Couple-Based Communication Interventions for Cancer Patient–Spousal Caregiver Dyads’ Psychosocial Adaptation to Cancer: A Systematic Review

(1) Background: Effective communication among couples in which one has been diagnosed with cancer is critical to improve their psychosocial adaptation to cancer. The objective of this review was to explore the characteristics and measurement outcomes of existing couple-based communication interventions in the cancer context. (2) Methods: Eight electronic databases were searched from database initiation to August 2022 to identify eligible articles. Hand searching was also performed on the included articles’ reference lists and authors. (3) Results: A total of 14 intervention studies were eligible to be included in this review. Cancer couples with distress or communication problems before intervention were more likely to benefit from the couple-based communication interventions. Positive outcomes were reported, including an improvement in relationship functioning (including mutual communication, intimacy, and relationship satisfaction) and individual functioning (including a decline of anxiety, depression and cancer-related concerns, and an increase in psychological adjustment and quality of life). (4) Conclusions: These findings supported the importance of improving mutual communication behaviors to promote cancer patient–spousal caregiver dyads’ psychosocial adaptation to cancer. While most included studies were conducted in western countries and the sample size was relatively small, more research is warranted to develop more efficacious couple-based communication interventions.


Introduction
According to updated global cancer statistics, approximately 19.3 million new cancer cases and an estimated 9.9 million cancer deaths occurred in 2020 for 36 types of cancers in 185 countries [1]. Furthermore, the global incidence of cancer is expected to increase to 28.4 million in 2040 [1]. According to an assessment by the World Health Organization (WHO), cancer was the first or second leading cause of death for the majority of countries in the world during the period from 2000 to 2019 [2]. The high incidence and mortality of cancer suggest that cancer-related issues are still a primary global disease burden. For those cancer patients who are married or in a committed relationship, patients and their spousal caregivers (SCs) generally experienced mutually related adaptation outcomes, such as selfefficacy, mental health, role adjustment, marital satisfaction, and quality of life (QOL) [3]. A study conducted by Lin et al., which focused on colorectal cancer patients (CPs) and their SCs, also found that CPs were positively related to their SCs, and vice versa [4]. Given the reciprocal influence among CPs and SCs, a growing number of researchers have advised to view cancer from a relational or dyadic perspective [5][6][7]. Kayser et al. [8] regard cancer as a "we-disease", which distinguished ownership of cancer. "We-disease" means that CPs and their SCs regard cancer as theirs and take efforts to communally cope with cancerrelated stress as a couple [8]. Badr et al. [9] explored the effect of pronoun use during the natural communication context on a couple's psychosocial adaptation to head and neck proaches and diverse measurements of outcomes [19]. Luo [20] provided multiple topics and comprehensive components for colorectal CP-SC dyads in her intervention study, and suggested that more focused content may make interventions more effective (e.g., focusing on couples' communication skills). At present, a systematic review assessing the effect of couple-based communication interventions on CP-SC dyads' psychosocial adaptation to cancer is lacking. In addition, a summary of the characteristics of the existing interventions is still needed to refine couple-based communication interventions for CP-SC dyads. Therefore, this review aims to (a) summarize the approach and contents of couple-based communication interventions in the cancer context, (b) explore these interventions' feasibility and acceptability, (c) review their impact on couples' relationship and/or individual functioning and (d) identify future research directions.

Search Methods for Eligible Articles
This review was performed in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A systematic search was performed to find all eligible studies which were published in English or Chinese. Six English databases (including CINAHL, Cochrane, Embase, Medline, PsyINFO, PubMed) and two databases (including CNKI and Wanfang Data) from China were searched from database initiation to August 2022. For the six English databases, the following key terms and their synonyms, which were limited to the title or abstract, were used: "cancer" or "tumor" or "oncology" AND "couple" or "spouse" or "partner" AND "communication" or "conservation" or "emotional disclosure" AND "program" or "training" or "education" or "intervention". Regarding the two Chinese databases, we used " 肿瘤" (tumor) or " 癌症" (cancer) AND " 夫妻" (couple) or " 配偶" (spouse) or " 伴侣" (partner) AND " 沟通" (communication) or " 交流" (conversation) AND " 干预" (intervention) or " 健康教育" (healthy education) as key words to search for eligible studies. For the purpose of finding as many relevant articles as possible, the outcome measures (e.g., "relationship satisfaction" or "psychological distress") were not included in the search terms, but they were considered during the selection process according to the inclusion and exclusion criteria. A manual search of reference lists of included studies and the relevant authors was also conducted to find additional eligible studies. Figure 1 shows a flowchart of the searching and selection process.

