illustrative example 2
Documents:
• Study: Golding, J., et al. (2019)
• Press release: University of Bristol (2019). Thinking
positive during pregnancy?
• Media reports (n=16)
Study summary:
Data from the ALSPAC longitudinal birth cohort study were
analysed to assess the association between maternal locus
of control (LOC), measured during pregnancy, and children’s
reasoning skills in maths and science. Maternal external
LOC was associated with poorer academic performance in
children. Three separate sets of factors were controlled for, to
assess the extent to which they mediated the association. Taken
together they identified at least 50% of the mechanism.
Comparison of study, PR and media reports:
Key changes at each stage of the risk reporting pathway are
summarised in Figure 8.
Description of the research and its findings: The study
design was reported accurately in the PR and all 16 media
reports. However, there were linguistic differences in the descrip-
tions of the exposure and outcome of interest. The exposure
discussed in the study was maternal LOC, which is defined as a
personality scale that “identifies individuals’ general attitude
to what happens to them as largely a matter of luck or fate or
of powerful others (externality) or whether they feel they can
influence the consequences (internality)”. However, the press
release presented this as “thinking positively during pregnancy”,
which is arguably inaccurate, since it implies optimism rather
than a sense of control. All 16 media reports mentioned both
LOC and “thinking positively”, but the latter was more promi-
nent, appearing in almost all the headlines (n=14). The PR
also stated that people with an external LOC are less “motivated
into action” or believe there is “little point in making an effort”,
which was echoed in 14 media reports. The temporal nature
of LOC is also presented misleadingly in the PR. In the study,
LOC is understood as a consistent personality characteristic,
which was measured during pregnancy as a baseline. Indeed,
its manifestation in parenting behaviours long after pregnancy
is a primary avenue of enquiry for the authors. Nonetheless,
the PR fails to explain this, and instead only discusses mater-
nal attitude during pregnancy, implying that the importance of
LOC is limited to those nine months, and that a mother’s
mindset during pregnancy can impact her children for years
to come regardless of her future behaviours.
The outcome of interest, maths and science reasoning, was
reported accurately in the PR and all 16 media reports, although
almost all reports used more generalised language such as
“bright”, “gifted” or “the next Einstein”. However, the PR intro-
duced additional outcomes which had not been included in the
study, including children’s health, confidence, independence,
sleep, emotional control, diet, school-related social difficulties,
and obesity. All of these additional outcomes were introduced
via quotes from two of the study’s authors, despite bearing no
clear relevance to the findings of this particular study. These
outcomes were mentioned in almost all (n=14) media reports.
Figure 8. Key changes along the risk reporting pathway (illustrative example 2).
Page 10 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
A primary objective of the study was to assess the extent to
which the association is mediated by factors including perinatal
exposures and parenting behaviours, which were found to explain
more than half of the mechanism. Some of these parenting
behaviours were mentioned in the PR, but their role as media-
tors of the association was not discussed. This was also absent
from all 16 media reports.
Caveats, criticism and study limitations: The PR does not
mention any of the study’s limitations or caveats, including the
differentiation of association and causation. However, the
majority of media reports (n=13) did reinstate this caveat.
Associated discussion topics: The primary discussion topic
in the media reports was past research on LOC. This theme was
present in both the PR and the study itself.
Advice, warnings and reassurance: The PR introduced
direct advice to the general public. Even the phrasing of its
headline and opening sentence (“Thinking positively during
pregnancy?”; “Your attitude during pregnancy…” etc.) is clearly
intended to address women directly. There is also discussion
in the form of quotes from the study’s authors about moth-
ers changing their outlook in order to improve their children’s
outcomes. Almost all (n=14) media reports included indirect
advice through discussion of mothers altering their LOC,
mainly through the inclusion of these quotes from the PR.
Summary: The press release altered the description of the
exposure of interest to “thinking positively during pregnancy”,
arguably misrepresenting the meaning of LOC and implying
that its importance is limited to the nine months of pregnancy.
Several child outcomes that were not explored in the study were
added to the PR via the inclusion of author quotes. Additionally,
the PR did not discuss the study’s limitations or state that
causation had not been established, and there was no discussion
in the PR of parenting as a mediating factor of the association.
