A randomized trial of internet-based versus traditional sexual counseling for couples after localized prostate cancer treatment




After treatment for prostate cancer, multidisciplinary sexual rehabilitation involving couples appears more promising than traditional urologic treatment for erectile dysfunction (ED). The authors of this report conducted a randomized trial comparing traditional or internet-based sexual counseling with waitlist (WL) control.


Couples were randomized adaptively to a 3-month WL, a 3-session face-to-face format (FF), or an internet-based format (WEB1). A second internet-based group (WEB2) was added to examine the relation between web site use and outcomes. At baseline, post-WL, post-treatment, and 6-month, and 12-month follow-up assessments, participants completed the International Index of Erectile Function (IIEF), the Female Sexual Function Index (FSFI), the Brief Symptom Inventory-18 to measure emotional distress, and the abbreviated Dyadic Adjustment Scale.


Outcomes did not change during the WL period. Of 115 couples that were randomized to FF or WEB1 and 71 couples in the WEB2 group, 34% dropped out. Neither drop-outs nor improvements in outcomes differed significantly between the 3 treatment groups. In a linear mixed-model analysis that included all participants, mean ± standard deviation IIEF scores improved significantly across time (baseline, 29.7 ± 17.9; 12 months, 36.2 ± 22.4; P < .001). FSFI scores also improved significantly (baseline, 15.4 ± 8.5; 12 months, 18.2 ± 10.7; P = .034). Better IIEF scores were associated with finding an effective medical treatment for ED and normal female sexual function at baseline. In the WEB2 group, IIEF scores improved significantly more in men who completed >75% of the intervention.


An internet-based sexual counseling program for couples was as effective as a brief, traditional sex therapy format in producing enduring improvements in sexual outcomes after prostate cancer. Cancer 2011;. © 2011 American Cancer Society.

Despite the promise of nerve-sparing prostatectomy1 and innovations in radiation therapy,2 surveys of diverse samples of prostate cancer survivors agree that only a minority are satisfied with their sexual rehabilitation outcomes.3-7 Not only are rates of erectile dysfunction (ED) higher than estimates from more selected cohorts,3, 4, 7 but medical treatments for ED after prostate cancer have had poor success rates and dismal adherence over time.8-11 Men are far more likely to receive ED treatments after radical prostatectomy than after radiation therapy.8-10, 12 In a population-based study of almost 39,000 prostate cancer survivors, only approximately 26% of men tried oral medication after radical prostatectomy, and 11% tried such medication after radiation therapy.12

Growing evidence suggests that “penile rehabilitation,” ie, using medical therapies to promote penile blood flow several times a week, protects epithelial cells in the penis and prevents fibrosis in the cavernosal tissues, allowing a more complete recovery of erections after radical prostatectomy13 or maintaining erectile function after radiation.14 Adherence to penile rehabilitation also is poor, however, with the high cost of medication, unrealistic goals for recovery of erections, and female partners who have lost interest in sex cited as barriers15, 16 Partners rarely are included in ED treatment of prostate cancer survivors, although having a partner who enjoys sex is associated with seeking medical help for ED17 and with better sexual rehabilitation outcomes.8

Pilot studies have suggested that combining medical management of ED with brief sexual counseling may improve adherence and increase men's satisfaction with urologic ED treatment in men unselected for health18-20 and in prostate cancer survivors.21-23 In contrast, interventions for men with prostate cancer that focus on general coping and include limited material on sexuality have failed to improve sexual function in men with prostate cancer.24-28

We published a 4-session pilot sexual counseling study of men who were treated for localized prostate cancer and their partners.29 The objectives of this study included enhancing both partners' sexual satisfaction and helping them integrate effective treatments for ED into their sex lives. Eighty-four heterosexual couples were randomized to have sessions either with both partners present or with the man alone. However, the partner participated in behavioral homework and completed questionnaires in both groups. Fifty-one couples completed all 4 sessions (drop-out rate, 39%). Men's use of ED treatments increased significantly from 31% at baseline to 49% at 6-month follow-up. Total scores on the International Index of Erectile Function (IIEF)30 for men and on the Female Sexual Function Index (FSFI) for women31 improved significantly from baseline to post-treatment. Men's distress levels decreased significantly (Brief Symptom Inventory-18 [BSI-18]).32 By the 6-month follow-up assessment, however, these gains had regressed toward baseline values.