Selection Criteria for Identifying Articles
The inclusion criteria were as follows: (1) target population of studies was couples in which one had a diagnosis of any type of cancer; (2) interventions specially focused on couple-based communication components; (3) both the patients' and spouses' relationship and/or individual functioning were measured; (4) studies were completely published in peer-reviewed journals in English or Chinese.
The exclusion criteria were as follows: (1) except for SCs, informal caregivers also included other family members or friends; (2) interventions focused on CPs or SCs, but not CP-SC dyads; (3) interventions paid attention to other components (e.g., mindfulness training), instead of focusing on couple-based communication skill training; (4) studies were protocols, reviews, theses, conference proceedings or editorials.

Data Extraction and Synthesis
Two tables were used to fully extract and synthesize important information in the included studies. In Table 1, we primarily summarized the characteristics of the intervention articles, including the author, published year, country where the study was conducted, study aims, study design, characteristics of cancer and participants, theoretical framework, intervention component and dosage and delivery format. In Table 2, we synthesized the included studies' measurement tools and intervention outcomes.

Selection Criteria for Identifying Articles
The inclusion criteria were as follows: (1) target population of studies was couples in which one had a diagnosis of any type of cancer; (2) interventions specially focused on couple-based communication components; (3) both the patients' and spouses' relationship and/or individual functioning were measured; (4) studies were completely published in peer-reviewed journals in English or Chinese.
The exclusion criteria were as follows: (1) except for SCs, informal caregivers also included other family members or friends; (2) interventions focused on CPs or SCs, but not CP-SC dyads; (3) interventions paid attention to other components (e.g., mindfulness training), instead of focusing on couple-based communication skill training; (4) studies were protocols, reviews, theses, conference proceedings or editorials.

Data Extraction and Synthesis
Two tables were used to fully extract and synthesize important information in the included studies. In Table 1, we primarily summarized the characteristics of the intervention articles, including the author, published year, country where the study was conducted, study aims, study design, characteristics of cancer and participants, theoretical framework, intervention component and dosage and delivery format. In Table 2, we synthesized the included studies' measurement tools and intervention outcomes.

Quality Assessment
We selected and used the Effective Public Health Practice Project (EPHPP) to assess the QR of included articles in this review. The EPHPP is a universal and comprehensive scale to examine bias in aspects of selection, design, confounders, data collection methods and withdrawals and dropouts for a series of study designs [21]. In addition, the EPHPP is regarded as a useful tool for systematic reviews that focus on intervention studies [22]. Table 3 provides detailed information of this assessment tool and the QR of included studies.
Notably, two trained reviewers independently extracted and synthesized data, and assessed the QR of the included studies. Discrepancies were resolved by the two reviewers through discussion until a consensus opinion was reached. -To assess feasibility and acceptability of the program; -To explore improved areas of participants after completing the program.

Single group study
Breast cancer (stages 0-3); All eligible participants regardless of level of relationship and individual functioning before intervention; Intervention Group (IG): 16 dyads.
Reciprocal self-disclosure: (1) disclosure of own and partner's preference and experience through answering a series of topics (including cancer). Partner responsiveness: (2) review and discuss own turning-toward and turning-away behaviors.

Relationship engagement:
(1) communicate individual and relationship strengths; (2) think metaphorically about cancer and create a visual representation of cancer to fortify sense of "we-ness"; (3) watch video to learn sharing self-concerns and understand partner's perspective;

Relationship engagement:
(1) discuss each physically pleasurable shared time and practice sensate focus exercise; (2) remaining contents were the same as the one (Fergus et al. [23]); six modules (once weekly) within eight weeks. Intervention contents and dosage were the same as the one (Fergus et al. [24]).
Reciprocal self-disclosure: (1) express own vulnerable feelings. Partner responsiveness: (1) patiently listen to partner's disclosure and response with empathy and validation.

