RESEARCH ARTICLE

Effect of Expressive Writing Intervention on

Health Outcomes in Breast Cancer Patients:
A Systematic Review and Meta-Analysis of
Randomized Controlled Trials
Chunlan Zhou1*, Yanni Wu1*, Shengli An2, Xiaojin Li1
1 Nanfang Hospital, Southern Medical University, Guangzhou, Guangdong, PR China, 2 Department of Bio-
Statistics, Guangdong Provincial Key Laboratory of Tropical Disease Research, School of Public Health and
Tropical Medicine, Southern Medical University, Guangzhou, Guangdong, PR China

These authors contributed equally to this work.

* lanchun200488@126.com (CLZ); 379249275@qq.com (YNW)

Abstract
OPEN ACCESS

Citation: Zhou C, Wu Y, An S, Li X (2015) Effect of Background

Expressive Writing Intervention on Health Outcomes
Numerous randomized controlled trials (RCTs) have arrived at conflicting conclusions on
in Breast Cancer Patients: A Systematic Review and
Meta-Analysis of Randomized Controlled Trials. expressive writing (EW) as an intervention for breast cancer (BC) patients, but there has
PLoS ONE 10(7): e0131802. doi:10.1371/journal. been no meta-analysis of these studies to assess the effectiveness of EW in BC population.
pone.0131802

Editor: Gozde Ozakinci, University of St Andrews, Methods

UNITED KINGDOM
PubMed, Web of Science, The Cochrane Library, EMBASE, and CINAHL and the www.
Received: September 12, 2014
clinicaltrial.gov database on ongoing clinical trials were searched to identify all the RCTs
Accepted: June 5, 2015 investigating efficacy of EW on the physical and psychological health in BC patients. The
Published: July 7, 2015 risk of bias of the original studies was assessed using the Cochrane Collaborations tool.
Copyright: 2015 Zhou et al. This is an open Our primary outcomes for physical and psychological health were respectively negative
access article distributed under the terms of the somatic symptoms and negative mood which were stratified by emotional, benefit-finding
Creative Commons Attribution License, which permits and multiple prompts in sub-group analyses. The data were analyzed using Review Man-
unrestricted use, distribution, and reproduction in any
ager 5.2 and Stata version 12.0 statistical software.
medium, provided the original author and source are
credited.

Data Availability Statement: All relevant data are Results

within the paper and its Supporting Information files. Of the 5,232 titles screened, we identified 11 RCTs with a total of 1,178 participants. The
Funding: The research was supported by the pooled results showed a significant effect of EW using either an emotional prompt or a ben-
Guangdong Provincial Department of Science and efit-finding prompt on reducing negative somatic symptoms in BC patients in the 3-month
Technology. The grant number is 2012B031800124.
follow-up group [Mean Difference (MD), -13.03, 95% CI, -19.23 to -6.83, P<0.0001; MD,
The funder had no role in study design, data
collection and analysis, decision to publish, or -9.18, 95% CI, -15.57 to -2.79, P = 0.005]. There was no significant effect of EW on physical
preparation of the manuscript. health in the >3-month follow-up group. There were no significant differences regarding psy-
Competing Interests: The authors have declared chological health indexes between EW intervention and control groups at any of the follow-
that no competing interests exist. up time-points (P>0.05).

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 1 / 19

 EW Intervention in Breast Cancer Patients

Conclusion
This systematic review and meta-analysis reveals that EW intervention may have a signifi-
cantly positive impact on the physical health but not the psychological health in BC patients,
but this benefit may not last long. However, further high-quality studies with more homoge-
neity are needed to confirm the current findings.

Introduction
Emotional expression, as a psychological or medical intervention, has been studied for many
years, demonstrating favorable impacts on physical and mental health [17]. Expressive writ-
ing (EW) as a form of emotional expression was first implemented in college students in 1986
by Pennebaker and Beall, who instructed respondents to write about their deepest emotions
and thoughts regarding traumatic/upsetting experiences for approximately 20 minutes over
four consecutive days [8]. Afterwards, work on EW as a potential intervention for physical and
psychosocial adjustment was extended to clinical and medical populations, including non-
patients [912] as well as patients with rheumatoid arthritis [13,14], asthma [15], HIV [16],
cardiovascular disease [17] or renal cell carcinoma [18]. Since Walker explored the feasibility
of using EW in a breast cancer (BC) cohort [19], there have been numerous randomized con-
trolled trials (RCTs) testing effectiveness of EW in BC patients [2029]. These studies have
been performed for reasons like: BC remains the second most frequently diagnosed type of can-
cer in women [30], many BC patients report feeling emotionally inhibited which has been
linked with worse psychological functioning [1,31,32], and physical problems are still highly
prevalent in BC population [3338].
Although EW was generally considered beneficial, some researchers questioned its utility in
light of failures to replicate the original findings [15,39]. Some studies reached a negative con-
clusion on the benefits of EW [4044] while others demonstrated positive effects of EW on the
physical and psychological health in various populations, mostly patients [13,16,4549]. Sys-
tematic reviews of the studies on the efficacy of EW in healthy and unhealthy populations also
led to various conclusions [5055]. Diversified concerns, inclusion of different populations,
measurement of different variables and different methodology may have been significant rea-
sons for their inconsistent findings. Harris and Mogk concluded that EW had little effect on
the subjects tested [53,54], Frisina and Frattaroli found EW was effective [50,52], but Boinon
and Merz could not make a definite conclusion about EW in cancer patients [51,55]. Similarly,
studies on EW in BC patients also arrived at conflicting conclusions. Some studies failed to
confirm the benefit of EW [19,22,23,26], but others found positive effects of EW on the physi-
cal or psychological health of this particular population [20,21,24,25,2729].
Our concern focuses on the benefits of EW on the physical and psychological health of BC
population. Firstly, EW is a low-cost, convenient and self-administered intervention that can
be routinely used in clinic if its therapeutic benefits can be confirmed. Secondly, BC victims are
overwhelmingly females who have to brave particular physical and psychological challenges
that may impact their therapeutic outcomes after diagnosis and treatment of BC [5663].
There is a great demand to develop all kinds of interventions which may help them cope with
their specific physical and psychological challenges on their way to combat BC. Moreover,
yearly increase of more than 1.3 million new cases has made BC the most frequently diagnosed
cancer in women worldwide [64]. However, numerous studies on EW in BC patients have pro-
vided inconsistent findings and there has been no meta-analysis of these studies so far.

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 2 / 19

 EW Intervention in Breast Cancer Patients

Conclusion
This systematic review and meta-analysis reveals that EW intervention may have a signifi-
cantly positive impact on the physical health but not the psychological health in BC patients,
but this benefit may not last long. However, further high-quality studies with more homoge-
neity are needed to confirm the current findings.

