Given that sexuality after cancer diagnosis remains an important part of maintaining quality of life,1-3 sexual oncology is gaining recognition as an essential area requiring attention in clinical practice and health research. Sexual health is a common concern for oncology patients, especially those with breast, gynecologic, and genitourinary cancers.4 Patients surgically treated for urogenital cancers and colorectal cancers, as well as patients treated with radiation to the pelvis or those undergoing hormonal treatments, experience significant physiologic dysfunction. Women experience vaginal dryness, occasional vaginal shortening, and consequent dyspareunia. Men experience erectile dysfunction and, occasionally, orgasmic disorders.5-7

These problems cause patient and partner suffering as well as difficulties engaging in sexual recovery.4,8 As these difficulties can be severe and persistent over time, psychological problems such as decreased libido, alterations in body image, and anxiety related to sexual performance are major concerns.1 Congruent with conceptions of masculinity, men are particularly affected, and they often enact psychological defenses against related vulnerabilities.9 Socially, cancer patients report difficulty maintaining sexual roles, emotional distancing from the partner, and changes in the partner's level of sexual interest.1

However, it is insufficient to simply provide prostate cancer patients with erectile function aids.9 Reviews of the literature10,11 have suggested that cancer patients benefit from information about their disease and corresponding treatments; however, it has been demonstrated that cancer patients are not given sufficient information on the effect of their disease and treatment on sexuality.1,11,12

Psychoeducational interventions have emerged in practice as adjunctive treatments for cancer in which patients and families are struggling with various life challenges, as they are based on strengths and focused on the present.13 Psychoeducation has been defined as a professionally delivered psychotherapeutic and educational intervention that utilizes a more holistic and competence-based approach, stressing health, collaboration, coping, and empowerment.14,15

Psychoeducational intervention has been demonstrated to reduce depression and anxiety and improve adjustment to illness and coping styles,16,17 compliance with vaginal dilation,18 and coital activities and sexual function.19,20 In psychoeducational interventions, participants are encouraged to discuss self-care strategies for treatment and its adverse effects. Intervention providers should provide information on addressing the effects of treatment and aim to correct any misconceptions about unsubstantiated self-care strategies.21 Furthermore, professionally facilitated groups appear to provide additional benefits to participants by providing evidence-based information and promoting higher rates of emotional disclosure from participants.22 Nevertheless, sexual health is still inadequately addressed after cancer diagnosis.23-25

The global incidence of cancer is approximately 12.7 million across more than 200 kinds of cancers.26 Given the high prevalence of prostate and breast cancer, large populations of cancer patients experience sexual problems. Furthermore, sexual health has been defined as a state of physical, emotional, mental, and social well-being in relation to sexuality. It is a complex integration of idea, emotion, and behavior.27 Therefore, past studies have varied considerably in gender, type of cancer, and outcomes measured. Most studies assessing psychotherapeutic interventions for cancer patients vary considerably in terms of participants assessed and research design (eg, different intervention delivery and methods).16,17,19,20 To our knowledge, no practical guidelines for sexual intervention exist, perhaps because of the multifaceted nature of sexual intervention among cancer patients.

Although individual-based interventions have been demonstrated as more effective than group-based therapies,28 some studies have suggested that a group setting is most effective for sexual intervention.11 Recently, an Internet-based sexual counseling program was introduced as a cost-effective alternative approach,29 and telephone-based intervention has been demonstrated as an effective intervention.30,31

To date, 2 systematic reviews have been conducted on gynecologic cancer. One reviewed gynecologic cancer survivors' experiences of a broad range of sexual concerns after diagnosis,1 and the other identified interventions for psychosexual dysfunction, but was limited in that the studies were of poor methodological quality.32 For healthy women, Shepherd et al33 conducted a systematic review of the effectiveness of a health education intervention regarding sexual risk reduction. A review of studies investigating the effects of sexual rehabilitation among prostate cancer patients/survivors was also conducted.34 In terms of intervention type, a review of couples-based intervention35 was conducted. However, some of the studies did not focus on psychosocial intervention. They summarized the results, but did not indicate the effectiveness of the intervention according to the variables and effect size.