Relationship engagement:
(1) highlight individual and relationship strengths; (2) communicate supportively to understand problems rather than transitioning too quickly to problem-solving communication; (3) identify frequency of received and preferred support through completing support in intimate relationship rating scale (SIRRS); (4) review supportive communication skills and discuss transition to survivorship phase.
Reciprocal self-disclosure: (1) improve couple's ability to comfortably share their cancer-related thoughts and feelings. Partner responsiveness: (1) improve mutual understanding.

Relationship engagement:
(1) improve constructive communication about cancer-related concerns, mutual support, and emotional intimacy; (2) complete in-session skill practice and home assignments; five sessions (each 90 min) within 8 weeks.
Australia [28] -To explore feasibility and acceptability of patients dignity inventory couple interview (PDI-CI) intervention; -To determine effect of PDI-CI on couples.
Single group study Cancer (advanced); All eligible participants regardless of level of relationship and individual functioning before intervention; IG: 9 dyads.

Not reported
Reciprocal self-disclosure: (1) individually complete PDI based on their own perception of patient's situation.

Relationship engagement:
(1) review results and identify concurrence and discordance with the psychologist, facilitate matched communication; one session (60 min).
-Psychologist; -In person. -To explore efficacy of partner-assisted emotional disclosure intervention comparing with providing cancer-related information.

Not reported
Unidirectional self-disclosure: (1) patients list their cancer-related concerns and disclose their own cancer-related events and feelings in as much detail as possible; Unidirectional partner responsiveness: (1) partners reflectively listen to patients' disclosure, try to understand cancer experience in patients' place, and avoid problem-solving, reassurance or advice giving; (2) identify helpful partner responses.

Relationship-compromising behaviors:
(1) identify unhelpful partner responses; four sessions (the first 75 min, the last three 45 min each) within four to eight weeks.
-Social worker or psychologist; -In person.

Not reported
Reciprocal self-disclosure: (1) express cancer-related experience, cognition and emotions; (2) patients and partners express their own bridges and barriers to communicate with the other separately. Partner responsiveness: (1) listen supportively to partner's disclosure and gain new insights into their experience.

Relationship engagement:
(1) discuss bridges and barriers as a couple; (2) co-construct a relationship line about up or down times, plan for their life post-cancer and share and discuss the relationship line with other couples; 2-session group intervention: two sessions (each four hours); 1-session group intervention: one session (four hours) which was eliminated part of "bridges and barriers".
Reciprocal self-disclosure: (1) disclosure of cancer experience and sources of distress and create a list of cancer concerns; (2) express support needs. Partner responsiveness: (1) respond with empathy and validation.

Relationship engagement:
(1) learn problem-solving model; (2) practice speaking, listening and problem-solving skill with a cancer concern during in-session exercise and home assignment, and practice sensate focus exercise at home; (3) review skills and challenges, and troubleshoot future issues; six sessions (each 90 min) and one booster call (30-45 min).
-Psychologist or social worker; -In person and telephone.

Not reported
Reciprocal self-disclosure: (1) share thoughts and feelings about a series of cancer-related topics. Partner responsiveness: (1) accept and affirm partners' feelings and perspectives.

Relationship engagement:
(1) try to make a cancer-related decision through communication skills; (2) review progress during treatment and identify future issues related to communicating about cancer; six sessions (each 60 min).
Porter et al.
Single group study GI cancer (advanced); Ones who presented holding back pattern before intervention; IG: 12 dyads.
Not reported Intervention contents and dosage were the same as the one (Porter et al. [34]).

Not reported
Reciprocal self-disclosure: (1) express experience, thoughts and feelings sincerely through a cancer-related topic. Partner responsiveness: (1) reflectively listen to partner's emotional disclosure, accept partner's feelings and avoid giving advice quickly.

Relationship engagement:
(1) practice communication skills in each session; (2) try to make a cancer-related decision through communication skills; (3) write and read the letter of thanks for each other, review progress and discuss coping strategies for future cancer-related communication problems.