Introduction
Emotional expression, as a psychological or medical intervention, has been studied for many
years, demonstrating favorable impacts on physical and mental health [17]. Expressive writ-
ing (EW) as a form of emotional expression was first implemented in college students in 1986
by Pennebaker and Beall, who instructed respondents to write about their deepest emotions
and thoughts regarding traumatic/upsetting experiences for approximately 20 minutes over
four consecutive days [8]. Afterwards, work on EW as a potential intervention for physical and
psychosocial adjustment was extended to clinical and medical populations, including non-
patients [912] as well as patients with rheumatoid arthritis [13,14], asthma [15], HIV [16],
cardiovascular disease [17] or renal cell carcinoma [18]. Since Walker explored the feasibility
of using EW in a breast cancer (BC) cohort [19], there have been numerous randomized con-
trolled trials (RCTs) testing effectiveness of EW in BC patients [2029]. These studies have
been performed for reasons like: BC remains the second most frequently diagnosed type of can-
cer in women [30], many BC patients report feeling emotionally inhibited which has been
linked with worse psychological functioning [1,31,32], and physical problems are still highly
prevalent in BC population [3338].
Although EW was generally considered beneficial, some researchers questioned its utility in
light of failures to replicate the original findings [15,39]. Some studies reached a negative con-
clusion on the benefits of EW [4044] while others demonstrated positive effects of EW on the
physical and psychological health in various populations, mostly patients [13,16,4549]. Sys-
tematic reviews of the studies on the efficacy of EW in healthy and unhealthy populations also
led to various conclusions [5055]. Diversified concerns, inclusion of different populations,
measurement of different variables and different methodology may have been significant rea-
sons for their inconsistent findings. Harris and Mogk concluded that EW had little effect on
the subjects tested [53,54], Frisina and Frattaroli found EW was effective [50,52], but Boinon
and Merz could not make a definite conclusion about EW in cancer patients [51,55]. Similarly,
studies on EW in BC patients also arrived at conflicting conclusions. Some studies failed to
confirm the benefit of EW [19,22,23,26], but others found positive effects of EW on the physi-
cal or psychological health of this particular population [20,21,24,25,2729].
Our concern focuses on the benefits of EW on the physical and psychological health of BC
population. Firstly, EW is a low-cost, convenient and self-administered intervention that can
be routinely used in clinic if its therapeutic benefits can be confirmed. Secondly, BC victims are
overwhelmingly females who have to brave particular physical and psychological challenges
that may impact their therapeutic outcomes after diagnosis and treatment of BC [5663].
There is a great demand to develop all kinds of interventions which may help them cope with
their specific physical and psychological challenges on their way to combat BC. Moreover,
yearly increase of more than 1.3 million new cases has made BC the most frequently diagnosed
cancer in women worldwide [64]. However, numerous studies on EW in BC patients have pro-
vided inconsistent findings and there has been no meta-analysis of these studies so far.

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 2 / 19

 EW Intervention in Breast Cancer Patients

According to the literature published, although Boinon and Merz both conducted a systematic
review [51,55] of the effectiveness of EW in cancer patients, they did not focus their concern
specifically on BC population and did not perform meta-analysis. Therefore, it is necessary to
perform a meta-analysis to review all the data from all the high-quality studies available on this
topic to make convincing up-to-date conclusions about EW in BC population. This study
aimed to test the hypothesis that EW might be a promising clinical intervention to improve the
physical and psychological health in BC patients by determining whether EW was beneficial
for BC patients, what were the benefits, and how effective was EW in relieving physical and
psychological symptoms.

Methods
Search methods
We searched the following databases to identify relevant studies for this meta-analysis and
adapted different search strategies according to the query requirements of the individual data-
bases. We limited our search by the time after the year 1986 when the first EW study using
Pennebakers prompt was published. We did not restrict our search by language. The following
databases were queried: PubMed (from 1986 to June 2014), Web of Science (from 1986 to June
2014), The Cochrane Library (from 1986 to June 2014), EMBASE (from 1986 to June 2014),
and CINAHL (from 1986 to June 2014) (S1 Appendix). We searched the database of ongoing
trials, www.clinicaltrial.gov. We also screened the references of included studies to identify
additional articles. We did not handsearch journals or conference proceedings, due to limited
time and resources.

Inclusion criteria
To ensure homogeneity across studies, we included studies that met the following criteria: (1) a
randomized controlled trial (RCT) with an experimental design that included an EW group
(expressive writing for at least a single 20-minute session using the Pennebaker and Beall para-
digm [8]) and a control group for comparison; (2) women participants with a BC diagnosis,
irrespective of their age, BC stage, treatment modality or treatment setting (including inpatient,
outpatient and primary care); (3) outcome measures that assessed factors relative to the physi-
cal and psychological health of BC patients. We excluded review articles or studies the com-
plete data of which were unavailable.

Assessment of methodological quality

The quality of included studies was assessed using the Cochrane Collaborations risk of bias
tool. Each study was assessed for random sequence generation (selection bias), allocation con-
cealment (selection bias), blinding of participants and personnel (performance bias), blinding
of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective
reporting (reporting bias), and any other potential sources of bias. Each factor would be rated
as low risk of bias (e.g., random sequence generation was computer generated), high risk of
bias (e.g., outcome assessment was not blinded) or unclear risk of bias (e.g., did not provide
specific information as to whether allocation concealment was used). Disagreements would be
resolved by consensus.

Data extraction
Two authors (YNW and XJL) independently extracted the data from each trial using a stan-
dardized data extraction form that included general information (author, title, source, contact

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 3 / 19

 EW Intervention in Breast Cancer Patients

address, and year of publication), the trial characteristics (randomization method, blinding,
duration of intervention period, length of follow-up, and method for handling missing data),
the patient characteristics (sample size, stage of disease, race, age, level of education, average
time since diagnosis, and inclusion criteria), the intervention (detailed description of the con-
trolled intervention, mode, and duration) and outcomes (outcome measures and scoring
range). When data were missing, one author (YNW) contacted the authors to request addi-
tional information. If further information could not be obtained, we coded the variables in
question as NR.

Data analysis
We used Review Manager 5.2 (Cochrane Collaboration, Oxford, UK) and Stata version 12.0
(Stata Corp, College Station, Texas, USA) for data analysis. Two investigators (YNW and SLA)
were involved in the statistical analysis. Measurement of outcomes was considered in terms of
original data at each follow-up time point (baseline scores not included). The mean difference
(MD) and 95% CI were calculated based on fixed-effect model for continuous variables. The z-
test was used to obtain the combined P-values of the included studies with a significance level
of P = 0.05. The statistical significance of heterogeneity among studies was assessed by calculat-
ing the chi-square test (a P-value of 0.10 was regarded as statistically significant). The I2 was
used to quantify the effects of heterogeneity. If statistical heterogeneity (P value 0.10 and
I250%) was identified, random effects meta-analysis was conducted before the causes of het-
erogeneity was further investigated by subgroup analysis; if not, a fixed-effects model was used
[65]. Eggers test and funnel plot were conducted to investigate the potential publication bias
influencing the analysis. To determine whether significant differences would exist between spe-
cific variables regarding the effectiveness of EW, subgroup analyses were carried out by sorting
the same specific variables reported in the RCTs.

Results
Description of the studies
Search results. We conducted the electronic searches in June 2014. A total of 5232 titles
and abstracts were screened, and 1622 duplicates were identified. Of the 3610 screened titles
and abstracts, 3587 were excluded. After we read the remaining 23 full-text articles, 12 full-text
articles were excluded [6677] and 11 studies included. No additional studies were identified
by searching the reference lists. There were no ongoing studies that we were aware of. The
study flow diagram is illustrated in Fig 1.
Included studies. Eleven studies examining the effectiveness of EW as an intervention on
the health outcomes in BC patients were included in this meta-analysis [1929]. Descriptions
of the studies and samples are presented in Table 1.
In all the 11 RCTs, the participants had been randomly divided into an EW group and a
control group. The intervention methods in the EW groups included EW with an emotional
prompt (cancer or any trauma), a benefit-finding prompt (cancer) or multiple prompts (emo-
tional, benefit-finding, cognitive appraisal and coping strategies), and those in the control
groups included neutral writing or no writing (usual care or non-cancer attention).
In total, 1178 BC patients were involved, with 613 in the EW intervention groups and 565
in the control groups. The BC stages ranged from 0 to 4. Six studies reported racial distribution
of the participants, with the Caucasian race accounting for more than 60% in three studies and
greater than 90% in the other three. Age distribution was provided in 10 studies, with an aver-
age age of approximately 50 years. Educational background of the subjects was reported in
seven studies, in six of which from 40% to 87% of the subjects had a university or higher degree

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 4 / 19

 EW Intervention in Breast Cancer Patients

Fig 1. Study flow diagram.