A meta-analysis is a formal statistical analysis of data from various subgroups of preexisting studies. It is conducted to obtain an overall estimate of the effectiveness of the selected subject/intervention. Therefore, it is necessary to estimate the effect size for psychoeducational interventions and compare their effectiveness across variables that affect intervention outcomes. The identification of effective interventions to address sexual concerns would be useful for oncologic health professionals. Accordingly, the aims of this study were to determine the effect size for psychoeducational interventions focused on sexuality and compare effect sizes according to intervention outcomes and characteristics.

Methods

Data Sources and Searches

In order to identify relevant studies, we explored quantitative evidence for the effects of sexual intervention for cancer patients or partners by using the electronic databases EBSCO, MEDLINE, ScienceDirect, and PQDT. We used the keywords "sexual intervention," "cancer," and "oncology" (1950 to September 2013).

Several potential studies were identified, and therefore, we restricted the search to the following inclusion criteria. First, we searched for studies that evaluated a psychoeducational intervention combining education with elements of behavioral or cognitive therapies to assist cancer patients in preventing or treating psychosexual problems after diagnosis. Second, studies must have used an experimental, quasi-experimental, or 1-group design. Finally, we focused on international literature written in English in order to gain insights into the effectiveness of sexual intervention for cancer patients or partners. Studies were excluded if the intervention did not focus on sexuality or if relevant quantitative outcomes were not reported. Intervention studies on combining medical therapies, such as phosphodiesterase-5 inhibitors or intracavernosal injections, with psychosocial intervention were also excluded. Because empirical data were required in order to compute an effect size, nonempirical studies and literature reviews were excluded.

We identified 537 potentially relevant articles, the titles and abstracts of which were scanned in order to identify potentially relevant studies for which full reports were subsequently obtained. Thirty-seven articles remained, and 5 articles were added through a manual search. Among them, we excluded 27 ineligible articles, resulting in a total of 15 (Figure 1).

Data Extraction

Data were extracted from studies that met our inclusion criteria. The methodological quality was assessed using RoBANS (Risk of Bias Assessment Tool for Nonrandomized Studies) and RoB (Risk of Bias) for randomized studies. RoBANS contains 6 domains: participant selection, confounding variables, measurement of exposure, blinding of the outcome assessments, incomplete outcome data, and selective outcome reporting.36 RoB also has 6 domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, and selective reporting.37 Two authors assessed the RoB for all eligible studies and rated eligible studies as having either high risk, low risk, or uncertain. Three or fewer domains were evaluated as high risk. Ten studies were rated as high risk for potential bias in the blinding of outcome assessments. Five studies among the 7 nonrandomized studies had a risk of measurement bias. Six studies among the 8 randomized studies had a risk of bias in blinding participants and personnel.

The 15 studies included in this analysis had different outcomes and study characteristics. The study outcomes and characteristics were coded to reflect potential moderating variables for the effect of psychoeducational intervention on sexual health. The coded characteristics were as follows. The concept of sexuality refers to the totality of being a person and being involved in another person's cognitive, affective, psychomotor, and social abilities and values.38-40 According to these defined characteristics of sexuality, the intervention outcomes in each study were classified as physical, psychological, cognitive, social, or compliance aspects. The outcomes, including sexual activity, comfort during sexual activities, sexual function, physical function, and physical symptom level, were coded as physical aspects. Outcome measures, such as anxiety, depression, mental health, and stress, were classified as psychological aspects. Variables such as body image, concern regarding sexual life, knowledge, sexual need, perceived positive experience and outcome, and sexual satisfaction were coded as cognitive aspects. Outcome measures involving relationships, including dyadic adjustment and marital intimacy, sexual role, and social support, were coded as social aspects. The numbers of the effect sizes are shown in Tables 4 and 5. Intervention characteristics were coded in terms of diagnosis or subjects, intervention providers, intervention method, random assignment, group or individual therapy, couple-based therapy, and intervention setting.

Data were extracted by 2 researchers using a standard data collection form that included information regarding effect size calculations as well as study and report characteristics. Differences between coders were resolved by discussion until a consensus was achieved.

