TrueNTH sexual recovery study protocol: a multi-institutional collaborative approach to developing and testing a web-based intervention for couples coping with the side-effects of prostate cancer treatment in a randomized controlled trial

The Movember Foundation has collaborated with national non-profit organizations, such as the Livestrong Foundation (United States) and the Prostate Cancer Foundation (Australia, the United Kingdom, the United States and Canada) to offer funding to selected institutions with a clinical and research track record in prostate cancer survivorship. A Request for Proposals in the United States was issued in the fall of 2013, and 15 institutions were selected: Johns Hopkins University, Emory University, Harvard University, Memorial Sloan Kettering Cancer Center, Moffitt Cancer Center, University of California San Francisco, Los Angeles and Davis, University of Michigan, University of Washington, Oregon Health Sciences University, University of Colorado-Denver, Wayne State University, Duke University and University of North Carolina. In March 2014, representatives from the 15 institutions met to define priorities for interventions, and agreed to participate in building those interventions in which they had the most experience. Sexual Recovery was one of the priorities selected for intervention building.

The Sexual Recovery Intervention was developed by a team (i.e., the Full Team) of sexual health experts, urologists, psychologists, health services researchers, epidemiologists, primary care physicians, radiation oncologists, medical oncologists, nurses (one of these professionals is a survivor and another one is a partner), and survivor and partner advocates from 11 institutions (Emory University, University of Colorado-Denver, University of California-Los Angeles, Johns Hopkins University, Fred Hutchinson Cancer Research Center, University of Michigan, Duke University, Memorial Sloan Kettering and Harvard University, University of North Carolina, University of California-San Francisco). A Design Team was established to write a proposal for intervention development and testing. It included a sexual health expert, a primary care physician, a urologist, and an internal medicine physician with expertise in information technology. The Full Team participated in monthly teleconferences and also provided regular electronic editorial contributions. After peer review by the Scientific Committee of the Movember Foundation, the protocol was amended and re-submitted for funding. Funding was assigned in March 2015.

The intervention was developed at the University of Michigan, given the institution’s expertise in sexual health and technology. However, the Design Team and the Full Team provided critical feedback during the development phase. The intervention is currently being tested in a multi-center randomized controlled trial (RCT) involving five institutions in the United States (Emory University, University of California-Los Angeles, Memorial Sloan Kettering Cancer Center, Johns Hopkins University and the University of Michigan). Plans are in progress to adapt the intervention in the United Kingdom.

The intervention, once tested, will be integrated into a support structure for prostate cancer survivors, partners, and families, which will be available on the Movember Foundation website. The prioritized US interventions have a collective over-arching identity as True NTH (pronounced ‘True North’). Its goal is “to improve the physical and mental well-being of men living with prostate cancer, together with their partners, caregivers and families.”

The proposed web-based sexual recovery intervention is underpinned by a theoretical model that benefits from existing research. It was tested in a mixed methods study to ensure content validity by the principal investigator and colleagues [42]. It is based on the biopsychosocial understanding of sexuality [11, 43]. Recognizing that ED and subsequent changed sexuality represent loss for both the man and the partner [44, 45], the model incorporates concepts from the grief literature to suggest that sexual recovery includes, at least in part, a grief process [46, 47].

The study’s first step was to develop the intervention and conduct formative testing with 3 focus groups which included 4 heterosexual couples, 8 survivors and 8 partners, respectively. Two non-heterosexual couples were interviewed separately as we were unable to recruit for a full focus group. Usability testing was conducted with 5 couples and 5 individual survivors who commented, while working through the intervention, on the ease of use, relevance, content and graphics. The majority of the participants endorsed the web-based format and suggested that the intervention include preparation for the sexual side-effects, rehabilitation, preparation for the emotional impact of erectile dysfunction and diminished libido on couples, education for partners about the importance of erections to men, sensitivity to gay sex and education of providers regarding sexuality of men who have sex with men [48].

Figure 1 illustrates the study process for the intervention group, starting prior to diagnosis with preparation for the sexual side-effects, their impacts and rehabilitation, and continuing with support for sexual recovery through 6 months after treatment (also the study completion). The figure also illustrates where the intervention modules of the TrueNTH Sexual Recovery Intervention will aim to affect the process of couples’ recovery. One of the modules (M10) coaches patients and partners to engage comfortably in discussions with healthcare providers about sexual problems and solutions. Finally, the figure includes measurement of the effect of the intervention on outcomes proposed in this protocol.

