Background

Breast cancer has become the most frequently diagnosed cancer in Chinese women (Fan et al., 2014; Ferlay et al., 2015). In 2012, the age-standardized incidence rate in China was 18.7/100,000, accounting for 11.2% of all newly diagnosed breast cancers worldwide (Ferlay et al., 2015). The active use of cancer treatments in China has increased the number of breast cancer survivors, but the side effects of cancer treatments have also brought physical and psychosocial problems, influencing the quality of life of Chinese breast cancer survivors (So et al., 2010). A recent study showed that breast cancer patients in China had a lower quality of life compared to those in the U.S. (Lu et al., 2016). However, few interventions have focused on improving quality of life among Chinese breast cancer survivors. The present study is thus designed to examine the effect of an expressive writing intervention on quality of life among Chinese breast cancer patients.

Expressive writing is a brief intervention instructing participants to write their deepest thoughts and feelings about a stressful life event (Pennebaker & Beall, 1986). Previous research has demonstrated that writing as little as 15–20 min for 3 days can lead to improvements in both psychological (Richards et al., 2000) and biological health (Petrie et al., 1995). Researchers have documented the physical and psychological benefits of expressive writing among non-patient (Burton & King, 2008; Sloan & Marx, 2004) and patient populations (Stanton et al., 2002; Zakowski et al., 2004). A recent meta-analysis of 11 randomized controlled trials revealed that expressive writing reduced negative somatic symptoms in breast cancer patients in ≤ 3-month follow-ups (Zhou et al., 2015). Psychological health improvements have also been reported among breast cancer patients after expressive writing interventions (Lu et al., 2012, 2017 Stanton et al., 2000, 2002). Contrary to the above studies that have demonstrated benefits of expressive writing, a recent meta-analysis of 16 randomized controlled trials did not find evidence for benefits of expressive writing on psychological, physical, or quality of life outcomes among cancer patients (Zachariae & O’Toole, 2015). However, the review suggested that those who experience low levels of emotional support may benefit from expressive writing.

Compared to individuals from Western societies, Chinese individuals, in general, are less likely to disclose their emotions and psychological distress. This may be because Chinese culture prioritizes collectivism and social harmony over individual expression (Wu & Tseng, 1985). In Chinese culture, displaying strong emotions is considered to be a weakness and was traditionally even thought to be a cause of illness (Wu & Tseng, 1985). The cultural values of emotional restraint and harmony have been associated with reluctance to solicit support from family and friends (Chang, 2014) and have led to higher levels of ambivalence over emotion expression (AEE) in Chinese breast cancer patients compared to European Americans and Chinese Americans in the United States (Ji et al., 2019). Expressive writing is a private intervention that facilitates emotional expression without damaging harmony with others and may be especially suitable for Chinese. Indeed, studies have found that Asian Americans with higher levels of AEE benefit more from expressive writing compared to those with lower levels of AEE (Lu & Stanton, 2010).

Previous studies of expressive writing on cancer patients have mainly been conducted among non-Hispanic Whites (Stanton et al., 2002; Zakowski et al., 2004) and only a few have been conducted among Asians in the U.S. (Lu et al., 2017, 2018). A typical expressive writing experiment usually requires participants to write about their deepest thoughts and feelings related to their cancer experience, but this has been found to be beneficial for Europeans but not Asians (Knowles et al., 2011; Lu & Stanton, 2010). This may be because Asians are not comfortable diving into emotions directly due to Asian cultures’ valuing emotional control and restraint over emotional expression (Lu et al., 2017). Correspondingly, Lu and Stanton (2010) developed a self-regulation model of expressive writing to emphasize the essential role of cognitive reappraisal on stressful events and the facilitating role of emotional disclosure in producing health benefits. The model was tested and shown to be particularly relevant for Asians (Lu & Stanton, 2010; Lu et al., 2017), suggesting that the cognitive component might be a more important ingredient than the emotional component for the expressive writing paradigm to benefit Asians. A cancer-facts writing condition to write about facts related to cancer diagnosis and treatment was always conceptualized as a control condition in previous studies, but it was found to be more beneficial than the self-regulation and emotional disclosure conditions for Chinese Americans in a recent randomized controlled trial (Lu et al., 2017). Asians tend not to openly talk about their cancer diagnosis and experience (Papadopoulos et al., 2010), and, as such, the opportunity to describe cancer in the cancer-facts writing task may be therapeutic because it provides an opportunity for them to process suppressed thoughts (Lu et al., 2017). These studies suggest that the effects of expressive writing may vary depending on writing instructions and population characteristics. As such, it is imperative to determine what types of expressive writing can deliver health benefits among mainland Chinese breast cancer patients.