Data Synthesis and Statistical Analysis

Fixed-effects models assume that the primary studies have a common effect size. By contrast, random-effects models attempt to estimate the distribution in the mean effect size, assuming that each primary study has a different population.41 Because the homogeneity assumption was violated, we used random-effects models to pool weighted effect sizes for the main effect and subgroup analysis. The subgroup analyses were conducted based on intervention outcomes and characteristics. Furthermore, the possibility of publication bias was assessed by checking symmetry of the funnel plot. In cases of asymmetry, the trim-and-fill method was used to obtain adjusted effect size estimates.

Two types of data format, 1-group studies (pretest-posttest) and control group (pretest-posttest), were used to calculate the effect size. The effect sizes and variances of each study were calculated by using the appropriate formulas. The findings were then evaluated based on Cohen's42 classification of effect sizes, where 0.2 is considered small, 0.5 medium, and 0.8 large.

Results

Study Characteristics

A variety of intervention types were found across the 15 studies, as shown in Table 1. Seven studies (46.6%) used random assignment, whereas 5 studies (33.3%) did not or did not feature a control group. Six (40%) assessed breast cancer patients, 3 (20.0%) assessed cervical and endometrial cancer patients, 3 (20.0%) focused on prostate cancer, and 3 (20.0%) focused on colorectal and various other types of cancers. In terms of intervention provider, 3 studies (20.0%) used registered nurses alone, 2 studies used peer providers, and the remainder used therapists or psychologists.

Table 1 Characteristics of Studies Included in the Analysis (n = 15)

Overall Analysis

The meta-analysis provided 133 effect sizes from 15 primary studies. As shown in Table 2, heterogeneity was observed when these studies were combined (Q = 698.64, P < .001). The analysis revealed significant improvements after intervention, with a pooled random-effects standardized mean difference of 0.75 (Table 3), which indicates a medium-to-large effect. The forest plot identified the precision of each study according to the size of the confidence interval (Figure 2).

Table 2 Results of the Homogeneity Test

Table 3 Overall Result of the Meta-analysis Using a Random-Effects Model

The funnel plot appeared to be symmetric, indicating minimal likelihood of publication bias. In the trim-and-fill analysis, no study imputed on the left of the mean, and although the effect size increased slightly (adjusted value, 0.93) for 3 trimmed studies to the right of the mean, this did not change the conclusions about whether the interventions were effective or the extent to which they were effective.41

Subgroup Analysis

Subgroup analysis within areas was performed according to intervention outcomes and study characteristics. The number of studies and effect size included in each analysis varied based on the specific outcome studied and its characteristics. Given that homogeneity tests demonstrated heterogeneity among subgroups, we performed a subgroup analysis using a random-effects model.43

Subgroup Analysis According to Intervention Outcomes

The results of the subgroup analysis according to study outcomes using the random-effects model are shown in Table 4. In terms of study outcomes, the effect sizes were as follows: compliance (2.40), cognitive aspect (1.29), psychological aspect (0.83), social aspect (0.45), and physical aspect (0.43). With regard to physical aspects, the effect size for sexual activities was the highest (0.63), followed by physical function (0.55), sexual function (0.44), physical symptom (0.27), and comfort during sexual activities (0.21). Regarding psychological outcomes, anxiety (2.12) and depression (1.41) demonstrated strong effects, mental health (0.78) showed a medium effect, and stress (0.41) a low effect size. In terms of cognitive aspect, the effect sizes were as follows: sexual need or concern (2.00), knowledge (1.65), body image (0.85), perceived positive experience or outcome (0.77), and sexual satisfaction (0.63). In the social domain of sexuality, social support (0.90) had the highest outcome effect size, followed by communication (0.68), sexual role (0.35), and relationship (0.29).

Table 4 Effect Sizes by Subgroup According to Intervention Outcomes

Subgroup Analysis According to Intervention Characteristics

Results for the subgroup analysis of intervention characteristics are shown in Table 5. With respect to study participants, the effect size of patients (0.76) was higher than that for the partner (0.45). Female participants (0.74) had a higher effect size than did the male participants (0.27).

