Study protocol and methods for Easing Pelvic Pain Interventions Clinical Research Program (EPPIC): a randomized clinical trial of brief, low-intensity, transdiagnostic cognitive behavioral therapy vs education/support for urologic chronic pelvic pain syndrome (UCPPS)

The Easing Pelvic Pain Interventions Clinical Research Program (EPPIC) trial is funded by the NIDDK under the R01 mechanism and registered on clinicaltrials.gov. All procedures described below have been approved by the University at Buffalo Institutional Review Board. At a minimum, this protocol comports with SPIRIT reporting guidance [68] (see supplemental materials for SPIRIT checklist).

Study setting

The clinical and administrative activities of EPPIC will take place in the clinic offices of the University at Buffalo’s Behavioral Medicine Division located at the Erie County Medical Center campus, an affiliated hospital of the Jacobs School of Medicine at the University at Buffalo. As a multi-site, EPPIC benefits from collaborations among experienced clinical scientists with subject matter expertise in biostatistics (Dr. Jaccard, NYU), UCPPS (Dr. Clemens, Michigan), and assessment (Dr. Naliboff, UCLA).

Participants and eligibility criteria

Planned enrollment is 240 adults between ages 18 and 70 (inclusive) of any gender, race, or ethnicity who have been formally assigned a diagnosis of IC/BPS or CP/CPPS (confirmed by a board-certified study urologist or urogynecologist) with clinically significant pelvic pain present for at least 6 months. Table 1 lists the inclusion and exclusion (urologic and general) criteria with corresponding rationales.

Table 1 Participant inclusion and exclusion criteriaRecruitment

We will adopt a two-pronged recruitment approach that emphasizes direct and indirect methods in an effort to optimize a sample representative of the range of patients with UCPPS. Indirect methods are aimed at enlisting support (e.g., referrals, generating positive word of mouth, building brand identity of the EPPIC) through an established network of “gatekeeper” physicians (e.g., primary, urologists and urogynecologists, OB/GYN), physical therapists, and other health care professionals who are in a position to engage participants in the EPPIC. Direct methods include efforts to promote EPPIC directly to patients in a wide variety of community locations including those frequented by patients with urologic conditions (urologists, urogynecologists, physical therapy offices, pharmacies, etc.) as well as at other community locales (coffee shops, places of worship, community centers, gyms), health fairs, and through local (e.g., newspapers, radio) media, social media (e.g., Facebook), study website (http://ubeppic.com/). We have also partnered with the Interstitial Cystitis Association in an effort to publicize the proposed trial through its clinical trial registry.

Trial interventionsMinimal-contact CBT

MC-CBT involves four 50–60-min individual clinic sessions delivered over a 10-week acute phase. While descriptive content is tailored to UCPPS, the protocol synthesizes [69] evidence-based CBT strategies into 4 modules targeting core transdiagnostic vulnerability factors [70,71,72,73] reflecting a rigid cognitive style expressed as discrete perceptual biases [74]. These include (a) a tendency toward self-immersive, abstract, and repetitive negative thought (RNT) [70] manifested in (b) the inclination to overestimate the probability of negative events (threat expectancy bias) [75,76,77,78]; (c) the tendency to inflate their costs or consequences when they occur (threat interpretative bias) [79,80,81]; (d) extreme negative self-schemas [82] (i.e., dysfunctional misconceptions or core beliefs like perfectionism [83,84,85]); and (e) a rigid, non-discriminative coping style characterized by an overreliance on control-oriented, problem-focused strategies deployed regardless of situational demands (e.g., controllability) [86,87,88,89,90]. Technical components include “real time” self-monitoring to generate a functional analysis of symptoms, their triggers, and responses across multiple domains (cognitive, emotional, somatic, behavioral), diaphragmatic breathing to reduce arousal and enhance personal control, worry control (e.g., evidence-based logic, decatastrophizing) to correct maladaptive information processing style, flexible problem solving, and relapse prevention skills to maintain gains after treatment discontinuation. MC-CBT content is introduced sequentially and reinforced through the provision of a workbook [91] with home exercises designed to facilitate skills acquisition. MC-CBT components and corresponding transdiagnostic processes are based on prior research with transdiagnostically designed CBT [47, 48] and refined through stakeholder involvement (e.g., patients, physicians, physcians assistants, and therapists) for the EPPIC. Contents addressed in of each week of treatment are presented in Table 2.