The press coverage echoed the PR closely with two exceptions:
the generalisation of the outcome of interest (maths and science
reasoning) to “bright”, “gifted” etc., and the reinstatement of
the caveat that causation was not established in almost all
reports (n=13).
Illustrative example 3
Documents:
• Study: Gignac, F., et al. (2019)
• Press release: IS Global (2019). Maternal Nut
Consumption During Pregnancy Linked to Improvements in
Neurodevelopment in Children
• Media reports (n=17)
Study summary:
Data from the Spanish Childhood and Environment (Infancia
y Medio Ambiente, INMA) Project, a population-based
birth cohort study, were analysed to assess whether maternal
nut consumption during pregnancy is associated with child
neuropsychological outcomes. First-trimester nut consump-
tion was associated with improved child performance in four
cognitive tests. After adjusting for confounders, reduced hit
reaction time standard error in the Attention Network Test
(ANT), a measure of sustained attention, remained significant.
Third trimester nut intake showed a weaker association.
Comparison of study, PR and media reports:
Key changes at each stage of the risk reporting pathway are
summarised in Figure 9.
Figure 9. Key changes along the risk reporting pathway (illustrative example 3).
Page 11 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
Description of the research and its findings: The study’s
design and the exposure of interest (maternal nut intake) were
described accurately in the PR and media reports. All 17 media
reports mentioned maternal nut intake, with the majority
specifying the first trimester of pregnancy and giving details
of serving size and/or frequency of consumption (n=13 and
n=10 respectively).
For the outcome, similarly to the previous examples, almost
all the media reports (n=16) used simpler language, express-
ing neuropsychological development as “intelligent”, “smart” or
“brainy”, although most (n=12) also mentioned the specific fac-
ets measured, i.e. attention, memory, or cognitive function.
Three media reports introduced additional outcomes, the inci-
dence of learning difficulties and super-high IQs, which had not
been mentioned in either the study or the PR. These were
included in the form of a quote from one of the study’s authors,
which was not present in the PR. The same three media reports
also stated that child performance in cognitive tests had
improved by an average of 3%, a figure that is not given in
either the PR or the study.
Caveats, criticism and study limitations: The study was clear
that while improvements were observed in all cognitive tests,
only one (hit reaction time standard error in the ART) remained
significant after fully adjusting for confounders. However, this
is not specified in the PR, which describes in its by-line “bet-
ter outcomes after birth in cognitive function, attention capac-
ity and long-term working memory”. None of the media reports
explained that only the ART results were significant after
adjusting.
The PR omitted any discussion of the study’s limitations, but
it did state that the findings should be interpreted with caution.
Only three media reports included any caveats in their report-
ing of the study. One media report included the caveat that cau-
sation had not been established since this was an observational
study, which was not stated explicitly in either the study or PR
(although both acknowledged that further research is required
to confirm the findings).
Associated discussion topics: Four media reports discussed
maternal nutrition and the benefits of eating nuts more
generally. Both these topics were present in the study and PR.
Advice, warnings, and reassurance: Just under half of the
media reports (n=8) used language that addressed advice to
women directly, for example, “eat nuts for a brainy baby” or
“mothers told” to eat nuts. This direct language was not present
in the study or PR, but the PR did contain indirect advice
through discussion of the benefits and the ideal amount of nuts
to eat. One media report sourced an expert quote from the Royal
College of Obstetricians and Gynaecologists, which discussed
the existing guidance on nut consumption during pregnancy.
Summary: The PR generally reflected the study’s findings
accurately, but omitted caveats, which remained absent in the
majority of media reports. Media reports sometimes introduced
direct advice, and almost always generalised the outcome of
interest to “brainy”, “smart”, etc., but otherwise the majority
reported the study’s findings accurately. However, three media
reports introduced a new figure (3% average improvement)
alongside two new outcome measures (incidence of learning
difficulties and super-high IQs) which had not been present
in the study or PR. All three appeared to source the lat-
ter information from one of the study’s authors. It is unclear
whether the 3% figure was obtained from the same source.