Table 5 Effect Sizes by Subgroup According to Intervention Characteristics

The effect sizes for studies according to diagnosis or subject were as follows: cervical and endometrial cancer (1.04), breast cancer (0.74), colorectal cancer (0.59), adolescent and young adults (0.74), and prostate cancer (0.31). Regarding intervention providers, the effect sizes were as follows: psychologist and registered nurse (2.38), registered nurse alone (2.22), psychologist or therapist (0.60), multidisciplinary approach (0.45), and peer (0.27). In terms of intervention delivery method, the effect size of face-to-face combined with telephone or Internet (1.04) was higher than for face-to-face (0.62) and telephone (0.58).

Studies using nonrandom assignment had a larger effect size (2.25) than those using random assignment (0.48). The effect size for studies with a high risk of bias in blinding outcome assessments (0.64) was low compared with studies with a low risk or uncertain bias (0.84).

Individual-based interventions (0.85) were more efficacious in improving study outcomes than group (0.50) and group combined with individual intervention (0.45). The effect size for the couple-based intervention (0.38) was lower than without partner (0.82) and that of mixed intervention (1.81). Clinical settings for intervention delivery (0.93) had a larger effect size than home (0.46) and clinic combined with home (0.45).

Metaregression by the Mean Age of Participants and Intervention Rate

A metaregression analysis was conducted to determine whether the mean age of participants and intervention rate had any effect on the sexual intervention outcome. The slope of the metaregression was negative for mean age of participants and positive for intervention rate, and both were statistically significant (Table 6).

Table 6 Fixed-Effects Regression Analysis by Age of Subject and Intervention Rate

Discussion

The aims of this study were to determine the effect size of psychoeducational interventions for sexuality among cancer patients and to compare the effect sizes according to relevant characteristics. These results could have implications for psychoeducational interventions for sexual health. The meta-analysis revealed a medium-to-large effect size for psychoeducational intervention focused on sexuality. This indicates that psychoeducational therapy offers numerous benefits to clinical practice and research intervention focused on the sexuality.

It was demonstrated that sexuality-based psychoeducational interventions after cancer diagnosis were effective for compliance as well as cognitive or psychological outcomes. Providing information accompanied by opportunities for communication about their feelings after cancer diagnosis could be effective in reducing depression and anxiety, as well as increasing adjustment to illness and coping.16,21,44 In this review, physical and social outcomes had a relatively small effect size. Healthcare professionals should also consider these outcomes when caring for cancer survivors. There is a need to develop an approach toward improving physical and social outcomes among cancer patients or survivors.

The studies included a heterogeneous sample of cancer patients, both in terms of intervention characteristics and outcome. It is a common criticism of meta-analysis that different kinds of studies are combined together for a comparison, although the effect size has the same meaning, regardless of the study design, from a statistical perspective.41

To overcome the variability between studies and their multifaceted analyses of sexual intervention among cancer patients, a sensitivity analysis was performed to evaluate subgroups of the studies sharing study characteristics.

Female participants showed a larger effect size than male participants. Furthermore, effect sizes for studies of prostate cancer patients were lower than the effect sizes for interventions for other cancer patients were. There is consistent evidence that psychological interventions specifically improve symptom severity among women, but not men, in the meta-analysis.45 Although prostate cancer is generally diagnosed in its early stages, many men experienced sexual difficulty as a result of adverse effects and treatments.5-7 There exist several barriers to receiving care regarding sexuality among men, such as men's cultural view of masculinity, which is subject to psychosocial adaptation among prostate cancer men.46 Furthermore, some African American men do not trust the health system and had poor relationships with healthcare professionals during prostate cancer treatment.12

In sexual rehabilitation treatment for prostate cancer patients, some men rejected erectile dysfunction treatments because of anticipated disappointments in the pursuit of a sexual relationship.9 In addition, health professionals must endeavor to change traditional male beliefs regarding the importance of erections in women's sexual satisfaction.4

In terms of the developmental stage, adolescent and young cancer patients demonstrated a medium effect size in psychoeducational interventions for sexuality. Counseling on fertility preservation is a high priority among newly diagnosed cancer patients.32 The options to preserve fertility in women and men undergoing treatment for cancer should be discussed at this age.33 In a qualitative study of young cancer survivors, conveying information regarding the effect of treatment on fertility and fertility preservation options was suggested, even if concerns are not expressed at diagnosis.47 Considering that adolescent cancer survivors are at risk for sexual impairment in adulthood and that they reported high levels of distress compared with the normative sample,48 psychoeducational interventions would be effective for these populations.