Relationship-compromising behaviors:
(1) point out couples' unreasonable communication behaviors; five sessions (each 60 min) within five to seven weeks.
-Psychologist and nursing postgraduate; -In person.

Acceptability
-Couples satisfied with the program, while females reported higher satisfaction than males (Cohen's d = 0.42, p = 0.01); -The favorable parts were varied (e.g., "the variety of activity", "the role playing" or "facing cancer as a unified front"), while limitations were lacking in-person contact, and novel learning.
Within-group comparison of pre-post change -Couples started to communicate with each other again, learned communication skills, spent quality time together, obtained insight into their relationship and perceived that they were "in this together".       Between-group comparisons of preand post-change -CPs and SCs reported more improvement in intimacy and relationship quality in IG than CG only when CPs had high holding back from discussing cancer-related concerns before intervention.

Not reported
Within-group comparison of pre-post change -CPs rarely directly talked about their emotion; -SCs generally provided instrumental support rather than responding to the CPs' emotion; -SCs were more possible to respond empathically when CPs were more expressive; -Empathic SCs were less likely to criticize their CPs; -CPs were rated as more expressive in IG when they reported lower relationship quality, higher partner avoidance, and higher holding back before intervention. Primary outcome measures -Interviews; During the intervention.

Not reported
Within-group comparison of preand post-change -CPs and SCs discussed to attempt to maintain normal relationship with each other rather than being in the role of "patient" and "caregiver"; -CPs and SCs shared mutual understanding and addressed conflicts; -CPs desired to provide support to their SCs; -CPs and SCs discussed symptom-related emotional and practical considerations and made likely future treatment decisions; -CPs and SCs discussed death-related topics; -Couples highlighted importance of couple relationship for each other.

Process of Study Selection
A total of 979 articles were found in eight databases and after additional manual searching, and 489 duplicates were excluded using EndNote 20. The titles and abstracts of the remaining 490 articles were reviewed based on inclusion and exclusion criteria. Of the 490 studies, 438 studies were excluded and 52 articles remained to be reviewed for the full-text component. Finally, 14 studies were included in this review for analysis. Figure 1 shows detailed information on the selection process. The most common reasons why studies were removed were that target population of studies were not couples, articles were not intervention studies or interventions did not focus on couple-based communication components. Table 3 illustrates the quality assessment of the involved articles. Three, seven, and four articles were rated as "strong", "moderate" and "weak", respectively. The rating of "weak" was mainly owing to a low recruitment rate of the target population, not using a blinding strategy and controlling confounders, and using tools without evidence of validity or reliability. Although the QR of the studies was varied, we included all of them because they basically completed their study aims and met the selection criteria of this review.

Characteristics of Intervention
Most intervention studies included in this review were carried out in western countries, including the USA (n = 9, 64.3%), Canada (n = 3, 21.4%) and Australia (n = 1, 7.1%), while one was from China (7.1%). The designs of the intervention studies were randomized controlled trial (RCT) (n = 8, 57.1%), single-group study (n = 5, 35.7%) and three-group study (n = 1, 7.1%). The single-group study included a pre-post single-arm study and interviews for participants from the treatment arm only.

Characteristics of Participants
Of the 14 articles involved in this review, 13 focused on a single type of cancer with varied stages, including gastrointestinal cancer (n = 5, 41.7%), breast cancer (n = 4, 28.6%), prostate cancer (n = 2, 14.3%), head and neck cancer (n = 1, 7.1%) and gastric cancer (n = 1, 7.1%), while one targeted any type of cancer with an advanced stage (n = 1, 7.1%). Notably, of the 13 articles testing a single type of cancer, seven focused on non-genderspecific cancer, while six targeted gender-specific cancer (e.g., breast or prostate cancer). The articles' sample sizes differed, ranging from nine dyads to 237 dyads. Across the target population, 10 studies focused on participants regardless of the level of their relationship or individual functioning before the intervention [23][24][25][26][27][28][29][30][31][32], while four studies recruited those who had distress or cancer-related communication problems before the intervention [33][34][35][36]. In addition, ethnicity and education level of the majority of the participants in the included studies were Caucasian and college or higher, respectively.