In the current report, we describe a randomized trial that compared face-to-face and internet-based versions of a revised and strengthened intervention, which we called the Counseling About Regaining Erections and Sexual Satisfaction (CAREss) Trial. We created an internet-based version because it would be convenient to disseminate and would take less therapist time. Not only do men routinely seek sexual content on the internet,33, 34 but patients with prostate cancer who desire information about the sexual impact of treatment view the internet as a preferred source.35 Our hypotheses were that the 2 formats of the intervention would have equal efficacy and that outcomes would be superior to those after a 3-month waitlist (WL) control condition.


This study was approved by the University of Texas MD Anderson Cancer Center Institutional Review Board, and yearly reports were submitted to the Data Monitoring Committee.

Eligibility and Recruitment

Couples were eligible if the male partner was heterosexual, aged ≥18 years, and had been treated for localized prostate cancer (T1-3N0M0) with either definitive surgery or radiotherapy between 3 months and 7 years previously. Couples had to be married or must have been living together for ≥1 year. Both partners agreed to participate in the assessments and intervention and had reasonable English fluency. Men were either unable to achieve and maintain an erection sufficient for sexual intercourse on ≥50% of attempts in the past 3 months or had not attempted intercourse for 3 months and also had not noted firm erections on waking from sleep. Men could not be using a satisfactory medical treatment for erectile dysfunction.

Couples were excluded if the man currently was using hormone therapy for prostate cancer because of the profound impact of such therapy on the desire for sex. Couples who entered the randomized trial had to be willing to come to The University of Texas MD Anderson Cancer Center 3 times during the 12-week treatment period.

Recruitment included mailing invitation letters to men in The University of Texas MD Anderson tumor registry who met eligibility criteria. We also relied on physician referral, fliers placed in outpatient clinics, and public service announcements in local media or on web sites of prostate advocacy groups. Special efforts to recruit African American couples included presentations at local churches, community health fairs, and announcements in newspapers and radio programs targeted to Houston's African American community. African American couples interested in participating were screened and treated by an African American psychologist.

Study Design

Couples were adaptively randomized to 1 of 3 groups using minimization.36 The following variables were used to assign participants, creating the best overall balance for each group: the man's age (<65 years vs ≥65 years), white non-Hispanic versus any minority ethnicity, and radiation versus surgery as primary therapy. The 3 arms were a 3-month WL group, an immediate intervention group that received 3 face-to-face sessions over 12 weeks (90 minutes for session 1 and 50-60 minutes for sessions 2 and 3) (FF), and an immediate intervention group that used an internet format of the intervention though e-mail contact with their therapist (WEB1). After the 3 months of the WL condition and completion of a post-WL assessment, WL participants were randomized a second time to 1 of the 2 intervention groups. During the course of the trial, we added a third intervention group (WEB2), which was recruited with separate materials. Couples met all entry criteria except that they lived too far away to enter the randomized trial. They were invited to enter directly into the internet-based intervention. We included the WEB2 group to have enough power to examine the association between the amount of web site use and outcomes. In the analyses described below, we have clearly specified whether WEB2 data were included.

Outcome Measures

All participants were mailed assessment questionnaires at baseline; after the 12-week treatment period; and at 3-month, 6-month, and 12-month follow-up. In addition, the WL group completed questionnaires at the end of their 3-month waiting period. Primary outcome measures were the total score on the IIEF,30 a 15-item assessment of sexual function and satisfaction for men, and the total score on the FSFI,31 a similar questionnaire for women. The BSI-1833 assessed current distress, and an abbreviated form of the Dyadic Adjustment Scale (A-DAS)37 measured relationship satisfaction. In the WEB groups, the duration, number, and content of visits were recorded electronically. A graphic on each participant's home page displayed the percentage of the intervention completed (pages visited or homework reports submitted).

Intervention: Content and Formats

The FF and WEB formats of CAREss included the same content and cognitive-behavioral homework. FF participants received printed handouts of materials from the web site, except for animations and videos. In the WEB groups, each partner had a unique user name and password and could not access the other's responses. Homework exercises had standardized report forms completed online (WEB) or on paper (FF) and were submitted by each partner. Therapists e-mailed feedback to the couple (WEB) or discussed homework in session (FF). If no homework reports were returned within 2 weeks of WEB entry, then e-mail and, subsequently, telephone reminders were given to 1 or both partners. WEB participants also could e-mail their therapist at any time. E-mail to the therapist was programmed within the web site to maintain privacy. The web site was encrypted and maintained on a secure server with access restricted to research participants and staff. Loaner laptop computers were provided for a few couples who did not have internet access at home.