doi:10.1371/journal.pone.0131802.g001

and in one of which only the average years of education (15.202.48) were provided. Five stud-
ies reported the time between diagnosis of BC and enrollment of the subjects was between 28.4
weeks and 1.9 years; one study included subjects who were enrolled 4 years after a stage-4 BC
diagnosis; one study involved subjects enrolled 15155 days after operation; the remaining
four studies provided no specific time.
The EW intervention in the 11 studies was based on the Pennebaker and Beall [8] paradigm.
The number of EW sessions ranged from 16 with a minimum of 20 continuous minutes of
writing per session. Overall, most of the 11 studies had more than 3 writing sessions. The fol-
low-up time ranged from 1 week to 9 months after the EW intervention, with 3-month follow-
up implemented in six studies.
The effect of EW intervention on physical health was evaluated using 11 indexes by eight
studies [2022,2426,28,29], three of which measured negative somatic symptoms using the
negative somatic symptoms scale [78] (Table 1). The effect of EW on psychological health was
evaluated using 18 indexes by all the 11 RCTs [1929], five of which observed negative mood
using Profile of Mood States (POMS). Therefore, this meta-analysis decided to take negative
somatic symptoms as the primary outcome for physical health and negative mood as the

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 5 / 19

 EW Intervention in Breast Cancer Patients

primary outcome for psychological health. The remaining indexes were regarded as the sec-
ondary outcomes for either physical health or psychological health, respectively.
Risk of bias in the included studies. The risk of bias in the 11 studies was assessed using
the Cochrane Collaborations tool. The results are summarized in Fig 2.

Table 1. Characteristics of the included studies

Reference Sample Breast Race Mean College Average time Task Follow-up Outcomes
(year) cancer age educated since diagnosis
stages (years)

Craft et al. 97 03 92.8% 5610.5 NR 13 months EW: EMO (n = 26), 1 and 6 Aspects of psychological
(2013) Caucasian EMO+ (n = 19); Control: months health measured using: 1.
NWc (n = 22), USUAL Functional Assessment of
(n = 30); Four sessions, Cancer Therapy-Breast
consecutive days Cancer Version (FACT-B)
De Moor 49 23 63% 53.5 78.5% NR (had nished EW: EMO (n = 24); 3 days pre- Aspects of physical health
et al. (2008) Caucasian 10.4 neoadjuvant Control: NWb (n = 25); surgery measured using: 1. Brief
chemotherapy) Four sessions over 7 and 2 Pain Inventory (BPI), 2.
days, 3 weeks prior to weeks Pittsburgh Sleep Quality
surgery post- Index (PSQI); Aspects of
surgery psychological health
measured using: 1. Brief
Symptom Inventory 18 (BSI-
18), 2. Perceived Stress
Scale (PSS)
Gellaitry 80 12 NR 57.9 NR NR (had received EW: MULTIPLE (n = 38); 1, 3, and 6 Aspects of physical health
et al. (2010) 9.9 radiotherapy Control: USUAL months measured using: 1.
treatment) (n = 42); Four sessions, Healthcare utilization;
consecutive days Aspects of psychological
(different prompt each health measured using: 1.
day) Prole of Mood States
(POMS), 2. Functional
Assessment of Cancer
Therapy-Breast Cancer
Version (FACT-B)
Henry et al. 80 14 NR 58.9 NR 18 months EW: BEN (n = 40); 3 and 9 Aspects of physical health
(2010) NR Control: USUAL months measured using: 1. Physical
(n = 40); One session health measure (included 18
physical symptoms);
Aspects of psychological
health measured using: 1.
Center for Epidemiologic
Studies-Depression Scale
(CES-D), 2. Prole of Mood
States (POMS)
Jensen- 507 12 NR 53.6 40% NR (days since EW: EMO+(n = 253); 3 and 9 Aspects of psychological
Johansen 9.1 surgery 15155) Control: NW(n = 254); months health measured using: 1.
et al. (2012) Three sessions over 3 Impact of Events Scale
weeks (IES), 2. Beck Depression
Inventory-Short Form
(BDI-SF), 3. Prole of Mood
States (POMS), 4. Passive
Positive Mood Scale(PPMS)
Low et al. 62 4 87% 53.8 74% 7.9 years EW: EMO (n = 31); 3 months Aspects of physical health
(2010) Caucasian 10.3 Control: NWc (n = 31); measured using: 1. Negative
Four sessions over 3 somatic symptoms scale
weeks (developed by Pennebaker,
includes 9 somatic
symptoms), 2. Pittsburgh
Sleep Quality Index (PSQI);
Aspects of psychological
health measured using: 1.
Center for Epidemiologic
Studies-Depression Scale
(CES-D), 2. Impact of
Events Scale (IES)
(Continued)

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 6 / 19

 EW Intervention in Breast Cancer Patients

Table 1. (Continued)

Reference Sample Breast Race Mean College Average time Task Follow-up Outcomes
(year) cancer age educated since diagnosis
stages (years)

Low et al. 60 12 NR NR NR NR (had EW: EMO (n = 21), BEN 3 months Aspects of physical health
(2006) completed primary (n = 21); Control: NWc measured using: 1. Negative
medical (n = 18); Four sessions somatic symptoms scale
treatments) over 3 weeks (developed by Pennebaker,
includes 9 somatic
symptoms), 2. Medical
appointments for cancer-
related morbidities; Aspects
of psychological health
measured using: 1. Prole of
Mood States (POMS)
Mosher et al. 86 4 81% 57.9 87% Average time since EW: EMO (n = 44); 8 weeks Aspects of physical health
(2012) Caucasian 12.1 diagnosis of stage Control: NW (n = 42); measured using: 1. The
4 breast cancer: 4 Four sessions over 47 Functional Assessment of
years weeks Chronic Illness Therapy
Fatigue subscale (FACIT-F),
2. Pittsburgh Sleep Quality
Index (PSQI); Aspects of
psychological health
measured using: 1. The
Functional Assessment of
Chronic Illness Therapy
Spiritual Well-being scale
(FACIT-Sp), 2. Distress
Thermometer (DT), 3.
Center for Epidemiologic
Studies-Depression Scale
(CES-D), 4. Anxiety:
Hospital Anxiety and
Depression Scale (HADS-A)
Park et al. 58 23 NR 48.2 57% 23 months EW: EMO (n = 29); 4 weeks Aspects of physical health
(2012) 7.5 Control: USUAL measured using: 1.
(n = 29); Six sessions Pennebakers Inventory of
over 6 weeks Limbic Languidness (PILL),
2. M. D. Anderson Symptom
Inventory (MDASI); Aspects
of psychological health
measured using: 1. Hospital
Anxiety and Depression
Scale (HADS), 2. Cancer-
Quality of Life (C-QOL)
Stanton 60 12 93% 49.5 NR (average 28.4 weeks EW: EMO (n = 21), BEN 1 and 3 Aspects of physical health
et al. (2002) Caucasian 12.2 education (n = 21); Control: NWc months measured using: 1. Negative
level: 15.20 (n = 18); Four sessions somatic symptoms scale
2.48 years) over 3 weeks (developed by Pennebaker,
includes 9 somatic
symptoms), 2. Medical
appointments for cancer-
related morbidities; Aspects
of psychological health
measured using: 1.
Functional Assessment of
Cancer Therapy (FACT), 2.
Prole of Moods State
(POMS)
Walker et al. 39 12 95% 53.6 79% NR (were EW: 1 session EMO 1, 46, 16, Aspects of psychological
(1999) Caucasian completing RT for (n = 11), 3 sessions and 28 health measured using: 1.
stage 1 or 2 breast EMO (n = 14); Control: weeks Positive and Negative Affect
cancer) ATT (n = 14); One or Scale (PANAS), 2. Impact of
three sessions over 14 Events Scale (IES)
days