Table 2 Content topics and transdiagnostic processes of MC-CBT and content topics of education/support by treatment sessionUCPPS Education

EDU is delivered in four 50–60-min individual clinic sessions and structured around information dissemination, support, and reflection. Content includes information about chronic pelvic pain and its clinical features, epidemiology, diagnostic criteria, medical tests, and treatment options as well as the role of stress, diet, and physical activity (Table 2). To control for home exercises of MC-CBT, subjects receive a science-based pelvic pain education book [92] that emphasizes the therapeutic value of knowledge ("the more you know about pain, the better off you'll be"), track UCPPS symptoms (but not corresponding thoughts, behaviors, and emotions), and complete a stress profile [93] without prescriptive behavior changes overlapping with CBT.

Explanation of choice of comparators

EPPIC’s UCPPS Education (EDU) condition conforms to the best practice [67, 94] of a nonspecific comparator structurally equivalent to MC-CBT (credibility, time, attention, therapist training, etc.). By comparing CBT to a non-specific education condition, we will be able to discern whether treatment effects reflect the benefits of specific CBT’s technical components above and beyond the generic effects of simply going to treatment that comes from feeling listened to and receiving support, mobilizing positive expectancy for improvement, establishing a therapeutic relationship around working toward shared goals with a trusted and knowledgeable clinician.

Sequence generation, concealment, and implementation

Simple randomization into one of the two equal-probability conditions in a 1:1 ratio on a continuous basis as participants qualify for allocation will be performed by the study coordinator who has no patient care responsibilities as a safeguard against selection bias. In addition, treatment assignment will be conducted using Randomization Module in Research Electronic Data Capture (REDCap) software [95], which generates random and unpredictable sequence of assignments. The details of its computer-generated randomization algorithm are unknown to members of the EPPIC research team. Allocation sequence concealment from study personnel is achieved because REDCap conceals the next treatment assignment from being known. In other words, neither participants nor members of the research team are aware of the generated sequence until (and only for) the participant is assigned to his/her respective condition. Allocation sequence is generated by a computer independent of research coordinator who implements the assignment. Because the two treatments have identical dosages (4 sessions), the condition to which the participant is assigned is not revealed to him/her until session 1 of the acute phase, further minimizing selection bias.

Concomitant care policy

To optimize the external validity of study findings and expedite accrual, participants of both conditions will be permitted to continue with or modify the treatment they were engaged through the acute phase with exception of ongoing pelvic pain-targeted psychological therapy which is disallowed for allocation. To strike a balance between methodological (rigor) and ethical (safety) concerns, patients will be encouraged to maintain concomitant care use during the baseline period for the purpose of establishing a stable reference for gauging treatment effects. It is our policy that an outright requirement of maintaining a stable dose through the baseline period exposes the participant to an unjustified level of risk should s/he experience a serious health event for which a change in medications or other therapies is medically necessary for and represents a higher order consideration than internal validity. Because the likelihood of serious health events is expected to be low for this population, encouraging patients to maintain stable doses through the baseline period does not diminish rigor and may actually increase participant engagement that optimizes overall internal validity. We also believe that our approach results in a more representative sample that includes participants with medical comorbidities for which medications are often prescribed. Beyond ethical issues, stabilizing concomitant therapies during the acute treatment phase may distort the therapeutic benefit of a self-management treatment for participants who learn to control symptoms for which they no longer require pharmacological, rehabilitative, or dietary interventions reported at baseline. A stronger methodological approach is to assess concomitant health care (e.g., dosage, type) which we will capture via self-report, factor into statistical analyses, and present in the final report.

Blinding

A board-certified urologist or urogynecologist will confirm medical eligibility using formal diagnostic criteria for IC/BPS or CP/CPPS at baseline for all patients and function as independent evaluators (“blind” to treatment assignment) of symptom improvement at immediate (week 12), 3 months and 6 months follow-ups. Participants will be unaware of study hypotheses and blind to treatment assignment through the pretreatment baseline period. The methodological criterion of blinding participants to assigned treatments is inapplicable to behavioral interventions [96]. To the extent that blinding controls for differential expectations and consequent demand characteristics they may generate, we will adopt the established, surrogate practice [96, 97] of having participants rate the credibility and expectancy of improvement of the treatment to which they were assigned using the Credibility/Expectancy Scale [98] at the end of session 1 (week 1). Statisticians will be blinded to treatment allocation during the study by analyzing deidentified data until data is unlocked [99].