Illustrative example 4
Documents:
• Study: Zhang, T., et al. (2019)
• Science Media Centre (2018). expert reaction to study
on caesareans and neurodevelopmental disorders
• Media reports (n=16)
NB no PR was published to accompany this study.
Study summary:
A systematic review and meta-analysis of 61 studies, com-
prising over 20 million births, found that caesarean birth was
significantly associated with both autistic spectrum disorder
(ASD) and attention-deficit/hyperactivity disorder (ADHD) in
children.
Comparison of study and media reports:
Key differences between the study and media reports are
summarised in Figure 10.
Description of the research and its findings: All 16 media
reports described the study accurately, reporting the exposure and
outcome of interest correctly. The majority of reports specified
the relative increased risk for ASD and ADHD (n=14 and n=13
respectively). One media report also specified the increase in
absolute risk for ASD, which was not mentioned in the study
but discussed in the SMC expert reaction summary.
Caveats, criticism and study limitations:
The majority of media reports (n=10) conveyed caveats
given in the study. Of the six that did not, five were passing
references to the study in articles about caesarean births
published several weeks after the study’s publication. Ten
reports stated explicitly that causation had not been estab-
lished, with one including this caveat in the headline. Half of the
reports (n=8) included the caveat that the mechanism by which
caesarean birth is associated with ASD and ADHD is not
understood, and eight mentioned confounding factors that
could explain the link. Five reports stated that Caesarean deliv-
ery is sometimes necessary to reduce other risks. Most reports
(n=10) also added expert criticism sourced from outside the
PR (mostly from the SMC summary).
Associated discussion topics:
Half of the media reports (n=8) discussed the rate of caesar-
ean births and seven discussed previous research on caesarean
births. Both topics of discussion were present in the study.
Page 12 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
Figure 10. Key changes along the risk reporting pathway (illustrative example 4).
Advice, warnings and reassurance:
No media reports advised that caesarean births should be
avoided, and most (n=14) omitted the recommendation from
the study that caesarean delivery should be used judiciously.
Nearly half of the media reports (n=7) included explicit reas-
surance from experts, sourced from the SMC, that women
should not be alarmed by the findings, with one report
including this in its headline.
Summary: All media reports described the findings accu-
rately, with the majority specifying the relative increase in risk
for both ASD and ADHD. Ten reports included caveats from the
study and also included expert commentary from other sources.
No warnings or advice were added by the media, and in fact the
study’s recommendation that caesarean delivery should be used
judiciously was mostly omitted. Instead, expert reassurance
to women was included in seven reports, including once in
the headline.
Discussion
We found that risk messages relating to pregnancy and fertility
are a common feature of UK print and online media reporting,
appearing in 43 headlines per month on average – more than
once per day. Each month there were, on average, 14 unique
stories based on a new scientific study, suggesting that roughly
every other day a novel scientific finding regarding pregnancy
or fertility risk is presented to the public. These messages
contribute to the context of pregnant women’s health-related
decisions. Since the majority of the science-based unique stories
we identified concerned lifestyle or health factors, particularly
food and drink, medication or underlying health conditions, their
cumulative effect may increase anxiety among the pregnant
population that any habit, lifestyle factor, or even emotion
they experience during pregnancy is now considered a risk
factor.
The majority of the unique stories focused on risk factors relat-
ing to the pregnant woman, which were framed in most cases
only in terms of their possible impact on fetal health, without
discussion of the health outcomes (either positive or nega-
tive) for the woman. The separation of maternal and child health
outcomes in this way positions pregnant women as vectors of
potential harm or benefit to their fetus, and frames their health
status, behaviours and mindset during pregnancy as determi-
nants of offspring wellbeing, both immediately and long into the
future. Previous authors have raised concerns about this
positioning of women as vectors, which has been used to jus-
tify increasing surveillance and scrutiny of pregnant women’s
choices without due consideration of their autonomy12–16.