Despite prior evidence for the significant impact of partners, this meta-analysis demonstrated that couple-based interventions had a small effect size. The literature suggested that couple-based interventions could lead to improvements in dyadic-level adjustment.49,50 Couple-based interventions were particularly efficacious in promoting intercommunication, sexual adjustment, and functional relationships with an increased understanding of the cancer diagnosis within couples.3,45,51 However, there were disagreements with regard to couple-based intervention that corresponded with different outcome measures. Fruhauf et al45 suggested that couple-based interventions are useful for decreasing symptom severity, but not effective for sexual satisfaction. However, these results should be interpreted with caution, as couple-based interventions in sexuality are so undeveloped that it is difficult to draw conclusions about how to use the results, and their use should be limited. Therefore, researchers and health professionals should endeavor to develop science-based interventions that help couples toward positive sexual outcomes.

Individual-based interventions were more effective in improving outcomes than group approaches and group combined with individual intervention. There was consistent evidence to suggest that individual-based cognitive behavioral therapy (CBT) interventions52,53 and individual psychosocial interventions for breast cancer patients54 were more efficacious than group-based interventions. When discussing sexuality with a healthcare professional, cancer patients indicated that they would have preferred an individual setting.28 However, the literature review regarding prostate cancer support groups suggested that a group setting was beneficial for providing information about their disease.10,11,55,56 Molton et al55 suggested that patients with higher levels of interpersonal sensitivity were more likely to perceive sexual dysfunction and had more effective outcomes and greater gains after radical prostatectomy. Therefore, researchers and health professionals should consider individual intervention according to the participants' preferences.

With regard to intervention providers, the peer approach to intervention showed a small effect size. Peer support had economic advantages,57 high satisfaction, and psychosocial support benefits among participants.58 However, Chambers et al30 reported that there were several challenges to the peer approach, namely, unfamiliarity and inherent difficulties in establishing rapport and understanding individual differences between couples and patients. In this study, the registered nurse only and team approach including the registered nurse had the highest effect sizes. It has been reported that nurse-led interventions for cancer patients' psychosocial cancer care are appropriate and effective.59 The nurses could play an important role in providing interventions that can be routinely integrated into care because of their close contact with patients.60,61 Sexual healthcare is recognized as an integral component of advanced oncology nursing,62 and the role of oncology nurses in sexual health counseling is an expected component in the United States and Korea.62,63 Some research has also supported professionally facilitated interventions because of benefits in receiving evidence-based information.22 Flynn et al32 suggested that interventions regarding the sexuality of women with gynecologic cancer should be delivered by a multidisciplinary team. Multidisciplinary teams consist of radiation oncologists, clinical nurses, behavioral scientists, psychologists, and radiation therapists with significant cancer-related experience, who are able to provide comments based on their clinical expertise and knowledge of cancer research.21,32,64

Recently, a training program was developed focusing on reproductive health, sexuality, cancer and parenting, barriers and facilitators, social issues, practical issues, and applications for oncology nurses.65 Considering that sexual healthcare training is associated with a more positive attitude toward sexual healthcare,66 training in sexuality and sexual issues should be implemented as part of the training of healthcare professionals.67

Face-to-face intervention combined with telephone or the Internet demonstrated a larger effect size than face-to-face and telephone interventions independently. Recently, it was demonstrated that an Internet-based sexual counseling program may be more cost-effective as it requires less of the therapist's time to respond to e-mails than to conduct traditional therapy sessions.29 Because of the barriers to delivering face-to-face counseling intervention, such as the lack of counseling professionals experienced both in treating sexual dysfunction and in psycho-oncology, and poor insurance coverage for sex therapy, such an intervention might be a viable option.68 However, given the weaknesses and difficulties inherent in using Internet- or telephone-based interventions independently, the combination of the face-to-face and online or telephone intervention would likely have more positive outcomes because of tailored psychological interventions.