Retention and compliance

To optimize session patient retention and compliance, we will provide reminders via their preferred method of text, telephone call, or email within 1 business day of the scheduled appointment. We will provide an honorarium for travel, time, and convenience for assessments: Initial assessment ($25); interim ($25); and 2 weeks ($50), 3 months ($50), and 6 months ($50) follow-ups. Patients who complete 75% of sessions will be regarded a priori as having received a clinically thorough regimen of treatment to which they were assigned (e.g., compliers). Additional strategies are codified in a retention plan that covers areas such as staff training for initial contact, early detection of patient behaviors that may “red flag” correctable adherence problems (e.g., work conflict), formal training in rapport building, and positive staff-patient communication skills, creating a welcoming and respectful environment for participants, and educating participants about their role as participants and the role of participation incentives. Other procedures to minimize attrition and non-adherence include engendering trust, maintaining relevance to clients’ needs, establishing routine while maintaining a degree of flexibility in scheduling to maintain engagement in both treatment and assessment phases of a study, therapist techniques for “rolling with resistance”, and other brief motivational enhancement strategies that are uniformly applied across conditions and therefore do not represent a source of bias. Secondary indices of compliance include the number of no-shows (failing to show without contacting office), and canceled appointments without rescheduling. Compliance with weekly home exercises will be measured using a 6-point clinician rating scale ranging from 1 (0%) to 6 (>100%) [48, 100]. For participants who prematurely discontinue treatment (dropouts), we will identify self-reported reasons for withdrawal and record them. Reason(s) for dropout will be coded using five categories: logistical (e.g., childcare coverage), treatment/program related (e.g., participant prefers different treatment, stopped because they felt better, it failed to meet their needs, not the right time to engage in treatment); influence of others (e.g., treating doctor advised against continuing); study staff reasons (e.g., eligibility failure due to non-disclosure of information that would have rendered patient ineligible); miscellaneous (e.g., death). Participants who discontinue treatment will be encouraged to complete follow-up assessments in an effort to optimize intent to treat (ITT) analyses.

Treatment fidelity

To optimize the quality of and adherence to CBT and EDU, therapists will receive extensive training in the components of each treatment under expert supervision before being assigned to study patients. Delivery will be optimized by treatment manuals that provide detailed session-by-session guidance to standardize intervention across therapists; the completion of checklists for session protocols after each session; and regularly scheduled supervision with senior clinicians. Sessions will be audio taped, 25% of which will be randomly selected per patient and rated for protocol adherence by an independent rater unassigned to treatment delivery.

Data collection and managementPrimary outcome measure

The patient version of the Clinical Global Impressions - Improvement Scale (CGI-I) is a 7-point centered scale that integrates symptom severity and improvement over time as the primary outcome measure. Specific UCPPS-based anchors points for rating the CGI will be appropriately added as is the convention for other multi-symptom disease states [101] including chronic pelvic pain [102]. Global ratings of UCPPS symptom improvement yield a measure of overall multi-symptom benefit from treatment and is a core outcome domain in pain RCTs [103]. Those who score 2 (much improved) or better at follow-up qualify as categorical responders. Patients will complete the CGI at the three follow-up assessments and, for process analyses (Aim 3), weeks 3, 5, 7, 8, and 10 of the acute phase. The clinician version of the CGI will be completed at follow-up by “blind” MDs masked to treatment assignment to minimize bias and establish the validity of the patient version [48].

Secondary outcome measures

Secondary outcome measures include pelvic pain, urinary symptoms, pain interference, emotional distress, quality of life, and patient satisfaction using measures with confirmed psychometric properties (see Table 3). Severity of urinary symptoms and pelvic pain will be assessed using factorially derived items from the Genitourinary Pain Index (GUPI) [104] and the Interstitial Cystitis Symptom (ICSI) and Problem Indices (ICPI) [105]. Pain interference will be assessed using the PROMIS - Pain Interference scale (PPI SF-6a) [106], a 6-item instrument of the consequences of pain on relevant aspects of one’s life, including social, cognitive, emotional, physical, and recreational activities. Emotional distress will be measured using the Brief Symptom Inventory-18 (BSI-18) [107] which measures the level of distress across three dimensions (i.e., anxiety, somatization, and depression). The Client Satisfaction Questionnaire (CSQ) [108], an 8-item instrument measuring patient satisfaction with treatment, will assess the quality of care at immediate post treatment. Quality of life and co-morbid COPCs will be measured with the 12-item version of the SF-36 Health Status Questionniare [109] and 41-item Complex Multi-Symptom Inventory (CMSI) [110] respectively. UCPPS symptom measures (GUPI, ICPI, ICSI) will be assessed at baseline, interim (weeks 3, 5, 7, 8, and 10 of acute phase), and at all follow-up visits. As non-urological secondary outcomes, the BSI-18, CSQ, CMSI, SF-12, and PPI SF-6a will be measured at baseline and post treatment.