However, to date, there has been only one study that tested the efficacy of expressive writing among breast cancer patients in mainland China (Lu et al., 2019). This study tested the effect of a positive thinking condition, a self-regulation condition, and a cancer-facts condition of expressive writing on mainland Chinese breast cancer patients undergoing chemotherapy. Participants in the cancer-facts condition and positive thinking condition had higher quality of life compared with the self-regulation condition from baseline to both the 1-month and 2-month follow-ups. Self-regulation was shown to be particularly relevant for Asians (Lu & Stanton, 2010; Lu et al., 2017). However, this study failed to test the benefits of self-regulation since there was no neutral control condition. Given that the cancer-facts control condition was found to be more beneficial for Chinese (Lu et al., 2017, 2019), it is essential to add a neutral control condition in future studies to better understand the potential benefits of expressive writing among mainland Chinese breast cancer patients.

Therefore, the current study aimed to investigate the effects of expressive writing on quality of life with a neutral control condition among mainland Chinese breast cancer patients. We hypothesized that the three experimental conditions (i.e., self-regulation, emotional disclosure, and cancer facts) would improve in QOL compared to the neutral control condition. We additionally explored the differences in QOL between the three experimental conditions and hypothesized that the self-regulation condition and the cancer-facts condition would have better QOL compared to the emotional disclosure condition.

Methods

Participants

Inclusion criteria: (1) being diagnosed with breast cancer between stage 0–III; (2) being within 2 years post-breast cancer diagnosis; (3) being able to read and write Chinese.

Based on the Lu et al. (2018) study and the assumption that women undergoing treatment may experience enhanced benefits of expressive writing, 26 participants per group were needed to detect d = − 0.8 between-group differences. Recruitment took place in the breast surgery ward in the Weifang People’s Hospital. Breast cancer patients who were receiving treatments for cancer or had follow-up visits were introduced to the study by being given a study flier by the nurses. The benefits of expressive writing and the procedure of the present study were briefly introduced in the flier. One hundred thirty-five patients who indicated an interest in the study were contacted by the research stuff to assess their eligibility. Among these patients, 2 did not meet the inclusion criteria, and 15 declined to participate because they were busy or because the tasks were too tedious. The remaining 118 women completed the baseline questionnaires and were randomized to the self-regulation condition (SR group, n = 31), the emotional disclosure condition (EMO group, n = 30), the cancer-facts writing condition (CTL group, n = 27), or the neutral control condition (CON group, n = 30) at the time of recruitment. Among the 118 participants, 103 completed the writing intervention (27 from the SR group, 23 from the EMO group, 23 from the CTL group, and 30 from the CON group) and 15 dropped out (see Fig. 1).

Fig. 1
figure 1

Flowchart

Full size image

Sample characteristics are shown in Table 1. Participants enrolled in the present study had a mean age of 48.0 years (SD = 8.8 years; range 28–75 years). 97.5% of the participants were within one year of receiving their cancer diagnosis and 2.5% were within 1–2 years. Average diagnosis duration was 3.6 months (SD = 2.8 months; range 1–17 months). 3.4% were diagnosed with stage 0 (carcinoma in situ) breast cancer, 20.5% with stage I, 47.0% with stage II, and 29.1% with stage III. 88.1% of participants had received a mastectomy and 11.9% a breast conservation. Most (71.2%) had undergone chemotherapy, and only 5.1% had undergone radiotherapy.