On first visiting CAREss, each WEB partner completed an online questionnaire about their sex history, current practices, and beliefs about sexual function and cancer. This information was elicited by interview for FF couples. CAREss included exercises to increase expression of affection, improve sexual communication, increase comfort in initiating sexual activity, and facilitate resuming sex without performance anxiety (using a sensate focus framework). Suggestions were provided to treat postmenopausal vaginal atrophy or cope with male urinary incontinence. Gender-specific exercises helped participants identify negative beliefs about sexuality and use cognitive reframing. Treatments for ED after prostate cancer were described, with suggestions on their efficacy and using them optimally.

A decision aid for choosing an ED treatment asked each partner to rate the importance of 7 characteristics (ie, not interfering with spontaneity, naturalness, producing reliable and firm erections, having few side effects, not causing permanent physical harm, not requiring ongoing expenses and physician visits, and not causing pain or discomfort). An algorithm generated a total score for each ED treatment (phosphodiesterase type 5 [PDE5] inhibitors, penile injections, vacuum pump, intraurethral suppository, penile prosthesis surgery, or having sex without firm erections) based on a score from 1 to 3 on each dimension multiplied by a participant's ratings on 10-point visual analog scales. Each partner received a numerical ranking of the 3 treatments that best met their expectations. Partners were asked to share their results, choose a first-line treatment, and make a plan for how to begin. Relapse prevention exercises were included near the end of the intervention. Booster telephone calls of 15 to 30 minutes were made to both groups at 1 month and at 3 months to discuss progress with end-of-treatment goals and ways to overcome remaining barriers. A therapist manual was used to train the therapists, who also had biweekly group supervision.

Statistical Analyses

Descriptive analyses were conducted on questionnaire data from baseline through 12-month assessments and on electronic recordings of web site usage. Descriptive statistics included frequencies, means, ranges, and standard deviations (SDs). Distributional characteristics of the variables were examined using box plots, histograms, and the Kolmogorov-Smirnov test of normality. Nonparametric tests, including the chi-square test, the Fisher exact test, and the Mantel-Haenszel test, were used to compare differences between treatment groups or across time, as appropriate. For repeated outcome measures, linear mixed-model (LMM) multivariate analyses were calculated based on all participants to adjust for within-subject correlations. The Akaike Information Criterion was used to select appropriate covariance structure and assess model fit. This approach is more accurate than the traditional intention-to-treat analysis. LMM uses all observed data and automatically handles missing data. It yields consistent estimates of effects, provided that the covariates in the model account for the probability of missing data.



Table 1 presents initial recruitment to each treatment group as well as attrition. During the 3-month WL period, 9% of couples dropped out. Remaining WL couples were randomized again to the FF group (N = 20) and the WEB1 group (N = 22). Thus, 115 couples entered the randomized trial, and 71 entered the WEB2 additional group. Recruitment was far more rapid for the WEB2 group, offering couples immediate entry into the internet-based intervention.

Table 1. Attrition in the Counseling About Regaining Erections and Sexual Satisfaction (CAREss) Trial
Time of AttritionWaitlist, N = 48FF, N = 60WEB1, N = 55WEB2, N = 71
  • Abbreviations: FF, face-to-face format; WEB1, internet-based format; WEB2, second internet-based group examining web site use and outcomes.

  • a

    Drop-outs during 3-month waitlist period completed baseline questionnaires but not post-waitlist questionnaires and, thus, never began an intervention.

  • b

    Drop-outs during intervention: either did not complete all 3 sessions of FF counseling, or both partners completed <25% of the internet intervention.

  • c

    Couples who were lost to follow-up completed an intervention but never completed any questionnaires afterward.

No. of drop-outs during waitlist period (%)a5/48 (9)
No. of drop-outs during intervention (%)b17/60 (28)7/55 (13)12/71 (17)
No. lost to follow-up (%)c3/43 (7)7/48 (17)16/59 (27)
Total attrition, %9332539

A chi-square test indicated that there were no statistically significant differences in the drop-out rate between the intervention groups (P = .44). Univariate logistic regression analyses indicated 2 significant associations with dropping out. Drop-outs were less likely if men had more positive baseline scores on the A-DAS (higher relationship satisfaction; P = .025). If the woman had a postgraduate degree, then the couple was more likely to drop out (P = .035). Although neither variable retained significance in a multivariate regression analysis, baseline A-DAS scores among men and education among women were controlled in other analyses of outcome to eliminate potential bias.