This study aims to evaluate the effect of 6 modules of the intervention in an RCT on patients’ and partners’ satisfaction with their sex life, compared to usual care at 6 month follow up. It is hypothesized that patients and partners who participate in the online intervention will report more satisfaction with their sex life than the control group. The intervention group is also expected to report greater knowledge about sexual recovery, less intense grief about sexual losses, more interest in sex, greater use of sexual aids, better dyadic coping, and a higher quality of life than the control group after treatment.

Recruitment will take place in urology and radiation oncology clinics at participating sites in the True NTH consortium: University of Michigan, Emory University, Johns Hopkins University, University of California –Los Angeles and Memorial Sloan Kettering Cancer Center. Sites were selected based on (1) patient volume which will impact feasibility and recruitment costs, (2) diversity of the patient population, and (3) current standard of care regarding sexual health. Each site will have a site PI and a study coordinator. The website’s survey and intervention couple engagement logistics data will be centrally managed at the University of Michigan which will also serve as the coordinating site for clinical data and a provider of technical support for all participating sites. Potential participants will be identified by a member of the study team at participating sites through screening clinic and surgical schedules as well through referral by the treating physician. The goal for each site will be to recruit a minimum of 15 couples and a maximum of 50 couples.

The patient/partner dyad will be randomized to either the intervention arm or the usual care arm after completion of the informed consents, and prior to initiation of treatment. Randomization will be performed using a computer-generated randomized block design to minimize the potential for sample size differences between the two arms. The randomized block design will be stratified by enrollment site to facilitate ease of administration of the trial (Fig. 2). Participants randomized to usual care will be routed to the American Cancer Society’s web page that provides information about the sexual side-effects of cancer and sexual rehabilitation, which can be considered a part of usual care. Participants will be blinded as to the arm that they are entering by language at recruitment which will indicate that we are comparing two supportive interventions for sexual recovery after prostate cancer treatment. Participants will be unblinded by letter after they complete their 6-month survey.

The full intervention will consist of 10 web-based modules, 6 of which are consistent with the trajectory of prostate cancer and its treatment (Table 1). Four modules will be thematically linked to the experience of sexual recovery after prostate cancer treatment.

The approach to the intervention is mindful of the a) biopsychosocial nature of sexuality; b) need to include information for gay men, c) representation of diverse cultural and ethnic groups, d) recognition that recovery does NOT mean returning to baseline, and d) the need to involve the partner at each stage of the sexual recovery process. The modules are tailored so that the content is relevant and appropriate according to (1) partnership, (2) prostate cancer treatment type and (3) sexual orientation. The language comprehension of the intervention reflects users with a 6th grade education. Although all modules were developed with the requisite tailoring during this project, only the 6 (1,2,3,4,5,10) modules, addressed to couples, will be tested initially in the RCT.

Each intervention module has the following sections: 1) video introduction featuring an individual or a couple who have gone through the prostate cancer experience or a relevant expert, such as a sex therapist, 2) educational content, 3) an activity for couples to engage in online, and 4) an opportunity to identify patient and partner concerns. Each intervention module lasts approximately 45 min or less. After each module, participants will receive, via email, tailored strategies, based on concerns they raised during the immediately preceding module’s activity.

Participants who pass through the screening process will be recruited in clinics or by telephone, depending on their availability. They will be consented during clinic appointments or by mail. Each couple will be given an envelope with their study identification number (ID) and login information with which to enroll into the True NTH website. The login password is coded to randomize the participating couple into the intervention or control arm of the study. After logging in, participants will provide basic demographic information. Patients will be asked to identify the gender of their partner and type of treatment they will undergo. Once this information is entered, the couple will be sent to surveys tailored to their sexual orientation and treatment type. They will then be asked to fill out baseline surveys. The participants who are randomized into to the intervention arm will access content tailored to their treatment and sexual orientation. Participants who do not fill out their baseline surveys will receive up to 3 phone calls from the study coordinator at participating sites. Those who do not fill out baseline surveys will not be able to continue in the study. Participants who do not log into a given module will receive up to 3 email reminders which are programmed into the intervention. Participants will be allowed to move forward in the intervention even if they do not log into the 3-month surveys to enable tracking their interest in the web-based support.