EMO, emotional (cancer) prompt; EMO+, emotional (any trauma) prompt; BEN, benet-nding (cancer) prompt; MULTIPLE, emotional, benet-nding,
cognitive-appraisal, and coping prompts; NW, neutral writing (trivial); NWb, neutral writing (health behavior); NWc, neutral writing (cancer); USUAL, usual
care; ATT, non-cancer attention; NR, not reported. The above format is cited in Merz [51].

doi:10.1371/journal.pone.0131802.t001

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 7 / 19

 EW Intervention in Breast Cancer Patients

The selection bias was considered low in six studies because a random serial grouping pro-
gram was used. Three of them were rated as completely low-risk because they employed a com-
puter-aided serial grouping program [22,25,28]; the random programs in the other three
studies were limited in most occasions, thus enhancing the predictability of grouping for the
researchers [21,23,26]. Three studies were considered as high-risk of bias because sequential
assignment [19,27] or matched controls [24] was utilized. The remaining two studies that did
not describe details of their randomization were rated as unclear risk of bias [20,29].
Three RCTs were identified as low selection bias because they used sequentially numbered
envelopes to conceal allocation of participants from the researcher or research assistants
[22,28,29]. Three RCTs were identified as high selection bias because they did not use condi-
tion allocation concealment of participants [19,24,27]. The remaining five studies did not pro-
vide specific information as to whether allocation concealment was used.
In the EW intervention, it was difficult to perform a fully blinded study because each
researcher and subject had to be informed of the writing requirements and methods. Therefore,
the bias risk was high in the 11 studies regarding the blinding of the participants and
personnel.
As for blinding of outcome assessment, three studies were considered low-risk because they
made efforts to mask the research purposes and writing tasks from research assistants and par-
ticipants [23,27,28]. One study was considered high-risk because the nurse in charge of the
baseline interview and assessment did not use a blinded method for allocation [24]. It was
unclear whether a blinded method was implemented in the outcome assessment for the
remaining seven studies.
Ten studies that stated the reason why and the time when a subject dropped out exhibited
low risk of attrition bias. Notably, Craft [27] used an intent-to-treat analysis to reduce the pos-
sibility of data bias. The remaining one study was regarded as unclear risk because no reason
was provided for a subject dropping out of the study [23].
All the 11 studies appeared to have reported on all measured outcomes, showing low risk of
reporting bias. No other potential sources of bias were noted in the included studies.

Effects of the interventions

To determine how long the effect of EW intervention would last in BC patients, we divided the
pooled outcome measures of physical health and of psychological health respectively into a
3-month follow-up group and a >3-month follow-up group. In studies that provided out-
come measures of more than one follow-up that met the criteria of grouping, the follow-up
data at one time point that satisfied the grouping criteria the most were included while the
other data were excluded. Specifically, the data of 3 days pre-surgery [26], 1 week [19] and 1
month [25,28] were excluded from the 3-month group while the follow-up results of 16
weeks [19] were excluded from the >3-month follow-up group.
Furthermore, to determine whether different means of EW intervention might have an
impact on the effectiveness of EW intervention we stratified the negative somatic symptoms
for physical health and negative mood for psychological health, which were primary outcomes
for our meta-analysis, by emotional prompt, benefit-finding prompt and multiple prompts.
However, since only one study [25] utilized multiple prompts but did not evaluate negative
somatic symptoms, actually it was only feasible for us to stratify negative somatic symptoms by
emotional prompt and benefit-finding prompt (Fig 3).
Expressive writing intervention and physical health. As shown in Table 2, seven studies
in the 3-month follow-up group employed a fixed-effect model to analyze the combined
effect of EW on the physical health indexes. The results showed a significant effect of EW using

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 8 / 19

 EW Intervention in Breast Cancer Patients

Fig 2. Risk of bias for the included studies.

doi:10.1371/journal.pone.0131802.g002

either an emotional prompt or a benefit-finding prompt on reducing negative somatic symp-

toms in BC patients in the 3-month follow-up group [Mean Difference (MD), -13.03, 95%

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 9 / 19

 EW Intervention in Breast Cancer Patients

Fig 3. EW intervention and physical health in 3-month follow-up group: negative somatic
symptoms.
doi:10.1371/journal.pone.0131802.g003

CI, -19.23 to -6.83, P<0.0001; MD, -9.18, 95% CI, -15.57 to -2.79, P = 0.005] compared with
the control group (Fig 3). A publication bias analysis using a funnel plot was performed on the
studies involving these indexes, and the results exhibited a symmetric distribution, indicating a
low publication bias (Fig 4). Furthermore, Eggers test also indicated a low publication bias
(P = 0.372).
As for secondary outcomes in the 22643-month follow-up group, medical appointments for
cancer-related morbidities (MD, -1.69, 95% CI, -2.30 to -1.08, P<0.00001), worst pain (MD,
1.76, 95% CI, 0.27 to 3.25, P = 0.02) and physical symptoms (MD, -0.26, 95% CI, -0.51 to -0.01,
P = 0.04) showed a significantly beneficial effect on physical health. However, no significant
effects were observed in the following outcomes: least pain, average pain, pain interference,
sleep quality and fatigue. In the >3-month follow-up group, there was only one study with a
9-month follow-up (Table 3). Analysis of the combined effect of EW on the physical symptoms

Table 2. EW intervention and physical health in 3-month follow-up group.

Number of Effect size Heterogeneity

Outcome Comparison participants MD 95% Cl P value 2 P value I2

(%)
1 Negative somatic symptoms scale, 4 sessions 5 Subtotals Subtotals
only only
1.1 emotional prompt vs control 3 140 -13.03 -19.23, <0.0001 0.88 0.64 0
-6.83
1.2 benet-nding prompt vs control 2 78 -9.18 -15.57, 0.005 0.16 0.69 0
-2.79
2 Medical appointments for cancer-related morbidities, 4 4 156 -1.69 -2.30, -1.08 <0.00001 0.99 0.80 0
sessions
3 Worst pain: Brief Pain Inventory, 4 sessions 1 49 1.76 0.27, 3.25 0.02 NA NA NA
4 Least pain: Brief Pain Inventory, 4 sessions 1 49 0.82 -0.22, 1.86 0.12 NA NA NA
5 Average pain: Brief Pain Inventory, 4 sessions 1 49 0.82 -0.34, 1.98 0.16 NA NA NA
6 Pain interference: Brief Pain Inventory (BPI), 4 sessions 1 49 1.28 -0.05, 2.61 0.06 NA NA NA
7 Pittsburgh Sleep Quality Index (PSQI), 4 sessions 3 199 0.65 -0.15, 1.46 0.11 0.42 0.81 0
8 The Functional Assessment of Chronic Illness Therapy 1 86 -2.20 -5.49, 1.09 0.19 NA NA NA
Fatigue subscale (FACIT-F), 4 sessions
9 M.D. Anderson Symptom Inventory (MDASI), 6 sessions 1 58 -11.17 -29.30, 6.96 0.23 NA NA NA
10 Pennebakers Inventory of Limbic Languidness (PILL), 1 58 -6.76 -21.35, 7.83 0.36 NA NA NA
6 sessions
11 Physical health measure, 1 session 1 80 -0.26 -0.51, -0.01 0.04 NA NA NA

NA, not applicable.

doi:10.1371/journal.pone.0131802.t002

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 10 / 19

 EW Intervention in Breast Cancer Patients

Fig 4. Funnel plot for EW intervention and physical health in 3-month follow-up group: negative
somatic symptoms.
doi:10.1371/journal.pone.0131802.g004

using a fixed-effect model demonstrated no significant impact of EW on the physical symp-

toms in BC patients (P>0.05).
Expressive writing intervention and psychological health outcomes. The influence of
EW on the psychological health of BC patients in all the 11 studies [1929] was analyzed using
a fixed-effect model. Our meta-analysis showed no significant effect of EW on the negative
mood, our primary outcome for psychological health, in the 3-month follow-up group
though one study showed that EW with a benefit-finding prompt significantly reduced the neg-
ative mood level in BC patients compared with the control group (P = 0.02; Fig 5). A publica-
tion bias analysis of the studies involving these indexes was performed via a funnel plot, and
the results followed a symmetric distribution, indicating a low publication bias (Fig 6). Further-
more, Eggers test also showed a low publication bias (P = 0.975). The secondary outcomes,
such as positive mood, stress, depression, intrusive thoughts, avoidance, anxiety and quality of
life, in the 3-month follow-up group showed insignificant effects on the psychological health
(P>0.05) (Table 4). In the >3-month follow-up group of five studies (Table 5), a fixed-effect
model was used to analyze the combined effect of EW on the psychological health. Their results
showed no significant effect of EW on the psychological health (P>0.05).