A metaregression analysis was conducted to determine whether the mean age of participants and intervention rate had any effect on sexual intervention outcome. Younger participants were more likely to experience positive intervention outcomes. There were some studies, however, that reported differences in intervention success according to age. Younger women under treatment of gynecologic cancer might be at greater risk for psychological difficulties because of early menopause or loss of fertility,69-71 but older patients often experience greater physical sexual dysfunction.72,73 Evidently, age is not always a predictor of intervention success, as age groups experience different issues related to sexual recovery. Perhaps the negative association between age and intervention success can be explained by the fact that interventions tend to be more successful in the psychological domains than in the physical domains.

There was a positive association between intervention rate and effect size. These results were consistent with those of Northouse et al,74 who concluded that a greater number of intervention sessions and a greater number of intervention hours had a positive influence on coping ability.

In terms of study quality, most of the studies used self-report scales for measuring the success of interventions. Although validated self-report scales have been used to assess sex-related variables, structured interviews might help minimize biases. Ten of the studies had the risk of bias in blinding for outcome assessment. As the participants were receiving the psychoeducational intervention from nurses, other specialists, or peers, the blinding of participants and researchers might be difficult, as it would be unfeasible or unethical.75,76 In the subgroup analysis, the effect size for studies with a high risk of bias in blinding outcome assessments was low compared with studies with a low risk or uncertain bias. Studies using nonrandom assignment had a larger effect size than those using random assignment. Through randomization, a study attempts to control or decrease as many confounding variables and other potential sources of bias as possible. This allows for confirmation that the observed effects are due to the intervention alone. Design strategies used to control for extraneous variables are considered the best source of experimental evidence, including random allocation of participants to an experimental or control group.77,78 However, it was demonstrated in a previous study that effect sizes did not significantly differ between studies with and without random assignment.79 Researchers examining educational interventions should use more rigorous research designs and methods for minimizing bias and gathering evidence based on current best practices. Randomized controlled trials are highly demanding for researchers and the clinical nursing staff.75

Strengths and Limitations of the Study

The data from the overall analysis might be considered a weakness or limitation of the analysis, as these data were derived from a number of different cancers. To overcome the differences between studies and their multifaceted conceptions of sexual intervention among cancer patients, we conducted an analysis of data from various subgroups of preexisting studies, obtained an overall estimate of the effectiveness of the selected subject/intervention, and compared its effectiveness across variables that affect intervention outcomes. These results should contribute to the establishment of practical guidelines for sexual intervention for cancer patients.

This study had some additional limitations. We did not include studies published in languages other than English or qualitative research. This study included a limited number of studies in the meta-analysis and a relatively small number of patients, as the review collected data from 15 reports with less than 130 patients. Furthermore, many studies were piloted with a small number of participants because of the undeveloped nature of interventions for sexuality in cancer patients. As a result, making recommendations for its effectiveness might be inappropriate, and drawing conclusions regarding practice might be premature.

Despite the above limitations, these results provide empirical data for evidence-based practice and inform the development of useful intervention programs through a comprehensive review and meta-analysis of the results from previous studies regarding sexuality-based psychoeducational intervention after cancer diagnosis.

Implications for Practice

From a clinical perspective, individual-based interventions appear to be useful for improving sexual outcomes. Given the weaknesses and difficulties inherent in using Internet- or telephone-based interventions independently, the combination of the face-to-face and online or telephone intervention might be effective because of the ability to provide tailored psychological interventions. The psychoeducational intervention would be more effective when the intervention rate is increased. Furthermore, registered nurses must play a key role in sexual healthcare for psychoeducational interventions focused on sexuality, as nurse intervention providers have been demonstrated as effective compared with other professionals or peers.

References