Development of a core outcome set for treatment studies for provoked vestibulodynia

Abstract

Background

There is an inconsistency in treatment outcomes used in clinical trials for provoked vestibulodynia (PVD), which makes it impossible to compare the effects of different interventions.

Aim

In this study, we completed the first step in creating a core outcome set (COS), defining what outcomes should be measured in clinical trials for PVD.

Methods

Identification of outcomes used in studies was done by extracting data from clinical trials in a recently published systematic review and via review of clinical trials for PVD registered on ClinicalTrials.gov. The COS process consisted of 2 rounds of Delphi surveys and a consensus meeting, during which the final COS was decided through a modified nominal group technique.

Outcomes

Consensus on what outcomes to include in a COS for PVD.

Results

Forty scientific articles and 92 study protocols were reviewed for outcomes. Of those, 36 articles and 25 protocols were eligible, resulting in 402 outcomes, which were then categorized into 63 unique outcomes. Participants consisted of patients, relatives/partners of patients, health care professionals, and researchers. Out of 463 who registered for participation, 319 and 213 responded to the first and second surveys, respectively. The consensus meeting consisted of 18 members and resulted in 6 outcomes for the COS to be measured in all treatment trials regardless of intervention: insertional pain (nonsexual), insertional pain (sexual), provoked vulvar pain by pressure/contact, pain-related interference on one’s life, pain interference on sexual life, and sexual function.

Clinical Implications

Critical outcomes to be measured in clinical trials will allow for accurate comparison of outcomes across treatment interventions and provide solid treatment recommendations.

Strengths and Limitations

The major strengths of the study are the adherence to methodological recommendations and the intentional focus on aspects of diversity of participating stakeholders (eg, status such as patients with lived experience and researchers, inclusiveness with respect to sexual identity), the latter of which will allow for broader application and relevance of the COS. Among the limitations of the study are the low rate of participants outside North America and Europe and the lower response rate (about 50%) for the second Delphi survey.

Conclusion

In this international project, patients, health care professionals, and researchers have decided what critical outcomes are to be used in future clinical trials for PVD. Before the COS can be fully implemented, there is also a need to decide on how and preferably when the outcomes should be measured.

Introduction

Provoked vestibulodynia (PVD) is the most common subtype of vulvar pain in women,¹ affecting approximately 8% of women aged <40 years.² PVD is characterized by pain during sexual and nonsexual contact of the mucosa around the vaginal opening. The condition has major impacts on sexual health, relationship well-being, psychosocial adjustment, and overall quality of life.^3–5 Although underlying pathologic mechanisms of biomedical and psychosocial factors have been investigated, the etiology of PVD is still not fully understood. It is likely multifactorial in nature,⁵ and this complexity is reflected in the variety of treatments for PVD.^6–9

The Swedish Agency for Health Technology Assessment and Assessment of Social Services (SBU) recently published a systematic review of clinical trials of the effectiveness of treatment approaches for PVD.⁶ The conclusion of the systematic review was that there is a need for more methodologically stringent trials on interventions for PVD, which includes trials that measure appropriate outcomes. Selecting meaningful and consistently measured outcomes is crucial when designing clinical trials; this process allows for accurate representation of critical outcomes and the ability to directly compare the effects of different interventions.¹⁰

Although Pukall et al¹¹ provided recommendations for self-reported outcomes in vulvodynia trials based on the domains outlined by the Initiative on Methods, Measurement and Pain Assessment in Clinical Trials (IMMPACT), their recommendations were not based on a systematic review of the treatment literature and did not incorporate input from a range of stakeholders. Rather, they relied on expert input and a general overview of measures used in the literature overall, basing their recommendations on frequently used measures that target the IMMPCT domains. Using a systematic approach with input from multiple stakeholders to make recommendations for what to measure in treatment trials for PVD would ensure that issues related to potential bias were addressed.

Indeed, there has been increasing concern among vulvodynia researchers about possible outcome selection bias in trials that hinders synthesis and limits the potential to combine the results of individual studies into summary estimates.^6,11,12 In addition, the number and type of outcome measures vary widely, complicating the ability to accurately assess treatment outcome across studies. For example, Sadownik et al reported that their literature review resulted in a range of 1 to >20 outcomes per study for vulvodynia trials.¹⁰ In addition, they found that self-reported and clinician-assessed outcomes were measured differently, significantly contributing to the heterogeneity of outcome measures. The lack of agreement on what outcomes to measure was also illustrated by Davenport et al.¹³ Their review evaluated treatment outcomes in 25 vulvodynia studies, in which 26 outcomes had been measured by 110 outcome instruments.¹³

One way to overcome the potential for bias and the demonstrated heterogeneity is to develop a core outcome set (COS).¹⁴ A COS is defined by the Core Outcome Measures in Effectiveness Trials (COMET) Initiative as an agreed standardized set of outcomes that should be measured and reported, as a minimum, in all clinical trials in specific areas of health or health care (https://www.comet-initiative.org/).¹⁵ In addition to the COS, researchers may include other outcomes of relevance in their study. In creating a COS, the first step is to decide what to measure. Once agreement has been reached regarding what to measure, the next steps are to define how and when the selected outcomes should be measured.¹⁴ In 2017, the COS-STAD project (Core Outcome Set–Standards for Development) established recommendations for how COS should be developed and reported in a standardized way: COS-STAR (Core Outcome Set–Standards for Reporting).^16,17 Although COSs have been developed for clinical trials of various conditions,^12,18,19 no COS has been undertaken for clinical trials of PVD.