Participants in the intervention arm will receive up to 3 email reminders when their successive modules become available. They will be able to continue in the intervention even if they do not access the modules. Participants in the control arm will access a link to a “Sexual Health after Cancer” web page of the American Cancer Society (ACS). They will receive 3 email reminders. The ACS website link will be available each time these participants access the True NTH website, for example when they fill out their 3 month and 6 month surveys. They can also return to the True NTH website at will throughout the time of the study. As participants will enter the study prior to treatment, their study participation will continue for approximately 7 months.

Measurement timeline that includes both the primary measure and secondary measures is outlined in Table 2. Description of the measures is included in this section.

Patient Reported Outcome Measure Information System (PROMIS) Global Satisfaction with Sex Life [49] measures patients’ and partners’ overall satisfaction with their sex life. Participants respond to seven items on a 5-point scale. A higher score reflects greater satisfaction (Cronbach alpha = 0.92).

PROMIS Interest in Sex Life [49] measures patients’ and partners’ interest in sex. Participants will respond to four items on a 5- point Likert scale. A higher score will indicate a higher level of interest (Cronbach alpha = 0.87).

PROMIS Sexual Activity [49] is a 13-item measure that evaluates the frequency and type of sexual activity. It is a descriptive measure.

PROMIS Use of Aids [49] is a 7-item measure that assesses the use of hormones, personal lubrications, medications, or devices intended to allow for or improve sexual function. These items are intended to be “stand alone” items and do not comprise a unidimensional scale.

Expanded Prostate Cancer Index Composite-26 (EPIC) [50] is a 26-item prostate cancer health-related quality of life (HRQOL) instrument that assesses urinary, bowel, sexual, and hormonal treatment-related symptoms of prostate cancer. Participants respond to a Likert scale. A higher score reflects higher function (Cronbach alpha = 0.93). Sexual function cut-off scores of the EPIC are: 0–33 (severe ED), 34–45 (moderate ED), 46–60 (mild/moderate ED), 61–75 (mild ED), and above 75 (no ED).

EPIC-Exp [50] is a measure parallel to the EPIC, but focused on expectations of functional recovery 1 year after treatment.

Female Sexual Function Index (FSFI) [51] is a 19-item questionnaire measure of sexual functioning in women. It assesses six domains of sexual functioning: sexual desire, arousal, lubrication, orgasm, satisfaction and pain. A total score is a sum of the domain scores. Participants respond to a Likert scale.

Dyadic Coping Inventory (DCI) [52] is a 16-item questionnaire that assesses stress communication and dyadic coping as perceived by (1) each partner about their own coping (what I do when I am stressed and what I do when my partner is stressed), (2) each partner’s perception of the other’s coping (what my partner does when he or she is stressed, and what my partner does when I am stressed), and (3) each partner’s view of how they cope as a couple (what we do when we are stressed as a couple). Participants respond on a Likert scale. A higher score reflects higher dyadic coping (Cronbach alpha for subscales = 0.71–0.92).

Prolonged Grief about the Loss of Sexual Function [53] is a 22-item scale that was adapted for sexual function loss from the Prigerson’s validated measure of prolonged grief in bereavement. Participants respond to a Likert scale.

The International Index of Erectile Function (IIEF) [54] is a 15-item scale with subscales that measure erectile function, orgasmic function, sexual desire, intercourse satisfaction and overall satisfaction. Participants respond on a Likert scale. A higher score reflects higher sexual function. (Cronbach alpha for subscales and overall IIEF is 0.91, 0.92, 0.92, 0.77, 0.73, 0.74, respectively). As the IIEF was validated for heterosexual sex with vaginal intercourse as the expected activity, several additional items will be added for gay participants, but not included in scoring to inform about gay sexual experience.

Expectations of Success of Treatment [55] is a 10-item measure, adapted from a mental health measure for confidence regarding treatment for sexual dysfunction. It measures confidence in the treatment and in the outcome of the treatment. Participants respond on a Likert scale. A higher score reflects greater confidence in the success of treatment.