Table 2 illustrates that the final groups were very similar in medical and demographic variables. All unmarried men had a sexual partner. No significant differences were observed between the FF and WEB1 groups. With the WEB2 group included, the FF group has a shorter duration of follow-up since cancer treatment than either the WEB1 group or the WEB2 group.

Table 2. Medical and Demographic Characteristics by Treatment Group
 Percentage of Men
CharacteristicFF, N = 40WEB1, N = 41WEB2, N = 43
  • Abbreviations: FF, face-to-face format; SD, standard deviation; WEB1, internet-based format; WEB2, second internet-based group examining web site use and outcomes.

  • a

    With all 3 groups included, P = .0128.

 <High school000
 High school859
 Some college201530
 4-Year college degree434933
 Postgraduate degree303228
 African American1057
Relationship status
 Never married000
Primary prostate cancer treatment
 Radical prostatectomy706884
 Radiation therapy303216
Age: Mean ± SD, y64 ± 864 ± 764 ± 8
Time in relationship: Mean ± SD, y31 ± 1435 ± 1532 ± 14
Time since treatment: Mean ± SD, moa17 ± 1425 ± 2426 ± 14


None of the outcome measures changed significantly for men or women during the 3-month WL period, confirming that, at a mean of 1.5 to 2 years post-treatment, sexual dysfunction is unlikely to resolve with time. For 43 men with baseline and post-WL IIEF total scores, the means ± SD scores were 27.3 ± 16.1 and 27.3 ± 17.6, respectively. Table 3 lists the means ± SDs for IIEF total scores in men by intervention group across time. A LMM was used to assess changes in IIEF scores and differences between intervention groups, controlling for covariates that were chosen with the Akaike Information Criterion. IIEF total scores improved significantly over time in all intervention groups, although, in the WEB2 group, IIEF scores began significantly higher at baseline than in the randomized trial groups (P < .05). Effect sizes are provided in the Table 3 footnote. IIEF total scores can vary from 0 to 75. In a comparison of 111 men with ED and 109 age-matched controls (average age, 55 years), the mean total IIEF for sexually functional men was 53.4 compared with 31.2 for men with ED.38 Although the current sample was 10 years older, many men did not attain optimal sexual function at follow-up.

Table 3. Mean International Index of Erectile Function Scores for Men Across Time by Treatment Group
 FFWEB1WEB2Whole Sample
Time of IIEF ScoreNo.Mean ± SD ScoreNo.Mean ± SD ScoreNo.Mean ± SD ScoreNo.Mean ± SD Score
  • Abbreviations: FF, face-to-face format; IIEF, International Index of Erectile Function; post-TX, post-treatment; SD, standard deviation; WEB1, internet-based format; WEB2, second internet-based group examining web site use and outcomes.

  • a

    The IIEF score improved across time (P < .0001; Cohen d, 0.35).

  • b

    The IIEF score improved across time (P = .0040; Cohen d, 0.35).

  • c

    The IIEF score improved across time (P = .0096; Cohen d, 0.27).

Baseline3926.4 ± 18.2a3327.4 ± 17.3b4034.8 ± 17.3c11229.7 ± 17.9
Post-TX3034.4 ± 22.22731.3 ± 20.43341.7 ± 20.19036.2 ± 21.2
3 Months2935.4 ± (22.02531.1 ± 23.22845.5 ± 20.38237.6 ± 22.4
6 Months2637.0 ± 22.42240.1 ± 23.53040.0 ± 22.67839.0 ± 22.5
1 Year2633.6 ± 23.12534.5 ± 22.52640.2 ± 22.07736.2 ± 22.4

Many researchers have focused on the 6-item Erectile Function subscale of the IIEF for measuring the severity of erectile dysfunction. A score of 25 has been used as a cutoff for normal function; however, in our work and that of others, scores ≥22 are associated with satisfaction with erections after prostate cancer.8, 39 Table 4 presents the percentage of men in each treatment group who attained this criterion over time. Between-group differences were not significant. Mantel-Haenszel tests indicated significant gains for all men between baseline and 6-month follow-up (P < .0006) and between baseline and 1-year follow-up (P < .0046). In addition to gains in the Erectile Function subscale, comparing baseline and 1-year scores, men improved significantly on orgasmic function (P = .001), intercourse satisfaction (P < .0001), and overall sexual satisfaction (P < .0001). Only sexual desire scores remained stable.