Table 3 Enrollment, intervention, and assessment schedule for primary outcomes, mediators, and predictorsMediators

The primary mediators are designed to tap into aspects of a rigid cognitive style central to a transdiagnostic conceptual model of centralized pain states such as UCPPS [111]. Mechanistic outcomes believed to drive CBT include context sensitivity, coping flexibility, repetitive thinking, self-distancing/perceived control, all of which are targeted by a different treatment module (see Table 2). All mediators will be assessed at baseline and follow-ups. The timing of additional mediator assessment is calibrated to the weeks when the corresponding skill believed to induce respective cognitive change is introduced and practiced. For example, because flexible problem solving is believed to improve symptom improvement by increasing sensitivity to contextual cues that promote coping (i.e., context sensitivity), the Context Sensitivity Questionnaire (CSI) [112] is administered at week 8 after the flexible problem-solving module is introduced and practiced. This approach differs from other assessment schedules of mechanistic studies when all putative mediators are assessed across different interim assessment periods (e.g., weeks 3, 5, 8) without regard to the mechanistic specificity of each strategy within a protocol [67]. By the same token, self-distancing, measured with the 11-item Experiences Questionnaire - Decentering (EQ-D) [113] and the Perspective Broadening scale of the Cognitive Emotion Regulation Questionnaire (CERQ-PB) [114] will be assessed at week 3 as will 4 scales (i.e., Noticing, Non-Distracting, Not-Worrying, Attention Regulation) of the Multidimensional Assessment of Interoceptive Awareness - 2 (MAIA-2) [115] measuring perceived control over aversive somatic sensations. Repetitive thinking will be assessed at week 7 and at all follow-ups with the Perseverative Thinking Questionnaire (PTQ) [116]. The Coping Flexibility Scale - Revised (CFS-R) [117], a 12-item instrument designed to measure discontinuation of ineffective coping strategies; re-coping, and meta-coping, will be assessed at baseline, week 10 and all follow-ups. Non-specific mediators common to both EDU and CBT include treatment expectancy (Credibility/Expectancy Questionnaire) [98] which will be assessed at the end of session 1 (week 1), while therapeutic alliance (Working Alliance Inventory) [118] which be assessed at weeks 1, 3, 5, 8, and 10.

Predictors and covariates

The primary theoretical-based predictor with prescriptive value (i.e., moderator) is trait self-regulation [119] which we will assess using the 75-item, self-report Behavior Rating Inventory of Executive Functioning - Adult (BRIEF-A) [120]. The BRIEF-A assesses habitual propensity with self-regulating or executive function within the context of everyday life. It is reliably associated with mental health, health behaviors, and physical health parameters [119]. Demographic variables, medication use, and disease characteristics (e.g., symptom severity, chronicity, treatment history, comorbidities) will be explored as general, non-specific predictors with prognostic value. Prescriptive and prognostic variables will be assessed at baseline.

Data collection and management

Sources of research material will include clinical data from structured interviews, self-report measures, physician assessments, and audio-recorded treatment sessions which will be used to establish therapist fidelity to treatment protocols. Clinical data will be captured using research electronic data capture (REDCap) software [95]. REDCap is a secure, Health Insurance Portability and Accountability Act compliant, web-based application designed to support data capture for research studies. Data will be collected for research purposes only and only with consent of the study volunteer released to a designated recipient (e.g., physician). Original hard copy source documents will be kept in study binders in locked cabinets or electronically stored on a secure server that will be encrypted and password protected. Data for analysis will be stored on a study-specific, password-protected database using subject numbers without personal identifiers. All data are backed up on external servers on a daily basis to a central secure data serve at UB. Access is password protected at multiple levels and no member of the EPPIC team apart from those with data management responsibility will have access to these passwords. Digital records of sessions will be stored in a secure, password-protected folder on the UB server. For all data, separate, an encrypted file linking names to trial ID will be kept and password-protected.

Statistical methods

We intend to interview all randomized individuals even if they drop out of treatment permitting straightforward ITT analyses. Missing data will be addressed using full information maximum likelihood (FIML) methods [121]. We expect that missing data are minimized by using electronic data capture systems that enable real-time data monitoring. We will test for attrition bias by comparing baseline scores for those who are lost to follow-up with those who complete follow-up. If analyses suggest missing data that violates missing at randomness, we will use Bayesian estimation or pattern modeling in place of FIML [122].