Discussion
In this systematic review of 11 RCTs exploring the influence of EW on the health outcomes of
BC patients, we tried to determine specific efficacy of EW as a potential therapeutic aid. Based
on the physical health indexes, we found that the negative somatic symptoms of BC patients
(measured by the Negative Somatic Symptoms Scale) were significantly relieved after EW inter-
vention for 3 months using either an emotional prompt or a benefit-finding prompt

Table 3. EW intervention and physical health in >3-month follow-up group.

Number of Effect size Heterogeneity

Outcome comparisons participants MD 95% Cl P value 2 P value I2 (%)

11 Physical health measure, 1 session 1 80 -0.02 -0.26, 0.22 0.87 NA NA NA

NA, not applicable.

doi:10.1371/journal.pone.0131802.t003

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 11 / 19

 EW Intervention in Breast Cancer Patients

Fig 5. EW intervention and psychological health in 3-month follow-up group: negative mood based
on POMS.
doi:10.1371/journal.pone.0131802.g005

compared with the control group This study also identified a significant effect of EW on medi-
cal appointments for cancer-related morbidities, worst pain (measured by Brief Pain Inventory)
and physical symptoms (measured by instrument Physical Health Measure) (Table 2).
Although negative somatic symptoms, medical appointments for cancer-related morbidities,
worst pain and physical symptoms are different indexes adopted by different assessment
instruments, they have, in fact, much in common. They all assess status of physical health in
BC patients from different points of view or with different focuses or using different terms.
Taken together, the RCTs that evaluated these indexes supported the efficacy of EW on reliev-
ing negative physical symptoms in BC patients. Consequently, we can safely conclude that EW
intervention may benefit the physical health of BC patients. However, we found that this bene-
fit became insignificant >3 months after EW intervention, which means the positive effect of
current EW intervention may not last long. This is in agreement with one of Mogks conclu-
sions about health effects of EW [54]. However, we believe that the duration of effectiveness of
EW intervention is an important issue that needs further investigation. It is associated with
intervention methods or dosages or an inborn limitation of EW itself? As for the psychological
health for BC patients, we found no significant effectiveness of EW intervention in either
>3-month follow-up group or 3-month follow-up group except that only one study reported
that EW with a benefit-finding prompt for 3 months significantly reduced the negative mood

Fig 6. Funnel plot for EW intervention and psychological health in 3-month follow-up group:
negative mood based on POMS.
doi:10.1371/journal.pone.0131802.g006

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 12 / 19

 EW Intervention in Breast Cancer Patients

Table 4. EW intervention and psychological health in 3-month follow-up group.

Number of Effect size Heterogeneity

Outcome comparisons participants MD 95% Cl P 2 P value I2

value (%)
1 Negative mood: Prole of Mood States (POMS), 34 4 Subtotals Subtotals
sessions only only
1.1 emotional prompt vs control 2 449 -0.54 -3.07, 1.99 0.68 0.03 0.86 0
1.2 benet-nding prompt vs control 1 35 -4.18 -7.62, -0.74 0.02 NA NA NA
1.3 multiple prompt vs control 1 80 -2.59 -18.68, 0.75 NA NA NA
13.50
2 Positive mood: Prole of Mood States (POMS), 4 sessions 2 71 0.16 -0.95, 1.26 0.78 0.24 0.62 0
3 Passive Positive Mood Scale (PPMS), 3 sessions 1 413 -0.40 -1.75, 0.95 0.56 NA NA NA
4 Brief Symptom Inventory 18 (BSI-18), 4 sessions 1 49 0.03 -3.20, 3.26 0.99 NA NA NA
5 Perceived Stress Scale (PSS), 4 sessions 1 49 0.84 -2.08, 3.76 0.57 NA NA NA
6 Beck Depression Inventory-Short Form (BDI-SF), 3 sessions 1 418 -0.50 -1.32, 0.32 0.23 NA NA NA
7 Intrusive thoughts: Impact of Event Scale (IES), 34 4 536 -0.50 -.68, 0.68 0.40 1.18 0.76 0
sessions
8 Avoidance: Impact of Event Scale (IES), 3 sessions 3 474 -1.09 -2.59, 0.41 0.15 2.42 0.30 17
9 Meaning/ peace: The Functional Assessment of Chronic 1 86 -0.98 -2.63, 0.67 0.24 NA NA NA
Illness Therapy-Spiritual Well-being scale (FACIT-Sp), 4
sessions
10Demoralization: The Functional Assessment of Chronic 1 86 1.57 -3.05, 6.19 0.51 NA NA NA
Illness Therapy-Spiritual Well-being scale (FACIT-Sp), 4
sessions
11 Distress Thermometer (DT), 4 sessions 1 86 0.16 -0.85, 1.17 0.76 NA NA NA
12 Anxiety: Hospital Anxiety and Depression Scale (HADS-A), 2 144 -0.59 -1.72, 0.54 0.31 0.12 0.73 0
4 sessions
13 Depression: Hospital Anxiety and Depression Scale 1 58 2.48 -1.09, 6.05 0.97 NA NA NA
(HADS), 6 sessions
14 Center for Epidemiologic Studies-Depression Scale 3 228 -0.23 -0.48, 0.03 0.08 0.04 0.98 0
(CES-D), 14 sessions
15 Positive affect: Positive and Negative Affect Scale 2 53 2.48 -1.09, 6.05 0.17 0.04 0.85 0
(PANAS), 13 sessions
16 Negative affect: Positive and Negative Affect Scale 2 53 2.14 -0.65, 4.92 0.13 0.18 0.67 0
(PANAS), 13 sessions
17 Cancer-Quality of Life (C-QOL), 6 sessions 1 58 1.31 -4.71, 7.33 0.67 NA NA NA
18 Functional Assessment of Cancer Therapy-Breast Cancer 3 229 -0.67 -5.01, 3.66 0.76 0.63 0.73 0
Version (FACT-B), 4 sessions

NA, not applicable.

doi:10.1371/journal.pone.0131802.t004

level in BC patients. This finding was rather surprising, because as a means of psychological
adjustment EW was expected to exert a positive effect on psychological health.
There have been only four meta-analyses available currently on the EW intervention. The
above findings are basically consistent with the meta-analysis by Frisina [52] who determined
that written emotional disclosure had a significant effect on the physical but not the psycholog-
ical health of various clinical populations, and with that by Frattaroli [50] who found EW was
effective in both healthy and unhealthy people. However, Harris [53] found EW was effective
in healthy people but not in samples defined by medical diagnosis or psychological criteria.
Our findings are inconsistent with the meta-analysis by Mogk [54] who concluded that EW
had minor or no effects on the healthy or unhealthy subjects in their study. Obviously, the
diversity of the populations included and the heterogeneity of the meta-analyses are likely a

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 13 / 19

 EW Intervention in Breast Cancer Patients

Table 5. EW intervention and psychological health in >3-month follow-up group.