Table 4. Percentage of Men Who Achieved Near-Normal Erectile Function Across Timea
 Percentage of Men
Treatment GroupBaseline6 Months1 Year
  • Abbreviations: FF, face-to-face format; WEB1, internet-based format; WEB2, second internet-based group examining web site use and outcomes.

  • a

    Near-normal function was considered a score ≥22 on the Erectile Function subscale of the International Index of Erectile Function.


Neither marital happiness nor overall distress changed significantly across time for men in any subgroup or in the total sample. However, the sample was not distressed initially. The mean ± SD score for dyadic adjustment was high at baseline (24.4 ± 4.7; possible range, 0-36) and remained the same (24.6 ± 4.5) at 1-year follow-up. General Severity Index scores on the BSI-18 also were low (4.6 ± 6.2 at baseline and 4.6 ± 5.6 at 1-year follow-up). A raw score of 5 on the General Severity Index is equivalent to a T-score of 50 in both oncology and community normative samples of men.

Sexual function/satisfaction (FSFI total score) did not improve significantly for women within any treatment group. The mean ± SD score for all women was 19.8 ± 9.9 at baseline and at 21.0 ± 10.6 at 1-year follow-up. A score <26.55 indicates dysfunction, suggesting that at least half of women had sexual problems at 1-year follow-up.31 However, when women were divided into the 26% with normal FSFI scores at baseline versus 74% with abnormal scores, women who initially had poor sexual function did have a statistically significant improvement over time (P = .0255) (Fig. 1). Scores declined slightly for women who initially were functional but recovered to baseline by 1 year (P = .0219).

Figure 1.

This chart illustrates the impact of initial Female Sexual Function Index (FSFI) level on changes in women's sexual function across time. Post-TX indicates post-treatment.

Like the men, the dyadic adjustment among women was high and was stable across time (mean ± SD: 24.7 ± 5.0 at baseline and 24.7 ± 5.2 at 1-year follow-up). General Severity Index scores also indicated good overall adjustment (5.6 ± 7.4 at baseline and 4.1 ± 5.9 at 1 year). In the WEB2 group, women's initial BSI-18 scores were significantly more distressed than in the FF or WEB1 groups (7.6 ± 9.1) but tended to decrease over time (4.4 ± 7.2 at 1 year; P = .064).

Women's baseline sexual function also predicted the efficacy of CAREss in improving men's IIEF scores. Figure 2 indicates that baseline IIEF scores were significantly higher when women had FSFI scores above the cutoff (P < .001). IIEF scores only improved significantly over time if the partner initially was sexually functional (P < .001).

Figure 2.

This chart illustrates the impact of baseline Female Sexual Function Index (FSFI) scores on men's sexual function over time. IIEF indicates International Index of Erectile Function; post-TX, post-treatment.

Use of Medical Treatments for ED

At each assessment point, ED treatment use was categorized as 1) none, 2) using oral medication only, or 3) using invasive ED treatment (vacuum device, penile injections, penile suppositories, or penile prosthesis surgery) with or without supplementary oral medication. Rates of ED treatment use did not change significantly within any treatment group or for the entire sample. At baseline, 45% of men used no ED treatment, 23% used PDE5 inhibitors, and 32% used an invasive method. By 1 year, the corresponding rates were 46%, 20%, and 34%, respectively. However, at all assessment points, the men who were using ED treatment (oral or invasive) had significantly higher IIEF scores than those who were not using ED treatment (P < .0001). At baseline, the mean ± SD IIEF scores were 19.0 ± 13.1 for men who were not using an ED treatment, 36.2 ± 17.7 for men who were using pills, and 40.3 ± 15.8 for users of invasive treatments. The corresponding values at 1-year follow-up were 18.4 ± 16.1, 52.5 ± 15.5, and 50.0 ± 15.0, respectively. Scores at follow-up for men who were using ED treatments were similar to the norms for healthy men.38