Quality of Life EQ-5D [56] is a standardized QOL instrument for use as a measure of health outcomes. It essentially consists of 2 pages – the EQ-5D descriptive system, and the EQ visual analogue scale (EQ VAS). The EQ-5D-3 L descriptive system comprises the following 5 dimensions: mobility, self-care, usual actives, pain/discomfort and anxiety/depression. Each dimension has 3 levels: no problems, some problems, extreme problems.

Comorbidities [57], such as heart disease, arthritis, diabetes and others, will be collected on a self- reported basis from both the patient and partner.

Clinical data will be abstracted from patients’ health records, including PSA and clinical stage of prostate cancer at diagnosis, Gleason score, comorbidities as rated per the Charlson Comorbidities Index [58], etc. For surgical patients, pathological stage, margin status (positive/negative), pathologic Gleason score and 3-month post-surgery PSA at 3 will be collected.

The primary endpoint of PROMIS Satisfaction with Sex Life at 6 months is expected to be a continuous measurement that is normally distributed. If not, we will transform the data for this outcome to approximate normality. Since the primary outcome is measured on both patients and partners and randomization was done at the level of each dyad, we will assess the impact of the intervention using the Actor-Partner Interaction Model (APIM) [59]. The APIM is a mixed-effects linear regression model that allows us to simultaneously estimate the association of partner characteristics with partner and patient outcomes, as well as simultaneously estimate the association of patient characteristics with both outcomes. The APIM incorporates random effects to account for the correlation between patient and partner outcomes from the same dyad. Our base model will contain (1) an indicator of intervention versus control, (2) baseline PROMIS Satisfaction with Sex Life, (3) a indicator of partner versus patient, (4) interaction terms of (3) with (1) and (2), (5) a random partner effect to account for variation among partners, (6) a random patient effect to account for variation among patients, and (7) a third random effect to account for the correlation (non-independence) between outcomes measured from the same dyad. From this base model, we can assess not only the relative impact of the intervention among dyads, but also the differential impact of the intervention on partners and partners.

We can then expand our base model to include other partner characteristics and/or patient characteristics to assess whether or not these characteristics moderate or mediate the effect of the intervention for the patient and the partner. For example, we could include baseline satisfaction with sex life of the partner and the patient and see if either or both factors moderate or mediate the effect of the intervention on 6-month satisfaction with sex life. We note that because dyads were randomized to the intervention, we expect dyad level characteristics to have little effect in our model because those factors should be nearly balanced among the two treatment arms. Thus, adjustment for them is unnecessary, except to gain statistical power. The random effects will allow us to estimate the remaining variability of outcomes among partners, among patients, and the correlation of outcomes within each dyad, that is not explained by our APIM.

All secondary endpoints will be analyzed with an APIM exactly as described for the primary endpoint. It is possible that some secondary endpoints, even after transformation, will have slightly skewed distributions. However, random effect models are fairly robust to slight deviations from normality.

Along with outcome data analyses, we will conduct an analysis of the logistics of participant activity. This will include participant engagement with the modules of the intervention on the web, the need for reminders, time spent on each module and number of times each module was accessed.

Basing power calculation on the average adjusted PROMIS Satisfaction with Sex Life score of 50 (standard deviation = 10) provided in the PROMIS scoring and anticipating Sexual Satisfaction scores for the intervention group vs. control group to be 50 vs. 45 based on expert opinion, a sample size of 128 patients and 128 spouse/partners will provide 80% power at the 5% significance level to detect this difference using a two-tailed t-test. Adding 10%, or 14 dyads (including an extra one to get to an even number to allow the same sample size in each arm), to our total to account for the anticipated dyads that will be lost to follow-up prior to study completion, we will recruit 140 dyads (280 individuals) for this study (expected to be 70 couples on each study arm).

The study will be coordinated by the University of Michigan, the protocol has been approved by the University of Michigan Institutional Review Board (IRB) and by the IRBs of all the other participating sites. The process of the study and protocol modifications, if needed, will be discussed in weekly meetings with all the sites and amendments filed to each institution’s IRB. Data will be collected by the University of Michigan Center for Healthcare Communications Research (CHCR) and stored securely on virtualized servers with a robust back up system. All participants must sign informed consent before enrolling in the study.