Table 1 Sample characteristics
Full size table

Procedure

This trial has been registered with trial number ChiCTR1900022045. Ethical approval (NO. 2014-317) was obtained from the Research Ethics Committee at Weifang Medical University following Chinese “Ethical review of biomedical research involving human beings” (Li, 2016). Eligible women who agreed to participate were asked to sign a consent form and complete a baseline questionnaire, and were then randomized to one of four conditions: SR, EMO, CTL, or CON. Simple randomization was conducted using a computerized random number generator. The first author conducted the randomization procedure and put group assignments into an envelope. After the participants completed the baseline questionnaire, the research staff opened the envelope and assigned participants to the corresponding condition. Participants in the experimental conditions were not aware of whether they were assigned to the SR, EMO or CTL condition, but participants in the CON condition knew they were assigned to the control group since there were no writing tasks for them.

Participants in the SR, EMO, CTL conditions were given two copies of a calendar and three sealed envelopes with corresponding writing instructions based on their condition assignment. They were asked to schedule three weekly writing sessions on the calendar, keep one copy of the calendar as a reminder, and give one copy back so that researchers could remind them of the next writing session. The three sealed envelopes given to the participants were labeled as “week 1,” “week 2,” and “week 3” with writing instructions for the corresponding week. On the scheduled writing day, researchers called the participants to remind them of the writing task. Participants were asked to only open the corresponding envelope at the scheduled writing time and to write continuously for up to 30 min or until they completed one page of writing.

Participants in the EMO condition were asked to write about their deepest thoughts and feelings about their cancer experience for 3 weeks. Participants in the CTL condition were asked to objectively write about their cancer diagnosis and treatment in detail for 3 weeks. Participants in the SR condition were asked to write about their deepest feelings and thoughts related to their breast cancer experience at week one, their coping strategies to deal with stressors caused by breast cancer at week two, and positive thoughts and feelings regarding their breast cancer experience at week three. Complete writing instructions are available from the online supplemental materials.

After each writing session, participants sealed their writing envelopes and brought them back to the research staff at their next clinic visit. At 3 and 6 months after the final writing, participants received follow-up questionnaires by mail. They were asked to complete and then mail the follow-up questionnaires back.

Measures

Functional assessment of cancer therapy-breast cancer

The Functional Assessment of Cancer Therapy-Breast cancer (FACT-B) version 4 was used to assess quality of life (QoL) of the participants. The FACT-B consists of 37 Likert-type questions covering five domains: physical well-being (PWB, 7 items), social/family well-being (SWB, 7 items), emotional well-being (EWB, 6 items), functional well-being (FWB, 7 items), and additional concerns for breast cancer (BCS, 10 items). The instrument asked respondents to rate how true each statement was for the last 7 days. Response scales ranged from 0 (not at all) to 4 (very much). Negatively worded items were recorded such that a higher score indicated a better QoL. The FACT-B total score is the sum of the scores from all five subscales. The psychometric properties of the FACT-B are well documented, and it was validated in Chinese (Wan et al., 2007). Cronbach α was 0.59–0.85 in the Chinese version of FACT-B (Wan et al., 2007).

Essay ratings

An independent rater unaware of writing group assignment read the essays in random order and judged whether each essay conformed to the writing instructions.

Linguistic analysis

To analyze essays written in Chinese, we used SCLIWC, a Chinese language psychological analysis system inspired by LIWC 2007 and C-LIWC (Gao et al., 2013). SCLIWC is a text analysis software program that calculates the degree to which people use different categories of words across a wide array of texts. The SCLIWC dictionary was developed and validated by Gao et al. from the traditional Chinese version of the LIWC dictionary (C-LIWC dictionary), which has been widely used in analyzing medical patients’ writing. Each lexical item in the C-LIWC dictionary was checked and validated manually to guarantee the accuracy of SCLIWC dictionary. The categories of the SCLIWC dictionary are compatible to C-LIWC. The reliability and validity of SCLIWC have been found to be good (Gao et al., 2013; Zhao et al., 2016).