Theoretical Framework of the Interventions
Six kinds of theoretical framework were used to instruct the research included in this review, including the developmental model of couple adaptation to illness [23][24][25], the theory of health relationship functioning [23][24][25], models of dyadic coping [23], systemicconstructive metatheory [24,25], social-cognitive processing theory [26] and RIM [26,27,33]. Among these theoretical frameworks, RIM was the most frequently demonstrated and was used to guide two kinds of intervention programs. Most included studies did not use a specific theoretical framework to guide their intervention, or some of those using a theoretical framework did not demonstrate how the framework guided their intervention in detail.

Intervention Dosage
The number of modules/sessions in the 14 intervention studies was varied, ranging from one to six, with an average of 3.5. Of the 14 articles included in this review, 10 articles illustrated that each session generally lasted 60 to 90 min [26,27,[29][30][31][32][33][34][35][36]. The length of each online module of three articles was chosen by the participants [23][24][25]. One study spent four hours in each session [28]. In addition, for the duration of the intervention, eight studies reported that the duration generally ranged from four to eight weeks, and the average value was six weeks [23][24][25]27,[29][30][31]36]. Five articles did not mention the duration [26,[32][33][34][35], while one study had only one session constructed during the first counselling session [28]. With regard to the follow-up periods, four studies had single follow-up timepoints, including eight weeks [30], three months [24,32] and six months after intervention [26], while another adopted multiple follow-up periods [33].

Feasibility
The feasibility of the interventions was assessed in 10 studies. Eight studies reported generally low recruitment rates that were less than 60% [23,[26][27][28][29][32][33][34], while two studies had relatively high recruitment rates above 60% (including one online intervention program conducted in Canada [24] and one in-person intervention conducted in China [36]). The reasons why participants refused to take part in the intervention mainly involved them being too busy, too far away, too sick, lacking interest or lacking need. It is worth noting that studies showed relatively high retention rates (ranging from 75% to 100%), and relatively high completion rates, which were demonstrated by the fact that 62.5-100% of participants completed most modules/sessions, 89-93.2% of participants completed relevant practice and 73-90% of participants completed outcome measurements.

Acceptability
Seven studies examined the acceptability of their interventions. CPs and their SCs were highly satisfied with interventions [23,25,26,28,33,34] or viewed the intervention as quite successful [27]. Particularly, for one online intervention program, females reported higher satisfaction than males (Cohen's d = 0.42, p = 0.01) [25]. The favorable aspects of programs for participants encompassed communicating their own feelings [26], listening skill [34] and "the variety of activity", "the role playing" or "facing cancer as a unified front" [24]. In addition, limitations were also pointed out, including that the online intervention program lacked in-person contact [23,25], the questionnaire was too long [26] and the intervention lacked profoundness [23,28]. Approximately 60% of participants in the online intervention programs regarded website usage as convenient [23,25]. Another study reported that the rates of participants' preferences for delivery formats were 77%, 20%, and 3% for video conference, in-person and telephone interventions, respectively [34].

Intervention Outcomes
All 14 studies measured relationship functioning, such as mutual communication, intimacy, and RS, and/or individual functioning, including anxiety, depression, psychological distress, cancer concerns, psychological adjustment and QOL for CPs and their SCs. The detailed information of intervention outcomes can be found in Table 2. In this review, the outcomes of relationship and individual functioning were separately analyzed according to within-and between-group comparisons of pre-post change.

Effects on Relationship Functioning for Within-Group Comparisons of Pre-Post Change
Eight studies reported that CPs and/or their SCs experienced a significant improvement in their mutual communication [23][24][25]28,31,35,36], intimacy [23,25,28] and RS [23,25,33,35,36] in the intervention group (IG). For these significant outcomes, it was found that the interventions had a short-term (up to five weeks) impact on relationship functioning [33]. In addition, three studies reported that CPs and/or their SCs with distress or communication problems (e.g., holding back) before the intervention seemed to be more able to benefit from the interventions [33,35,36]. On the contrary, one online intervention [24] and another offline group intervention [32], which viewed RS as the primary outcome, did not improve the RS of either CPs or their SCs during multiple follow-up periods.