Number of Effect size Heterogeneity

Outcome comparisons participants MD 95% Cl P 2 P I2

value value (%)
1 Negative mood: Prole of Mood States (POMS), 34 sessions 2 511 -2.19 -5.93, 0.25 0.03 0.85 0
1.55
2 Passive Positive Mood Scale (PPMS), 3 sessions 1 431 0.10 -1.19, 0.88 NA NA NA
1.39
3 Intrusive thoughts: Impact of Event Scale (IES), 34 sessions 3 483 0.56 -0.78, 0.41 2.41 0.30 17
1.90
4 Avoidance: Impact of Event Scale (IES), 3 sessions 3 483 -0.57 -2.07, 0.46 0.35 0.84 0
0.94
5 Positive affect: Positive and Negative Affect Scale (PANAS) 2 53 2.84 -0.64, 0.11 0.58 0.45 0
6.31
6 Negative affect: Positive and Negative Affect Scale (PANAS), 2 53 3.00 -0.33, 0.08 0.00 1.00 0
6.33
7 Center for Epidemiologic Studies-Depression Scale (CES-D), 1 1 80 -0.08 -0.35, 0.56 NA NA NA
session 0.19
8 Beck Depression Inventory-Short Form (BDI-SF), 3 sessions 1 435 -0.30 -.14, 0.48 NA NA NA
0.54
9 Functional Assessment of Cancer Therapy-Breast Cancer Version 3 229 2.26 -1.85, 0.28 1.58 0.45 0
(FACT-B), 4 sessions 6.37

NA, not applicable.

doi:10.1371/journal.pone.0131802.t005

major reason for the inconsistency in the conclusions on the EW intervention. This was also
the reason why we decided to conduct a meta-analysis on EW intervention specifically in BC
population.
The poor homogeneity of the limited number of current meta-analyses on EW intervention,
including ours, reflects a fact that the researches on EW as an adjuvant therapy for patients are
still diversified. The assessment instruments, indexes, outcome measures, samples included,
indications, intervention methods, and conclusions are all various. Notably, EW has been
tested or used chiefly in Caucasians rather than in oriental races. There is a long way to go
before EW can be used as a well-developed intervention in clinic.
The poor homogeneity of the RCTs investigating EW intervention also led to the chief limi-
tations of our meta-analysis. Firstly, because of inconsistent outcome measures by different
instruments and incomplete data, the majority of the indexes were unable to be pooled for
meta-analysis, leading to inclusion of N = 1 tests in the results. For example, of the 11 studies
included in this meta-analysis, only five addressed negative mood as an index of psychological
health using the same instrument POMS with emotional prompt, benefit-finding prompt or
multiple prompts. Of the five studies, only three provided detailed data we were able to retrieve
for sub-group analysis of whether different prompts in EW intervention might have an effect
on the intervention outcomes (Fig 5). Other studies measured the effects of EW on psychologi-
cal health using other instruments or other indexes (Table 4). In addition, the limited number
of studies available made it very difficult for us to investigate the potential publication bias
influencing the analysis. The power of Eggers test is too low to distinguish chance from real
asymmetry. Secondly, inconsistent factors or indexes, such as characteristics of the subjects
and writing prompt and cycle of EW intervention, might have caused biases in results. Thirdly,
it was infeasible for us to determine more specific efficacy of EW by further sub-group analyses
according to age, education or BC staging because the data were lacking or because the sample

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 14 / 19

 EW Intervention in Breast Cancer Patients

size was very limited. Moreover, we did not perform manual retrievals due to time and condi-
tion limitations. In addition, three studies failed to provide complete data for five observation
indexes even after consultation with the authors [24,25,28], and the full text and data of three
conference papers were unavailable even after we tried to contact their authors. All these might
have missed some important data on EW as an intervention in BC patients, leading to possible
biases in our results.
In summary, this meta-analysis has found that EW intervention may have a positive effect
on the physical health rather than the psychological health of BC patients. It proves to chiefly
relieve negative physical symptoms in a short term rather than in a long term. The non-lasting
momentum of EW intervention is an interesting issue to explore. As a convenient adjunctive
intervention for BC patients, EW has much more to be clarified. To determine if personality,
race, age, education, BC stage, intervention time since diagnosis and other possible factors
would be associated with the intervention effect of EW, it is necessary to do more research in
large populations of BC patients to develop EW as an established intervention.

Supporting Information
S1 Checklist. PRISMA 2009 checklist in this meta-analysis.doc.
(DOC)
S1 Appendix. Search strategy.doc.
(DOC)

Acknowledgments
We would like to thank Professor Xu Xia, Department of Medical Informatics, Southern Medi-
cal University, for developing the search strategies and for helps with the search tools. We also
thank Professor Ping Allen Liang and Doctor Da Wang for their careful revision of this
manuscript.

Author Contributions
Conceived and designed the experiments: CLZ YNW SLA. Performed the experiments: CLZ
YNW SLA XJL. Analyzed the data: YNW SLA. Contributed reagents/materials/analysis tools:
YNW SLA XJL. Wrote the paper: CLZ YNW SLA XJL. Designed the search strategy and per-
formed and updated the search: YNW XJL. Screened the reference files and extracted the data:
CLZ YNW XJL. Contacted the authors: YNW.

References
1. Servaes P, Vingerhoets AJ, Vreugdenhil G, Keuning JJ, Broekhuijsen AM. Inhibition of emotional
expression in breast cancer patients. Behav Med. 1999; 25: 2327. PMID: 10209695
2. Spiegel D, Bloom JR, Kraemer H, Gottheil E. Psychological support for cancer patients. Lancet. 1989;
2: 1447.
3. Rachman SJ. Emotional processing. Behav Res Ther. 1980; 18: 5160. PMID: 7369988
4. Alexander F. Psychosomatic medicine: Its principles and applications. 1st ed. New York: Norton.
1950.
5. Alexander F. Emotional factors in essential hypertension. Psychosom Med. 1939; 1: 173179.
6. Gross J. Emotional expression in cancer onset and progression. Soc Sci Med. 1989; 28: 12391248.
PMID: 2660280
7. Cepeda MS, Chapman CR, Miranda N, Sanchez R, Rodriguez CH, Restrepo AE, et al. Emotional dis-
closure through patient narrative may improve pain and well-being: Results of a randomized controlled
trial in patients with cancer pain. J Pain Symptom Manage. 2008; 35: 623631. doi: 10.1016/j.
jpainsymman.2007.08.011 PMID: 18359604