PRs had a strong influence on media coverage, indicating that
academic institutions have a large degree of control over the
media’s reporting of their research. Our findings support
previous research that exaggeration in PRs is echoed in media
reports1,8. To a degree, journalists fulfilled a corrective role,
reinstating caveats omitted at the PR stage and sourcing addi-
tional expert criticism. However, certain misleading elements
introduced at the PR stage (such as the framing of external
LOC in example 2 as “thinking positively during pregnancy”)
remained present in the media coverage, and some caveats or
other pertinent contextualising information (such as the lower
FASD estimates in example 1, or the role of parenting behav-
iours in example 2) were not recovered at the media stage to
the risk reporting pathway. Since media reports inform the
public’s health behaviours, understanding and decisions4–6, these
inaccuracies matter. The portrayal of external LOC as “thinking
Page 13 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
positively during pregnancy”, for example, may conceiv-
ably cause women who have difficult pregnancies or antenatal
depression to feel that they have caused irreversible harm to their
child, and this is compounded in that example by the lack of
explanation in the PR and media reports that parental behaviour
is a mediating factor of the association. Efforts at the PR stage
to ensure that accuracy, caveats and context are preserved
may help to avoid such instances of needless anxiety.
PRs were not the only route by which information reached
journalists: several media reports included additional expert
commentary sourced from the SMC, other experts, or directly
from the study authors. In illustrative example 4, which had no
PR, we observed a noteworthy reliance on expert quotes sourced
from the SMC and elsewhere, suggesting that in the absence of a
PR journalists may choose not to add interpretation themselves,
preferring to rely on expert opinion. Conversely, in illustra-
tive example 1, the SMC materials – which were highly critical
of the study – were not frequently used. This may indicate
that time-pressed journalists are less compelled to consult
additional sources for interpretative expert commentary if a
comprehensive PR is available. If true, this would constitute
another way in which PRs exert influence on media report-
ing. Reduced journalist reliance on external expert commentary
is material because, in our examples, the heavy use of exter-
nal sources appeared to yield more accurate and reassuring
reporting. The media coverage in illustrative example 4 deviated
the least from the underlying study, and was particularly meas-
ured and reassuring, incorporating expert quotes that empha-
sised the study’s limitations and explicitly told women not to
worry. Notably, the lack of PR in illustrative example 4 did not
appear to hamper media coverage, and neither was coverage
hampered by the publication of critical SMC round-ups: exam-
ples 1 and 4 each generated 16 headlines, despite the cautionary
nature of the accompanying SMC commentaries.
To conclude, the PRs analysed here omitted caveats and
other pertinent information, introduced advice to women, and
occasionally included misleading elements. Greater reliance on
SMC (and other third-party) expert materials appears conducive
to reliable, measured reporting. We suggest that a new model of
science dissemination may be desirable, whereby journalists
are encouraged to consult expert opinion from a variety of
sources or make greater use of comment round-ups compiled by
independent third parties such as the SMC.
Strengths and limitations
Our study did not analyse broadcast media or social media, both
of which are important streams by which health information
is disseminated. Our four illustrative examples were selected
based on volume of coverage, which ensured a richness of
material to analyse, but also rendered them “outliers” among
our unique stories (which mostly only generated a small number
of headlines) and therefore possibly atypical. Lastly, this is an
observational analysis based on correlational relationships and
an assumption that the risk reporting pathway was followed.
It is possible that similarities between PRs and media reports
were not causal and occurred by coincidence.
Future research
The media landscape does not necessarily reflect the nature
of research undertaken in the field, since not all studies are
reported. Future research could analyse risk messages dissemi-
nated by journals or research institutions to examine whether the
fetus-centric framing observed in the media is also present ear-
lier on the risk reporting pathway. It would also be interesting
to repeat this analysis for risk headlines relating to infant care,
including breastfeeding. On the role of PRs, further research
could compare media coverage of scientific papers published
with or without a PR, to assess whether PRs alter the volume or
nature of coverage. The use of SMC commentary by journal-
ists could also be examined to see whether its uptake is reduced
by the availability of a comprehensive PR, and whether its use
generally leads to more balanced and/or reassuring reporting.
Data availability
Figshare. Risk messages relating to pregnancy and fertility:
a media content analysis (dataset). https://doi.org/10.6084/
m9.figshare.14480784.v122
Data are available under the terms of the Creative Commons
Zero “No rights reserved” data waiver (CC0 1.0 Public domain
dedication).