Effects on Relationship Functioning for Between-Group Comparisons of Pre-Post Change
Gremore et al. [26] reported that CPs and their SCs, regardless of their level of relationship and individual functioning before the intervention, experienced more improvement in intimacy and RS in the IG than the control group (CG), even at six months follow-up. Another five studies showed that the intervention positively impacted relationship functioning only when CPs and/or their SCs had a low or negative functioning before the intervention [27,29,30,34,36]. Of the five studies, one demonstrated that CPs reported higher scores of self-disclosure (t = 3.50, p < 0.001), perceived partner disclosure (t = 3.27, p = 0.002) and perceived partner responsiveness (t = 3.26, p = 0.034), while their SCs had higher scores of mutual constructive communication (t = 3.70, p < 0.001), intimacy (t = 3.42, p = 0.001) and RS (t = 3.94, p < 0.001) in the IG than the CG only when these variables were low before the intervention [27]. The remaining four studies illustrated that CPs and/or their SCs obtained a greater increase in communication, intimacy, and/or RS in IG than CG only when at least one of them had communication problems (e.g., holding back) before interven-tion [29,30,34,36]. Notably, the offline group intervention had no any significantly different influence on couples' RS between treatment terms regardless of at post-treatment or three months follow-up [32]. Unexpectedly, Manne, et al. [27] found that the intervention may diminish level of self-disclosure of CPs, and intimacy and RS of their SCs when these variables were high before the intervention.

Effects on Individual Functioning for Within-Group Comparisons of Pre-Post Change
Two studies explored the effect of interventions on individual functioning in withingroup comparisons of pre-post change [24,33]. One online intervention program reported that couples perceived less anxiety at follow-up periods in both the IG and the CG [24]. The other study, which focused on participants with distress before the intervention, reported that CPs and their SCs experienced an improvement in psychological distress, depression, cancer concerns and psychological adjustment during varied follow-up periods regardless of treatment arms [33].

Effects on Individual Functioning for Between-Group Comparisons of Pre-Post Change
Seven studies investigated the different influence of treatment arms on psychological adaptation and/or QOL. Greater improvements in post-traumatic stress, anxiety, depression and psychological adjustment at both post-treatment and six-month follow-up were reported by CPs and their SCs in the IG than the CG [26]. Moreover, Manne et al. [27] demonstrated that couples had fewer cancer-related concerns (t = −2.34, p = 0.022) or cancer-specific distress (t = −2.31, p = 0.024) in the IG than the CG, but only when these variables were high before the intervention. The offline group intervention found that CPs had more visible improvement in psychological adjustment in the two-session group than the other two treatment arms, but SCs did not [32]. In addition, Porter et al. [34] illustrated that couples seemed to gain more improvement in post-traumatic growth and self-efficacy in the CG than the IG, when the CG aimed to provide cancer-related information over six sessions [34]. Two studies that were designed to promote CPs' unidirectional self-disclosure to their SCs reported no significant change in the psychological distress of couples between treatment arms, regardless of follow-up periods [29,30]. As for QOL, CPs had a greater increase in QOL at six-month follow-up (Cohen's d = 0.48), while their SCs experienced more improvement at post-treatment (Cohen's d = 0.22) and six-month follow-up (Cohen's d = 0.31) in the IG than the CG.

Discussion
A total of 14 eligible articles were included in this review according to the inclusion and exclusion criteria. Through extracting and synthesizing characteristics and outcomes of these included studies, we have demonstrated the approaches and contents used in couple-based communication interventions, their feasibility and acceptability, and their effect on CP-SC dyads' psychosocial adaptation to cancer. To synoptically discuss the results of these included studies, we have arranged the construction of the Discussion section as four "Ws" (including Who, What, How, and When), efficacy, recommendations and limitations to provide some enlightenment for future research.