Statistical analysis

Analyses were conducted using SPSS 22 as intent-to-treat analyses to include all randomized participants (n = 118). Expectation maximization procedures were used to accommodate missing data. Preliminary multivariate analyses of variance for continuous variables and Chi square analyses for categorical variables were conducted on demographic variables, cancer-related parameters, and dependent variables at baseline. Repeated measures analysis of variance was used to investigate the main effect (time) and the interaction effect (time × group) on QoL and its subdimensions. Three sets of regression models were then specified to evaluate the effects of group on QoL and its subdimensions at the final assessment while controlling for baseline data. For these models, QoL and its subdimension scores at the final assessment were regressed on baseline data, with three dummy code variables representing condition. We conducted three models in which the neutral control condition, cancer-facts condition, and emotional disclosure condition were specified as the reference condition, respectively, so as to permit comparisons between all four groups.

Results

Sample characteristics

Participant characteristics at the beginning of the study are shown in Table 1. Fifteen of the 118 participants dropped out the study, leaving 7.6% missing data. A Little’s MCAR test revealed that the data were missing at random (χ2 = 532.5, df = 515, P = 0.29). The completers and non-completers did not differ on demographic variables (i.e., age, education level, employment, household income, religious belief, and marital status) (Ps > 0.33), cancer-related parameters (i.e., time since the first diagnosis, stage, surgery, treated with chemotherapy and radiotherapy) (Ps > 0.31), and variables of interest (i.e., quality of life dimensions) at baseline (Ps > 0.29). Participants in the four groups were compared to determine the success of randomization using ANOVA and Chi square tests. There were no significant group differences in demographic variables, cancer-related parameters, or baseline QoL between the four groups (all P > 0.05). However, post hoc analyses revealed that the EMO group had a higher baseline level of social wellbeing (d = 2.97, P = 0.02) and emotional wellbeing (d = 2.74, P = 0.035) than the SR group. Emotional wellbeing in the EMO group was also higher than that of the CTL group (d = 3.31, P = 0.012).

Essay ratings

The independent rater correctly judged the writing assignment for 85.3% of the 76 sets of essays (80% SR, 86.4% EMO, and 90.4% CTL), indicating good adherence to the writing instructions.

Linguistic analysis

To determine group differences in words used in the written essays, SCLIWC was used to examine word counts in two emotion categories: positive emotions (e.g., confident, satisfied, blessed) and negative emotions (e.g., worried, suspicious, sad). It was also used for two categories that reflect cognitive processes: insight (e.g., think, know, consider) and causation (e.g., because, effect, hence). As expected, the EMO and SR groups used more negative emotion words compared to the CTL group (all P < 0.05), and there was no difference between the EMO and the SR groups. No differences were found in the use of positive emotion and cognitive words among the three writing groups.

Analyses on FACT-B quality-of-life

Repeated measures ANOVA showed significant effects of time and the time × group interaction on overall QoL (F = 13.9, P < 0.001, η2 = 0.20; F = 3.5, P < 0.01, η2 = 0.08), physical well-being (F = 13.08, P < 0.001, η2 = 0.19; F = 5.35, P < 0.001, η2 = 0.12), emotional well-being (F = 3.84, P < 0.05, η2 = 0.06; F = 2.20, P < 0.05, η2 = 0.06), and additional concerns for breast cancer (F = 11.42, P < 0.001, η2 = 0.17; F = 2.72, P < 0.01, η2 = 0.07) (Table 2).