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 15 / 19

 EW Intervention in Breast Cancer Patients

8. Pennebaker JW, Beall SK. Confronting a traumatic event: toward an understanding of inhibition and
disease. J Abnorm Psychol. 1986; 95: 274281. PMID: 3745650
9. Smyth JM. Written emotional expression: effect sizes, outcome types, and moderating variables. J
Consult Clin Psychol. 1998; 66: 174184. PMID: 9489272
10. Sloan DM, Marx BP. A closer examination of the structured written disclosure procedure. J Consult Clin
Psychol. 2004; 72: 165175. PMID: 15065952
11. King LA. The health benefits of writing about life goals. Pers Soc Psychol Bull. 2001; 27: 798807.
12. Burton C, King L. Effects of (very) brief writing on health: the two-minute miracle. Br J Health Psychol.
2008; 13: 914. doi: 10.1348/135910707X250910 PMID: 18230223
13. Smyth JM, Stone AA, Hurewitz A, Kaell A. Effects of writing about stressful experiences on symptom
reduction in patients with asthma or rheumatoid arthritis: a randomized trial. JAMA. 1999; 281: 1304
1309. PMID: 10208146
14. Broderick JE, Stone AA, Smyth JM, Kaell AT. The feasibility and effectiveness of an expressive writing
intervention for rheumatoid arthritis via home-based videotaped instructions. Ann Behav Med. 2004;
27: 5059. PMID: 14979863
15. Harris AH, Thoresen CE, Humphreys K, Faul J. Does writing affect asthma? A randomized trial. Psy-
chosom Med. 2005; 67: 130136. PMID: 15673635
16. Rivkin ID, Gustafson J, Weingarten I, Chin D. The effects of expressive writing on adjustment to HIV.
AIDS Behav. 2006; 10: 1326. PMID: 16421649
17. Manzoni GM, Castelnuovo G, Molinari E. The written-heart study (expressive writing for heart healing):
rationale and design of a randomized controlled clinical trial of expressive writing in coronary patients
referred to residential cardiac rehabilitation. Health Qual Life Outcomes. 2011; 9: 51. doi: 10.1186/
1477-7525-9-51 PMID: 21740564
18. Milbury K, Spelman A, Wood C, Matin SF, Tannir N, Jonasch E, et al. Randomized Controlled Trial of
Expressive Writing for Patients With Renal Cell Carcinoma. J Clin Oncol. 2014; 32: 663670. doi: 10.
1200/JCO.2013.50.3532 PMID: 24470003
19. Walker BL, Nail LM, Croyle RT. Does emotional expression make a difference in reaction to breast can-
cer? Oncol Nurs Forum. 1999; 26: 10251032. PMID: 10420420
20. Park EY, Yi M. Development and effectiveness of expressive writing program for women with breast
cancer in Korea. J Korean Acad Nurs. 2012; 42: 269279. doi: 10.4040/jkan.2012.42.2.269 PMID:
22699176
21. Mosher CE, DuHamel KN, Lam J, Dickler M, Li Y, Massie MJ, et al. Randomised trial of expressive writ-
ing for distressed metastatic breast cancer patients. Psychol Health. 2012; 27: 88100. doi: 10.1080/
08870446.2010.551212 PMID: 21678181
22. Low CA, Stanton AL, Bower JE, Gyllenhammer L. A randomized controlled trial of emotionally expres-
sive writing for women with metastatic breast cancer. Health Psychol. 2010; 29: 460466. doi: 10.1037/
a0020153 PMID: 20658835
23. Jensen-Johansen MB, Christensen S, Valdimarsdottir H, Zakowski S, Jensen AB, Bovbjerg DH, et al.
Effects of an expressive writing intervention on cancer-related distress in Danish breast cancer survi-
vorsresults from a nationwide randomized clinical trial. Psychooncology. 2012; 22: 14921500. doi:
10.1002/pon.3193 PMID: 22991162
24. Henry EA, Schlegel RJ, Talley AE, Molix LA, Bettencourt BA. The feasibility and effectiveness of
expressive writing for rural and urban breast cancer survivors. Oncol Nurs Forum. 2010; 37: 749757.
doi: 10.1188/10.ONF.749-757 PMID: 21059586
25. Gellaitry G, Peters K, Bloomfield D, Home R. Narrowing the gap: the effects of an expressive writing
intervention on perceptions of actual and ideal emotional support in women who have completed treat-
ment for early stage breast cancer. Psychooncology. 2010; 19: 7784. doi: 10.1002/pon.1532 PMID:
19194996
26. de Moor JS, Moye L, Low MD, Rivera E, Slingletary SE, Fouladi RT, et al. Expressive writing as a pre-
surgical stress management intervention for breast cancer patients. J Soc Integr Oncol. 2008; 6: 59
66. PMID: 18544285
27. Craft MA, Davis GC, Paulson RM. Expressive writing in early breast cancer survivors. J Adv Nurs.
2013; 69: 305315. doi: 10.1111/j.1365-2648.2012.06008.x PMID: 22494086
28. Stanton AL, Danoff-Burg S, Sworowski LA, Collins CA, Branstetter AD, Rodriquez-Hanley A, et al. Ran-
domized, controlled trial of written emotional expression and benefit finding in breast cancer. J Clin
Oncol. 2002; 20: 41604168. PMID: 12377959
29. Low CA, Stanton AL, Danoff-Burg S. Expressive disclosure and benefit finding among breast cancer
patients: mechanisms for positive health effects. Health Psychol. 2006; 25: 181189. PMID: 16569109

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 16 / 19

 EW Intervention in Breast Cancer Patients

30. American Cancer Society. Cancer facts and figures 2010. Atlanta: American Cancer Sociey; 2010.
31. Zakowski SG, Harris C, Krueger N, Laubmeier KK, Garrett S, Flanigan R, et al. Social barriers to emo-
tional expression and their relations to distress in male and female. Br J Health Psychol. 2003; 8: 271
286. PMID: 14606973
32. Tamagawa R, Giese-Davis J, Speca M, Doll R, Stephen J, Carlson LE. Trait mindfulness, repression,
suppression, and self-reported mood and stress symptoms among women with breast cancer. J Clin
Psychol. 2013; 69: 264277. doi: 10.1002/jclp.21939 PMID: 23280695
33. Aranda S, Schofield P, Weih L, Yates P, Milne D, Faulkner R, et al. Mapping the quality of life and
unmet needs of urban women with metastatic breast cancer. Eur J Cancer Care (Engl). 2005; 14: 211
222.
34. Stefanic N, Caputi P, Iverson DC. Investigating physical symptom burden and personal goal interfer-
ence in early-stage breast cancer patients. Support Care Cancer. 2014; 22: 713720. doi: 10.1007/
s00520-013-2026-x PMID: 24193221
35. Cleeland CS. Symptom burden: multiple symptoms and their impact as patient-reported outcomes. J
Natl Cancer Inst Monogr. 2007; 37:1621. PMID: 17951226
36. Cella D, Fallowfield LJ. Recognition and management of treatment-related side effects for breast can-
cer patients receiving adjuvant endocrine therapy. Breast Cancer Res Treat. 2008; 107: 167180.
PMID: 17876703
37. Fallowfield L, Cella D, Cuzick J, Francis S, Locker G, Howell A. Quality of life of postmenopausal
women in the Arimidex, Tamoxifen, Alone or in Combination (ATAC) Adjuvant Breast Cancer Trial. J
Clin Oncol. 2004; 22: 42614271. PMID: 15514369
38. Hunter MS, Grunfeld EA, Mittal S, Sikka P, Ramirez AJ, Fentiman I, et al. Menopausal symptoms in
women with breast cancer: prevalence and treatment preferences. Psychooncology. 2004; 13: 769
778. PMID: 15386641
39. Earnhardt JL, Martz DM, Ballard ME, Curtin L. A writing intervention for negative body image: Penneba-
ker fails to surpass the placebo. J College Stud Psychother. 2002; 17: 1935.
40. Zakowski SG, Ramati A, Morton C, Johnson P, Flanigan R. Written emotional disclosure buffers the
effects of social constraints on distress among cancer patients. Health Psychol. 2004; 23: 555563.
PMID: 15546223
41. Broderick JE, Stone AA, Smyth JM, Kaell AT. The feasibility and effectiveness of an expressive writing
intervention for rheumatoid arthritis via home-based videotaped instructions. Ann Behav Med. 2004;
27: 5059. PMID: 14979863
42. Fernandez I, Paez D. The benefits of expressive writing after the Madrid terrorist attack: Implications for
emotional activation and positive affect. Br J Health Psychol. 2008; 13: 3134. doi: 10.1348/
135910707X251234 PMID: 18230227
43. Wagner LJ, Hilker KA, Hepworth JT, Wallston KA. Cognitive adaptability as a moderator of expressive
writing effects in an HIV sample. AIDS Behav. 2010; 14: 410420. doi: 10.1007/s10461-008-9427-8
PMID: 18607714
44. Lumley MA, Keefe FJ, Mosley-Williams A, Rice JR, McKee D, Waters SJ, et al. The effects of written
emotional disclosure and coping skills training in rheumatoid arthritis: a randomized clinical trial. J Con-
sult Clin Psychol. 2014; 82: 644658. doi: 10.1037/a0036958 PMID: 24865870
45. Kelley JE, Lumley MA, Leisen JC. Health effects of emotional disclosure in rheumatoid arthritis
patients. Health psychol. 1997; 16: 331340. PMID: 9237085
46. Klein K, Boals A. Expressive writing can increase working memory capacity. J Exp Psychol Gen. 2001;
130: 520533. PMID: 11561925
47. Broderick JE, Junghaenel DU, Schwartz JE. Written emotional expression produces health benefits in
fibromyalgia patients. Psychosom Med. 2005; 67: 326334. PMID: 15784801
48. Pennebaker JW, Colder M, Sharp LK. Accelerating the coping process. J Pers Soc Psychol. 1990; 58:
528537. PMID: 2324942
49. Paradisi A, Abeni D, Finore E, Di Pietro C, Sampogna F, Mazzanti C, et al. Effect of written emotional
disclosure interventions in persons with psoriasis undergoing narrow band ultraviolet B phototherapy.
Eur J Dermatol. 2010; 20: 599605. doi: 10.1684/ejd.2010.1018 PMID: 20605769
50. Frattaroli J. Experimental disclosure and its moderators: a meta-analysis. Psychol Bull. 2006; 132:
823865. PMID: 17073523
51. Merz EL, Fox RS, Malcarne VL. Expressive writing interventions in cancer patients: a systematic
review. Health Psychol Rev. 2014; 6: 3741.
52. Frisina PG, Borod JC, Lepore SJ. A meta-analysis of the effects of written emotional disclosure on the
health outcomes of clinical populations. J Nerv Ment Dis. 2004; 192: 629634. PMID: 15348980