Who?-Choosing the Target Population
For CPs and their SCs, the diagnosis of specific types of cancer would induce specific communication needs. For example, CPs with colorectal cancer would like to communicate about ostomy-related issues and changed bowel function with their SCs [37]. There are different communication topics among couples during different cancer stages. Facing early diagnosis, CPs and their SCs may be more concerned with the treatment effect, a healthy diet and physical exercise. Meanwhile, for advanced cancer stages, there will be more difficult topics, such as making the CPs comfortable rather than cured and anticipatory grief of losing loved partners [38]. In addition, half of the articles included in this review targeted a population with gender-specific cancer (breast or prostate cancer). Gender may act as a possible factor impacting adjustment outcomes for CPs and SCs. For instance, in the context of colorectal cancer, studies reported that female CPs and female SCs usually experienced more anxiety, depression, fear of cancer recurrence [39][40][41][42] and less marital satisfaction [43] than their male counterparts. Although a systematic review and meta-analysis conducted by Hagedoorn et al. found that individual levels of distress were attributed more to gender compared with role (CP or SC) [44], focusing on participants with gender-specific cancer still makes it difficult for researchers to distinguish whether gender or role has an impact on adjustment outcomes [45]. In addition, according to intervention outcomes in this review, interventions seemed to be more beneficial for participants who experienced distress or communication problems before the intervention. Notably, one study led by Manne et al. reported that a couple-based communication intervention may reduce participants' level of self-disclosure, intimacy and RS when they had relatively high levels of relationship functioning before the intervention [27]. More research is warranted to explore whether it is necessary to assess couples' relationship or individual functioning before the intervention and select those with distress or communication problems as participants.

What?-Communication Topics
Communication topics mentioned in the included studies consisted of cancer-related experience, thoughts, feelings and relationship issues. Encouraging cancer-related emotional disclosure among CP-SC dyads is a common recommendation in couple-based interventions [46]. Badr reported that there is great variability in topics when couples talk about cancer [17]. Except for emotional disclosure, health-related topics (e.g., symptom management, treatment issues, daily care and prognosis) and relationship topics (e.g., role change, marital relationship and social/family issues) also need to be given more attention [17,47]. For example, talking about symptom management/treatment issues possibly helps couples reassert a sense of control over their cancer situation [17], talking about daily care (e.g., healthy diet and exercise) may promote couples going back to a routine life, while discussing relationship topics may activate support resources from the relationship network [11]. Therefore, it is necessary to emphasize the importance of varied topics for CP-SC dyads' psychosocial adaptation to cancer, including cancer-related emotional disclosure, healthy issues and relationship topics.

How?-Communication Methods
Teaching couples how to effectively communicate with each other was the critical content in these interventions in this review, mainly comprising of improving beneficial reciprocal self-disclosure, partner responsiveness and relationship engagement. In this review, unidirectional self-disclosure and partner responsiveness may have had limitations because it overlooked the process of SCs' self-disclosure and CPs' responsiveness. Manne et al. [48] reported that, regardless of role (CP or SC), self-disclosure and partner disclosure positively promoted the perceived partner responsiveness, and further improved their intimacy, which may be attributed to reciprocal self-disclosure and partner responsiveness being important for improving a couple's members' sense of feeling understood, validated and cared for [11]. In addition, distinguishing and declining relationship-compromising behaviors are similarly important because of their consistent association with negative outcomes [17]. As for ways of expression, except for oral discussion, this review showed that written expression was also used and was effective. With regard to communication channels, some participants regarded themselves as "face-to-face kind of person" [25], while other participants preferred using social media to communicate [47]. Therefore, it is more reasonable to provide varied communication channels to be chosen by participants.

When?-Communication Timing
With regard to intervention dosage, in general, the included studies implemented interventions in 3-4 modules/sessions (each 60-90 min in length) within six weeks, and had up to six-month-long follow-up periods. For some participants with specific situations, such as a low capability to understand, more need to communicate and disruption due to treatment, fittingly lengthening the intervention dosage is necessary. As for when to start the coupled-based communication interventions, previous studies suggested that it is better to start before some important treatment/life timepoints, such as before making a decision about treatment, surgery, the transition to survivorship and end-of-life care [11,47].