Table 5 categorizes men's use of ED treatments at baseline and at 1-year follow-up as follows: 1) used no ED treatment at either time point or used an ED treatment at baseline but none at 1 year; 2) stable level of ED treatment, ie, oral treatment at both time points, invasive treatment at both time points, or invasive treatment at baseline and oral treatment at 1 year; and 3) intensified ED treatment, which was defined either as no treatment at baseline but oral or invasive treatment at 1 year or oral treatment at baseline and invasive treatment at 1 year. Men who intensified their treatment had large, significant gains in IIEF scores across time, and their partners also had significant improvement on the FSFI. The stable group had a smaller but still significant improvement in IIEF scores but no change for female partners. Neither partner improved on average if men were not using an ED treatment at 1 year. For 24 men who were not using an ED treatment at either time point, the mean ± SD change in IIEF was 2.09 ± 7.87. For 14 men who were using an ED treatment at baseline but none at 1 year, the IIEF change was −4.18 ± 9.7. Men who achieved an IIEF score ≥22 did not have a significantly longer duration of follow-up than men who had lower IIEF scores, suggesting that time alone was not associated with erectile recovery. Marital satisfaction and emotional distress were not correlated significantly with using ED treatment.

Table 5. Relation of Erectile Dysfunction and Erectile Dysfunction Treatment Use to Changes in Mean ± Standard Deviation Scores on the International Index of Erectile Function and Female Sexual Function Index Across Time (N = 83)
 Change in Use of ED Treatment From Baseline to 1-Year Follow-Up 
Change Scores Across TimeNone or Dropped, N = 38Stable, N = 34Intensified, N = 11Overall P
  1. Abbreviations: ED, erectile dysfunction; FSFI, Female Sexual Function Index; IIEF, International Index of Erectile Function; SD, standard deviation.

IIEF0.0 ± 8.98.4 ± 13.028.1 ± 17.8<.0001
FSFI−0.8 ± 5.9−0.8 ± 6.37.6 ± 8.4.0072

Optimal Internet Use and Outcome

For both WEB1 and WEB2 and for the combined groups, we analyzed outcomes related to 2 measures of electronically monitored web site use: the number of visits to the web site and the percentage of the intervention completed. Both indices were correlated significantly within couples, with a range of coefficients from 0.4 to 0.5. However, women made significantly fewer visits than men (P < .0001) and tended to complete significantly less of the web site (P < .0906). Fifty-two percent of men and 44% of women completed >75% of the web site. Figure 3 indicates that men who met this criterion had higher baseline IIEF scores but also improved slightly more over time. According to an LMM analysis, the difference reached significance only for the WEB2 group (P = .0427).

Figure 3.

This chart illustrates mean International Index of Erectile Function (IIEF) scores across time by men's level of web site completion (≤75% vs >75%). Post-TX indicates post-treatment.

Predictors of Good Sexual Outcome

LMM analyses (proc mixed in SAS), which included all 186 couples who participated in the study, were performed to identify the factors associated with improvement in sexual outcomes, which was defined as change across time in IIEF total scores for men and FSFI total scores for women. In these models, factors that were associated in univariate analyses with dropping out (the woman's age and baseline FSFI total score) were included as covariates. Other factors were male baseline A-DAS, treatment group, and type of ED treatment at 1 year. The significance of factors was almost identical whether the WEB1 and WEB2 groups were included separately or were combined; therefore, we report LMM analyses by sex, including all participants,.

Change in total IIEF score was significant across time (P < .001). Treatment group was not a significant predictor. In the final multivariate analysis, men had more improvement in IIEF scores if their baseline marital satisfaction score was higher (P = .001), the women's initial FSFI score was higher (P < .001), the female partner was younger (P = .007), and the man was using any treatment for ED at 1-year follow-up (P < .001).

The final multivariate analysis of FSFI scores at 1 year indicated that they increased slightly but significantly across time (P = .034). Improvements in FSFI scores were associated with higher IIEF scores for men at 1 year (P < .0001), the partner having less than a college education (P = .014); better dyadic adjustment among women at baseline (P = .014), and not being menopausal (P = .020).


Traditional face-to-face sexual counseling and an internet-based format of the CAREss intervention program that depended on e-mail for contact with the therapist produced equally significant gains in men's sexual function and satisfaction. In contrast, none of the outcome measures changed during a 3-month WL period. Outcomes peaked at 6-month follow-up, suggesting that couples improved their sex lives beyond the intervention period itself. Despite a slight regression to the mean, improvements remained statistically significant 1 year after the intervention, which is encouraging given the problems of nonadherence and dissatisfaction with ED treatment after prostate cancer.3-7 Men who are using ED treatments at 1 year have IIEF scores similar to those of healthy controls in normative samples. Compared with our pilot study,29 the drop-out rate was slightly less (39% vs 33%), and the benefits were much more enduring.