We will evaluate non-normality, variance heterogeneity, specification error, and outlier effects in all analyses. We generally will rely on robust methods of analysis (e.g., Huber-White robust standard errors in Mplus or bootstrapping) [122]. We will make clustering adjustments as implemented in the Mplus software, as needed. For all multi-item measures, we will evaluate composite reliability [123], concurrent validity, and discriminant validity. We will routinely test for unidimensionality and explore the factor structure of all multi-item scales. These analyses will dictate the formation of latent variables to accommodate variable inter-correlations and collinearity, as necessary. We will adjust for familywise error rates using a Holm-modified Bonferroni method [122, 124, 125] and will compare results with unadjusted contrasts in the spirit of sensitivity analyses.

Data analysis: statistical power analysis

The field has generally operationalized a clinically meaningful effect as a Cohen’s d of 0.50 or a correlation equivalent to it of 0.23, the latter of which represents about 5% explained variance [126,127,128]. Rather than framing power analysis as the probability of correctly rejecting a null hypothesis, we approach it in terms of effect size sensitivity. If a minimal clinically meaningful effect size is set at d = 0.50, it does not matter if we “miss” effect sizes less than d = 0.50 by failing to reject the null hypothesis for them because they are judged to be non-meaningful [129, 130].

Assuming conservatively a 12% failure to assess individuals, our final sample size of complete case data will be about 211 or about 105 per condition. For a traditional single degree of freedom contrast of means to evaluate paths from treatment to mediator and assuming a power of 0.80, the effect size sensitivity for a sample size of 100 per group is a population Cohen’s d of 0.39, which represents about 3.5% explained variance. This sensitivity will decrease somewhat if there is clustering, but the decrease will be offset by the use of baseline covariates. For a regression analysis to estimate paths from mediator to outcome with 8 predictors (to account for the simultaneous entry of mediators and covariates to control), and assuming a squared multiple correlation of 0.40, the effect size sensitivity for a given predictor will be the detection of a population coefficient that represents 3% unique explained variance. If we add a cluster adjustment representing an interclass correlation coefficient of 0.10, the sensitivity increases to 5% unique explained variance. These same statements apply for the analysis of moderation. The full information estimation structural equation model (SEM) we reference below will generally have equivalent if not greater statistical power than the above, so they, by definition, are adequately powered [131].

Data analysis: analysis of aims

Given the centrality of mediation and moderation in this proposal, we will use analytic methods based in SEM for RCTs [132]. We refer to our approach as a randomized explanatory trial (RET) because of its emphasis on the explanation of treatment effects on outcomes through mediation and moderation. The exogenous treatment condition is identified by a two-valued variable (0 = EDU, 1 = CBT). The endogenous outcome, our primary endpoint, is the CGI. Mediators are the primary transdiagnostic change mechanisms of the intervention. A powerful feature of RET analysis is that it pinpoints the specific facets where a program succeeds and where it falls short of eliciting mechanistic change that drives symptom relief. RET is not concerned so much with omnibus mediation effects but rather focuses on each link in a given mediational chain and identifies the link(s) where a chain is “broken” so that we learn what needs to be addressed technically to boost the therapeutic impact of CBT. See Fig. 2 for the RET schematic, which will be analyzed overall using SEM.

Fig. 2figure 2

Randomized experimental trial model of mediation analyses

Also of interest in an RET is the significance and magnitude of paths linking each mediator to the outcome. For example, perhaps a path associated with coping flexibility is not significant, suggesting that, contrary to our assumptions, coping flexibility does not meaningfully elicit UCPPS symptom improvement. In this case, we might consider streamlining the program so that it is simpler by eliminating the flexible problem solving. We will conduct state-of-the-art dominance analyses that allow us to order the relative strength of paths from the mediators to the outcome variable [133].

Analyses for Aim 1: Evaluate the efficacy of MC-CBT for UCPPS as compared to a nonspecific control intervention (EDU) in relieving pain and related symptoms. H1a states: Patients randomized to MC-CBT will show greater global symptom improvement on primary endpoint (CGI) compared to those randomized to EDU. This hypothesis will be tested using a covariate-adjusted single degree of freedom contrasts at each time point (immediate posttest 2 weeks, 3 and 6 months) comparing MC-CBT versus EDU on the full-scale CGI. Standard covariates will include medication history, gender, age, and indicators of symptom severity as measured at baseline. H1b states: MC-CBT will be superior to EDU on key secondary endpoin

留言 (0)

沒有登入
gif