Exact Efficacy of Couple-Based Communication Interventions
Generally, the feasibility and acceptability of existing couple-based communication interventions are acceptable. The relatively low recruitment rates may not only have been due to the difficulty of conducting the psychological program, but also obstruction of recruiting CPs and their SCs simultaneously. The relatively high retention and completion rates may suggest that the interventions could greatly meet participants' communication needs, which, in turn, supports the view that it is necessary to carry out this kind of intervention to help couples adjust to cancer better. The effect sizes of interventions were identified as small (Cohen's d = 0.20-0.30), medium (Cohen's d = 0.30-0.60) and large (Cohen's d > 0.80) [26]. In this review, two studies reported small effect sizes, ranging from 0.22 to 0.30 [24,33], which were consistent with the result of a review conducted by Badr and Krebs [19]. Another two studies showed medium [34] to large effect sizes [26], ranging from 0.31 to 1.17. The relatively high effect sizes may be attributed to the clear and singular purpose of couple-based communication interventions, or the fact that they focused on participants with distress or communication problems before the intervention. It is still difficult to conclude what the reasons were for the medium to large effect sizes of interventions due to insufficient studies.
In addition, Badr, et al. suggested that researchers should move beyond a "one size fits all" approach [48,49], explore nuances of couples' communication and develop more efficacious interventions for promoting CP-SC dyads' psychosocial adaptation to cancer [17]. For example, different interactions of gender and role seem to result in varied communication performance during dyadic coping with cancer. As for female patients with male partners, male partners usually initiated cancer-related communication (e.g., treatment options), but most of them avoided communicating emotional reactions [50]. Lyons et al. [51] found that female patients experienced less depression when their male partners engaged in a lower level of protective buffering (e.g., hiding worries or waving patients' worries aside). When it came to male patients with female partners, both of them tended to deny, avoid, refuse and hold back cancer-related discussion, which may be because they wanted to protect each other, or they assumed their partner's situations, thoughts and feelings rather than assessing and resolving their partner's actual emotional problems [50]. Although male patients perceived less depression when their female partners engaged in a high level of protective buffering [51], female partners' desire to gain more emotional reaction and information from male patients needed to be given more attention [50]. According to social role theory [52], the gender difference of communication performance may be attributed to gender stereotypes, that is, men are good at instrumental behaviors while women prefer expressive behaviors. Exploring reasons or motivations for specific communication performance induced by role and gender would be helpful for researchers to better understand this specificity and provide CP-SC dyads with more personalized communication support.

Recommendations for Future Research
According to the characteristics and outcomes of intervention studies included in this review, we give the following recommendations in the hope of helping future relevant interventions: (1) Conducted country: 93% of couple-based communication interventions were carried out in western countries, which reminds us that future interventions should be investigated in other regions, such as Asia; (2) Target population: more research is needed to explore the necessity of screening distress or communication difficulties of CPs and/or SCs before the intervention; (3) Study design: more longitudinal RCTs with large enough sample sizes are needed to explore the efficacy of couple-based communication interventions; (4) Theoretical framework: a detailed combination of the theoretical framework and intervention design adopted should be reported in order to support future replicative studies; (5) Intervention contents: focusing on varied topics, teaching couples to improve mutual communication behaviors, activating couples' relationship resources and considering specificity of gender and role during the intervention may promote the efficacy of couple-based communication interventions; (6) Intervention delivery: to meet the CP-SC dyads' different preferences for delivery format, a combination of in-person and web-based intervention delivery is recommended to be adopted.

Limitations of this Review
It is necessary to point out that there are several limitations in this review. First, only studies published in English or Chinese were searched for and included in this review. Potential studies in other languages or published forms (e.g., conference proceedings, dissertation or editorials) may have been overlooked. Second, the heterogeneity of the included studies, such as differences in the type and stage of cancer studied, the study design, and the varied measurement tools used, may have impacted the generalization of outcomes. Third, we should be cautious to interpret the effect sizes of the interventions because they had small sample sizes and different baseline functioning. In sum, more research is needed to improve understanding and develop more efficacious couple-based communication interventions.

Conclusions
CP-SC dyads may experience communication difficulties while coping with cancer. Studies included in this review reported that a couple-based communication intervention improved CP-SC dyads' relationships and individual functioning, that is, to improve couples' psychosocial adaptation to cancer. More research is warranted to understand and develop more efficacious couple-based communication interventions, such as exploring the impact of the interaction of gender and role on dyads' mutual communication behaviors.

Conflicts of Interest:
The authors declare no conflict of interest.