Table 2 Repeated measure analysis of variance on quality of life
Full size table

As there was a significant time × group interaction, a series of regression models were conducted to detect the group differences in QoL and its subdimensions over time. Because four conditions were included in the present study, three regression models were conducted for QoL and each dimension as outcomes, with the first regression model testing the effects of CTL, SR, and EMO in comparison to the CON condition, and the second model testing the effects of EMO, SR, and CON in comparison to the CTL condition, and the third model testing the effects of SR, CTL, and CON in comparison to the EMO condition.

Results from these residualized change regression analyses are presented in Table 3. Results show that the CTL, EMO, and SR groups reported higher levels of overall quality of life (b = 23.20, P < 0.001; b = 24.66, P < 0.001; b = 17.80, P < 0.001) than the control group (see Fig. 2). The three writing group also had higher physical well-being (b = 8.40, P < 0.001; b = 7.47, P < 0.001; b = 6.06, P < 0.001), emotional well-being (b = 5.45, P < 0.001; b = 4.85, P < 0.001; b = 3.93, P < 0.001), and additional concerns for breast cancer (b = 5.73, P < 0.001; b = 6.99, P < 0.001; b = 5.17, P < 0.001) than the control group. The CTL group had higher physical well-being (b = 2.35, P < 0.05) and the EMO group had higher social well-being (b = 1.97, P < 0.05) than the SR group.

Table 3 Residual change models at 6-month follow-up
Full size table
Fig. 2
figure 2

Overall quality of life across time within the four conditions

Full size image

Discussion

The present study aimed to examine the health benefits of expressive writing among mainland Chinese breast cancer patients. Findings suggest the potential benefits of expressive writing and the importance of further testing the expressive writing paradigm among patients in mainland China. Participants in the SR, EMO, and CTL conditions reported higher quality of life than those in the control condition at the 6-month follow-up, which indicates that all three active writing conditions were beneficial for mainland Chinese breast cancer patients.

The mechanisms through which expressive writing is beneficial to one’s health could be explained by several theories. According to inhibition theory, the inhibition of thoughts or feelings regarding a stressful event is detrimental, and, therefore, disclosure of those inhibited thoughts or feelings may alleviate stress and improve physical and psychological health (Freud, 1954). Repeated expression may decrease the attendant negative emotions and physiological arousal, which, in turn, may diminish negative emotions. Thus, the situation is not experienced as dire as originally conceived, and benefit can be extracted from adversity (Stanton et al., 2000). According to cognitive-processing theory, expressive writing may allow for the cognitive reorganization of a stressful event, ultimately reducing physiological activation associated with inhibition and obsessive thinking (Pennebaker, 1993). According to self-regulation theory, expressive writing may allow people to observe themselves expressing and controlling their emotions. This may lead to a stronger sense of emotional regulation self-efficacy and help people feel that their traumas or stressors are more controllable, which could reduce negative affect and produce health benefits (Lepore et al., 2002). Furthermore, Creswell et al. (2007) found that self-disclosure through expressive writing facilitates self-affirmation, which could buffer the stress of traumatic life events and positively affect functional health status. Future studies should further test these mechanisms.

The study also showed that the cancer-facts condition reported a significantly higher level of physical well-being than the self-regulation group at the 6-month follow-up. This is consistent with a previous study showing that mainland Chinese patients undergoing treatment benefited more from cancer-facts writing than self-regulation writing (Lu et al., 2019). This finding is also similar to the Lu et al. (2017) study that showed that the CTL was the most beneficial among Chinese immigrant cancer survivors in the U.S., and also that the CTL facilitated cognitive processes and gaining insight. Asian cultural norms that prioritize emotional restraint to maintain group harmony may have restricted Asians from talking openly about their cancer diagnosis and experience (Papadopoulos et al., 2010), and, therefore, their cancer-related thoughts may have been suppressed. As discussed above, the inhibition of thoughts regarding a stressful event is harmful (Pennebaker, 1993). Consequently, the opportunity to describe the cancer diagnosis and experience in detail in the cancer-facts writing task could be therapeutic as it presents the opportunity to deal with inhibited thoughts in a non-threatening way (Lu et al., 2017).