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 17 / 19

 EW Intervention in Breast Cancer Patients

53. Harris AH. Does expressive writing reduce health care utilization? A meta-analysis of randomized trials.
J Consult Clin Psychol. 2006; 74: 243252. PMID: 16649869
54. Mogk C, Otte S, Reinhold-Hurley B, Kroner-Herwig B. Health effects of expressive writing on stressful
or traumatic experiencesa meta-analysis. Psycho-social medicine. 2006; 3: Doc06.
55. Boinon D, Charles C, Dauchy S, Sultan S. Effects of emotional expression on adjustment to cancer: a
systematic review of research. Psycho-Oncologie. 2011; 5: 173190.
56. Goldberg JA, Scott RN, Davidson PM, Murray GD, Stallard S, George WD, et al. Psychological morbid-
ity in the first year after breast surgery. Eur J Surg Oncol. 1992; 18: 327331. PMID: 1521623
57. Kain ZN, Sevarino F, Alexander GM, Pincus S, Mayes LC. Preoperative anxiety and postoperative pain
in women undergoing hysterectomy. A repeated-measures design. J Psychosom Res. 2000; 49: 417
422. PMID: 11182434
58. Kissane DW, Grabsch B, Love A, Clarke DM, Bloch S, Smith GC. Psychiatric disorder in women with
early stage and advanced breast cancer: a comparative analysis. Aust N Z J Psychiatry. 2004; 38:
320326. PMID: 15144508
59. Palesh OG, Collie K, Batiuchok D, Tilston J, Koopman C, Perlis ML, et al. A longitudinal study of
depression, pain, and stress as predictors of sleep disturbance among women with metastatic breast
cancer. Biol Psychol. 2007; 75: 3744. PMID: 17166646
60. Knobf MT. Clinical update: psychosocial responses in breast cancer survivors. Semin Oncol Nurs.
2011; 27: e1e14.
61. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global Cancer Statistics. CA Cancer J Clin.
2011; 61: 6990. doi: 10.3322/caac.20107 PMID: 21296855
62. Hunter MS, Coventry S, Mendes N, Grunfeld EA. Menopausal symptoms following breast cancer treat-
ment: a qualitative investigation of cognitive and behavioural responses. Maturitas. 2009; 63: 336340.
doi: 10.1016/j.maturitas.2009.06.003 PMID: 19616392
63. Nyrop KA, Muss HB, Hackney B, Cleveland R, Altpeter M, Callahan LF. Feasibility and promise of a 6-
week program to encourage physical activity and reduce joint symptoms among elderly breast cancer
survivors on aromatase inhibitor therapy. J Geriatr Oncol. 2014; 5: 148155. doi: 10.1016/j.jgo.2013.
12.002 PMID: 24495696
64. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer
in 2008: GLOBOCAN 2008. Int J Cancer. 2010; 127:28932917. doi: 10.1002/ijc.25516 PMID:
21351269
65. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002; 21: 1539
1558. PMID: 12111919
66. Lu Q, Zheng D, Young L, Kagawa-Singer M, Loh A. A pilot study of expressive writing intervention
among Chinese-speaking breast cancer survivors. Health Psychol. 2012; 31: 548551. doi: 10.1037/
a0026834 PMID: 22229930
67. Carlson LE, Doll R, Stephen J, Faris P, Tamagawa R, Drysdale E, et al. Randomized controlled trial of
Mindfulness-based cancer recovery versus supportive expressive group therapy for distressed survi-
vors of breast cancer. J Clin Oncol. 2013; 31: 31193126. doi: 10.1200/JCO.2012.47.5210 PMID:
23918953
68. Lee HS, Cohn LD. Assessing coping strategies by analysing expressive writing samples. Stress Health.
2009; 26: 250260.
69. Jensen-Johansen MM, Christensen S, Valdimarsdottir H, Zakowsk S, Bovbjerg D. An expressive writ-
ing intervention improved self-reported social support among early stage breast cancer patients-results
from a danish population-based, randomized clinical trial. Psycho-Oncology. 2011; 20 (Suppl.2): 30
104.
70. Bardwell WA DJ, Rock CL, Pierce JP. To express or not to express: negative emotional expression
(NEE) vs mental and physical health in women treated for early-stage breast cancer. Psycho-Oncology.
2006; 15: S4.
71. Dubenko L, Greenberg M, Altree PK, Yutsis M, Golant M. Emotional expression in breast cancer sup-
port groups and emotional information available in observational and text-based coding systems. Psy-
cho-Oncology. 2008; 17: S92S93.
72. Bauer-Wu S, Norris R, Powell M, Healey M, Habin K, Partridge A, et al. A web-based expressive writing
intervention for young women with newly diagnosed breast cancer. Oncol Nurs Forum. 2007; 34: 202.
73. Johansen MB, Zachariae R, Valdimarsdottir H, Bovbjerg D, Zkowski S. Expressive writing and breast
cancer: Associations between cognitive and positive emotional words and changes in perceived social
support. Psycho-Oncology. 2006; 15: S278.

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 18 / 19

 EW Intervention in Breast Cancer Patients

74. Bauer-Wu S, Norris R, Healey M, Powell M, Habin K. An innovative expressive writing intervention for
young breast cancer patients: Feasibility, preference and psychological effects. Psycho-Oncology.
2007; 16: S78S79.
75. Craft M. Expressive writing in newly diagnosed breast cancer patients. Oncol Nurs Forum. 2007; 34:
507.
76. Laccetti M. Expressive writing in women with advanced breast cancer. Oncol Nurs Forum. 2007; 34:
10191024. PMID: 17878130
77. Ridner SH, Bonner CM, Deng J, Sinclair VG. Voices from the shadows: living with lymphedema. Can-
cer Nurs. 2012; 35: E1826. doi: 10.1097/NCC.0b013e31821404c0 PMID: 21558848
78. Pennebaker JW. The psychology of physical symptoms. New York: Springer-Verlag; 1982.

PLOS ONE | DOI:10.1371/journal.pone.0131802 July 7, 2015 19 / 19