Men experience a range of sexual problems after prostate cancer treatment, including low desire, difficulties reaching orgasm, and problems with painful sex.7, 40, 41 The IIEF total score assesses the full range of sexual function and satisfaction, suggesting that the intervention is broadly effective.30, 38

Although partners had a more modest improvement in FSFI scores as a group, 26% had normal sexual function at baseline. Women who were dysfunctional initially did improve significantly. Womens' change scores for sexual function/satisfaction were highly correlated with the men's improvements. CAREss produced the largest gains when the male partner had discovered a successful ED treatment by the 1-year follow-up. It is unclear why more men did not achieve this goal, although the 54% rate of ED treatment use at 1 year is almost double the 26% to 30% reported in uncounseled cohorts.9, 12, 29 Past research suggests8, 16, 17 that the woman's sexual function and both partners' relationship satisfaction are important prognostic factors for the success of sexual rehabilitation. The association of better outcomes with a woman's premenopausal status and younger age suggest that CAREss may not be effective if the female partner grew up in an era of more conservative sexual values, has untreated postmenopausal dyspareunia, has lost interest in sex, or is unhappy in the relationship. More intensive treatment may be needed in such couples.

Men with less than a college education experienced greater improvements with CAREss. Perhaps they had less knowledge at baseline about cancer treatment and sexual function and made more use of novel information. In an Australian study, prostate cancer survivors with less education had more unmet sexuality needs.42 Low socioeconomic status has been associated with ED in general even after adjusting for age, comorbid diseases, health behaviors, and ethnicity.43

Recruitment rates for African American couples were disappointing despite our use of community partnerships and African American research staff. African American prostate cancer survivors report better sexual function than Caucasian men but perceive themselves as more impaired, reporting more distress about sexual problems and more disappointment with their cancer treatment modality.44-47 African American men are likely to have serial relationships compared with men of other ethnicities.48 Even within committed relationships, African American men are highly concerned about sexual function, but women focus on survival from cancer.49 The combination of privacy concerns about discussing sexuality in the African American community,50 fewer men in committed relationships, lack of partner motivation to focus on sexuality, and stigmatization of counseling51 may have interfered with recruitment. Recently at Duke University, 40 couples (far below the target number) entered a randomized trial of a psychoeducational program to improve quality of life for African American prostate cancer survivors.52 Only 12 couples in the experimental condition provided usable data. African Americans are less likely to use online cancer support groups and are more comfortable getting support from family and friends of the same ethnicity.53 A man-to-man peer-counseling model, like 1 we have used with African American breast cancer survivors,54 may be more appealing for these men than the CAREss program.

Our data cannot rule out the possibility that recovery of autonomic nerves after prostatectomy contributed to the positive outcomes. However, this seems unlikely. Duration of follow-up was not correlated with quality of erections, but gains in sexual function and satisfaction were strongly related to finding an effective ED treatment at 1 year. Furthermore, men were more likely to improve if they completed a higher percentage of the internet-based intervention. Most of the gains occurred during the intervention and in the subsequent 6 months, with some mild deterioration in outcomes between 6-month and 1-year follow-up, a finding more consistent with intervention efficacy than with natural healing.

We believe that an internet-based version of CAREss has the potential to be more cost-effective than traditional sex therapy. It took far less therapist time to respond to e-mails than to conduct traditional therapy sessions. Because private insurers rarely reimburse marital or sex therapy, an internet-based intervention could make treatment more affordable as well as accessible. The 3-fold faster recruitment rate for couples who were invited to enter directly into the internet-based condition may reveal a preference for this format, although the difference also may reflect reluctance to enter a randomized trial. A limitation of internet-based interventions, however, is high drop-out rates.55 Fewer than half of couples in the internet groups completed ≥75% of the program. The drop-out rate in the FF group also was more than optimal at 33%. Further empirical research is needed on methods to engage both partners in making behavioral changes.56


This project was supported by grant TURG-02-189-01-PBP (L. R. Schover, principal investigator) from the American Cancer Society, National Office.


The authors made no disclosures.