Acknowledgements
Special thanks are due to Dr Heather Trickey of Cardiff
University for her contributions and counsel. We also thank
members of the WRISK Oversight Committee for their sup-
port, and particularly the committee’s chair, Dame Cathy
Warwick.
References
1. SumnerP,Vivian-GrithsS,BoivinJ,et al.: The association between
exaggeration in health related science news and academic press releases:
retrospective observational study. BMJ. 2014; 349: g7015.
PubMed Abstract | Publisher Full Text | Free Full Text
2. RieschH,SpiegelhalterDJ:‘Careless pork costs lives’: Risk stories from
science to press release to media. Heal Risk Soc. 2011; 13(1): 47–64.
Publisher Full Text
3. TullochJC,ZinnJO:Risk, health and the media. Heal Risk Soc. 2011; 13(1): 1–16.
Publisher Full Text
4. AdamsRC,SumnerP,Vivian-GrithsS,et al.: How readers understand causal
and correlational expressions used in news headlines. J Exp Psychol Appl.
2017; 23(1): 1–14.
PubMed Abstract | Publisher Full Text
5. GrilliR,RamsayC,MinozziS:Mass media interventions: eects on health
Page 14 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
services utilisation. Cochrane Database Syst Rev.2002;(1):CD000389.
PubMed Abstract | Publisher Full Text
6. ThirlawayKJ,HeggsDA:Interpreting risk messages: Women’s responses to a
health story. Heal Risk Soc. 2005; 7(2): 107–121.
Publisher Full Text
7. LeeE,SuttonRM,HartleyBL:From scientic article to press release to
media coverage: advocating alcohol abstinence and democratising risk in
a story about alcohol and pregnancy. Heal Risk Soc. 2016; 18(5–6):
247–269.
PubMed Abstract | Publisher Full Text | Free Full Text
8. BrattonL,AdamsRC,ChallengerA,et al.: Causal claims about correlations
reduced in press releases following academic study of health news
[version 1; peer review: 1 approved, 1 approved with reservations].
Wellcome Open Res. 2020; 5: 6.
Publisher Full Text
9. SumnerP,Vivian-GrithsS,BoivinJ,et al.: Exaggerations and Caveats in
Press Releases and Health-Related Science News. PLoS One. 2016; 11(12):
e0168217.
PubMed Abstract | Publisher Full Text | Free Full Text
10. AdamsRC,ChallengerA,BrattonL,et al.: Claims of causality in health news:
a randomised trial. BMC Med.2019;17(1):91.
PubMed Abstract | Publisher Full Text | Free Full Text
11. BottL,BrattonL,DiaconuB,et al.: Caveats in science-based news stories
communicate caution without lowering interest. J Exp Psychol Appl.2019;
25(4): 517–542.
PubMed Abstract | Publisher Full Text
12. Lupton D: ‘Precious cargo’: Foetal subjects, risk and reproductive
citizenship. Crit Public Health. 2012; 22(3):329–340.
Publisher Full Text
13. Ruhl L: Liberal governance and prenatal care: Risk and regulation in
pregnancy. Econ Soc.1999;28(1):95–117.
Publisher Full Text
14. Lowe P, Lee E, Yardley L: Under the Inuence? The Construction of Foetal
Alcohol Syndrome in UK Newspapers. Sociol Res Online. 2010; 15(4):75–84.
Publisher Full Text
15. LoweP,LeeE,MacvarishJ:Growing better brains? Pregnancy and
neuroscience discourses in English social and welfare policies. Health Risk
Soc. 2015; 17(1):15–29.
PubMed Abstract | Publisher Full Text | Free Full Text
16. Lowe P: Idealized Pregnancy. In: Reproductive Health and Maternal Sacrice.
(PalgraveMacmillanUK).2016;109–137.
Publisher Full Text
17. Timmermans S, Leiter V: The Redemption of Thalidomide: Standardizing the
Risk of Birth Defects. Soc Stud Sci. 2000; 30(1): 41–71.
PubMed Abstract | Publisher Full Text
18. RichardsonSS,DanielsCR,GillmanMW,et al.: Society: Don’t blame the
mothers. Nature. 2014; 512(7513): 131–132.