On the other hand, unexpectedly, we found that the EMO group had higher levels of social well-being than the SR group. This finding was in contrast with the Lu et al. (2017) study that found the EMO group had lower emotional well-being than the SR group. There are at least two plausible explanations. First, the baseline social wellbeing and emotional wellbeing score was higher in the EMO condition; although we examined the group effect after controlling for the baseline data, it is still possible that those who have a higher quality of life to begin with may benefit more from expressive writing. Second, these divergent findings might be explained by sample differences, such as time since diagnosis. In the Lu et al. (2017) study, the average time since diagnosis was 19 months. In contrast, in the present study, the average time since diagnosis was 3.6 months and most participants were newly diagnosed and were receiving radiotherapy or chemotherapy. One study in mainland China found that women newly diagnosed with breast cancer used more catastrophizing and less positive reappraisal compared to healthy women (Li et al., 2015). It is possible that newly diagnosed patients have a lot of emotional trauma and can benefit from emotional disclosure; however, they may not be ready to switch topics from a traumatic experience to a positive one as the SR condition asked of them. This explanation is in line with a previous study showing that the writing condition switching from disclosing trauma to benefit finding did not yield as much benefit as the conditions that focused on either trauma disclosure or benefit finding alone (King & Miner, 2000). Stanton et al. (2002) also discusses that whether benefit finding writing is helpful or not depends on the trajectory of the breast cancer experience. Writing instructions to promote finding benefits shortly after diagnosis or in the midst of treatment might have been much more difficult for women to perform effectively (Stanton et al., 2002). However, finding benefits in another expressive writing study with mainland Chinese breast cancer survivors undergoing treatment did deliver benefits (Lu et al., 2019). Together, these findings suggest that benefit finding could be beneficial; however, future studies need to examine how and when to switch writing topics from emotional disclosure to benefit finding during treatment in order to exceed the benefits of emotional disclosure alone. In summary, the inconsistent results regarding differences between EMO and SR between the current study and the Lu et al. (2017) study might be explained by baseline differences in QOL in the current study or sample differences in the time since diagnosis between the two samples. This study highlights the importance of further investigating the expressive writing paradigm among cancer patients in different survivorship stages.

Study limitations

Limitations of the present study deserve mention. First, one limitation was a primary reliance on participant self-report; however, questionnaire measures were psychometrically sound and empirically validated. Second, the participants in the present study were newly diagnosed mainland Chinese breast cancer patients receiving cancer treatment. Generalizability of our findings to long-term survivors, other cultural groups, and individuals with other types of medical illnesses may not be warranted. Third, due to the small sample sizes for each condition, the study was more hypothesis generating than testing. Future studies are warranted with larger sample sizes to replicate the findings. Finally, previous expressive writing studies with cancer survivors suggest that studies with a no-writing control condition are more likely to find benefits of writing (Gellaitry et al., 2010; Henry et al., 2010). It is possible that the benefits of this study might have been smaller if the control condition was a writing condition as people may expect benefits from writing.

Clinical implications and conclusions

The current study was one of the first expressive writing studies with a neutral control condition conducted among mainland Chinese breast cancer patients. We found that mainland Chinese breast cancer patients in SR, EMO, and CTL conditions reported higher levels of quality of life than those in the neutral control condition at the 6-month follow-up, suggesting potential health benefits of expressive writing. Furthermore, our findings suggest that mainland Chinese breast cancer patients shortly after diagnosis tend to benefit more from cancer-facts writing and emotional disclosure rather than self-regulation. This result contrasts with previous studies conducted in Western countries and suggest that the benefits of expressive writing vary as a function of sample characteristics, such as cancer survivorship stage and culture. Thus, cultural adaptation and evaluation is critical for applying validated psychosocial interventions to new populations.