PubMed Abstract | Publisher Full Text
19. SharpGC,LawlorDA,RichardsonSS:It’s the mother!: How assumptions
about the causal primacy of maternal eects inuence research on the
developmental origins of health and disease. Soc Sci Med.2018;213: 20–27.
PubMed Abstract | Publisher Full Text | Free Full Text
20. HaugenAC,SchugTT,CollmanG,et al.: Evolution of DOHaD: the impact of
environmental health sciences. J Dev Orig Health Dis. 2015; 6(2): 55–64.
PubMed Abstract | Publisher Full Text | Free Full Text
21. Wriskproject:Risk reporting pathway. 2020.
Reference Source
22. MarshallO,BlaylockR,MurphyC,et al.: Risk messages relating to pregnancy
and fertility: a media content analysis (dataset). 2021.
http://www.doi.org/10.6084/m9.gshare.14480784.v1
Page 15 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
Open Peer Review
Current Peer Review Status:
Version 1
Reviewer Report11 June 2021
https://doi.org/10.21956/wellcomeopenres.18465.r43901
© 2021 MacKendrick N. This is an open access peer review report distributed under the terms of the Creative
Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium,
provided the original work is properly cited.
Norah MacKendrick
Department of Sociology, Rutgers University, New Brunswick, NJ, USA
This is an important study that examines how the news media reports scientific research related
to pregnancy and fertility. The research objectives are clear and the methodologies are
appropriate. The manuscript would benefit from revisions, namely to fill in missing details across
several sections of the paper.
Introduction
The manuscript has two general objectives: assess i) risk framing in PRs and news media reports
and, ii) the accuracy of news reporting of scientific studies related to fertility and pregnancy. The
introduction could state more clearly these two general objectives. They appear only at the end of
the introductory section (p. 3).
In this section, the authors refer to the literature on the framing of risk messages in PRs and the
impact of this framing on the public’s understanding of health risks and attribution of blame.
These are important points that the authors re-visit in the discussion. The summary of this
literature is very short and could be expanded upon. On page 3, for example, the authors cite
several key articles (e.g. Reisch and Spiegelhalter; Tulloch and Zinn) but say little about the larger
findings from this literature. This is important context that helps to set the stage for the study and
the discussion section.
Methods
The methods section is well organized and the methods used are appropriate. I appreciate the use
of figures to show the reader the various stages of data collection and analysis. The authors could
say more about why certain kinds of news articles were important and others were excluded. This
section should also explain the rationale for the time period of 2018-19.
Analysis
The analysis is detailed and the use of figures is helpful. The specific approach for evaluating
accuracy remains vague. On pages 3-4, the authors refer to risk reporting pathways as a measure
of reliability. If this is the method of measuring accuracy, the authors could tell us more about it.
Page 16 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
Figures 5 and 6 capture some elements of framing, as well as article content, but these figures
could be explained in more detail in the text.
Discussion
This section summarizes the main findings and their implications. The role of PRs in shaping
media reports is an important finding. The authors could say even more about what they learned
about the framing of pregnancy and fertility risks. Were risk messages in PRs the same as those in
the media reports? What do we learn about risk framing in cases where the framing between PR
and article diverged?
Finally, the authors could revise the manuscript to add more text to the paper to guide the reader
through each section of the manuscript. There are few transition and introductory sentences in
each section, which gives the paper a “choppy” flow. Adding some of these would help the reader
follow the paper from section to section.
I appreciated the opportunity to review this paper and hope to cite it in my own work.
Is the work clearly and accurately presented and does it cite the current literature?
Yes
Is the study design appropriate and is the work technically sound?
Yes
Are sufficient details of methods and analysis provided to allow replication by others?
Partly
If applicable, is the statistical analysis and its interpretation appropriate?
Yes
Are all the source data underlying the results available to ensure full reproducibility?
Yes
Are the conclusions drawn adequately supported by the results?
Yes
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: media framing, gender and risk, sociology of health, sociology of
environmental risk
I confirm that I have read this submission and believe that I have an appropriate level of
expertise to confirm that it is of an acceptable scientific standard.
Page 17 of 17
Wellcome Open Research 2021, 6:114 Last updated: 11 JUN 2021