As a first attempt to provide guidelines for what should be measured as outcomes in vulvodynia clinical trials, recommendations to use the framework for relevant domains have been proposed as suggested by IMMPACT.^11,19 However, these recommendations have not been formally integrated into vulvodynia treatment trials. Given that the publication of SBU’s systematic review was part of a larger Swedish governmental assignment focusing on women’s health and that this review recommended that a COS be developed for PVD clinical trials,⁶ SBU was further assigned to develop a COS for PVD. A final COS includes what, how, and when to measure. In this study, we focused on the first step in creating a COS, defining what outcomes should be measured in future clinical trials for PVD. The COS is intended for clinical trials for all types of interventions and settings in those with PVD. In the continuation of the project, measurement instruments related to the different outcomes will be identified, and their psychometric properties will be assessed to inform recommendations on how and when to measure.

Methods

The project group consisted of staff from the SBU with expertise in COS, 2 external experts in PVD, and 1 patient representative. Briefly, the project consisted of the following steps: outcomes used in trials for PVD were identified, and relevant stakeholders (ie, patients, health care professionals, and researchers) took part in 2 rounds of Delphi surveys before the final COS was decided in a consensus meeting (Figure 1).

The methodology in the current study followed the guidelines detailed in “The COMET Handbook”¹⁴ and COS-STAD.¹⁶ The project was reported according to COS-STAR¹⁷ and was registered in the COMET Initiative database (http://www.comet-initiative.org/Studies/Details/2085),¹⁵ after a search of the database ensured that no past or current COS study for PVD was registered. Although 1 registered COS for PVD was identified, the SBU team confirmed with the lead investigator that it was no longer active. The study was approved by the Swedish ethical application board (2022-02091-01), and a protocol was developed a priori (https://osf.io/b9vcy/).

Systematic overview for identification of outcomes

All treatment outcomes in the relevant studies from the previous systematic review published by SBU were extracted.⁶ For the review, databases were searched from January 1, 1990, to January 29, 2021 (for complete search strategies, see reference Bohm-Starke et al).⁶ An additional search on ClinicalTrials.gov (https://www.clinicaltrials.gov/) was completed on February 3, 2022, for outcomes in study protocols for PVD (Table S1). Two persons in the project group (C.H., M.Ö.) screened the full text of all studies assessed by the previous systematic review as well as all study protocols from clinical trials for relevance according to the following eligibility criteria: population (ie, individuals aged ≥18 years with PVD), interventions (ie, interventions for treatment, cure, or relief of symptoms, such as pharmacologic, surgical, psychotherapeutic, and physical therapy options, were considered alone or combined), control interventions (no limitations were placed for control interventions), and study design (ie, randomized or quasi randomized controlled studies).

Data extraction

The following data from the studies and protocols were extracted for tabulation: reference information, study design, publication year, intervention, outcomes (defined as primary or secondary), measurement instrument, time of measurement, and percentage of participants who had a registered result of outcome measurement. If a study was published as a protocol in clinical trials as well as in a published article, the article was used. After extraction, all outcomes were listed and duplicates were removed. Similar outcomes in the list were combined; for example, “pain during insertion of vaginal dilator,” “tampon test pain intensity,” and “pain during insertion of a small speculum” were combined to “insertional pain (nonsexual).” The outcomes were then categorized according to the core and supplemental IMMPACT recommendations as well as IMMPACT-adapted outcomes for vulvodynia.¹¹ Additionally, some categories were added (eg, “other” was added to some categories to be as inclusive as possible). The categories consisted of the following: pain (pain intensity, quality and affect, pain temporality, other), physical functioning (health-related quality of life, sexual function, sexual distress, sexual function interference, sexual satisfaction, other), emotional functioning (depression, anxiety, responses to pain, other), participant ratings of improvement and satisfaction with treatment, symptoms and adverse events, participant disposition (eg, adherence, reasons for premature withdrawal from the trial), role functioning (work and educational activities), interpersonal functioning (relationships and activities with family, friends, and others), pharmacoeconomic measures and health care utilization, biological markers (eg, assessments based on quantitative sensory testing, imaging, genetic markers, pharmacogenomics, and punch skin biopsy), coping, clinician or surrogate ratings of global improvement, neuropsychological assessments of cognitive and motor function, suffering and other end-of-life issues, and other. Some of the categories (eg, suffering and other end-of-life issues, neuropsychological assessments of cognitive and motor function) did not map onto any of the outcomes, so they were not included in the survey. For the prioritization process, outcomes related to participant disposition were removed. The experts in the project group added definitions of each outcome before the list was sent to participants. The outcome list and all information were described in Swedish and English to enable international participation.

Recruitment of participants

Participants were recruited via social media advertisements (Twitter, LinkedIn, and Facebook) targeting national and international stakeholders, and relevant organizations were contacted to be part of the study and were asked to share information with their members (Appendix S1). Those interested in participating were directed to an application form on the SBU website (https://www.sbu.se), where they were provided with detailed information about the study and were asked to register for it by filling out their contact details, country of residence, and perspective (ie, stakeholder status). The last question offered the following choices: (1) individual aged >18 years with present or past history of PVD, (2) relative/partner of individual with present or past history of PVD, (3) health care professional (ie, physician [gynecologist, dermatologist, general practitioner, others]; physiotherapist; psychologist, sexologist, behavior therapist, etc; midwife; and other relevant health care professional), (4) researcher, and (5) other (eg, health care authority, research financier, health technology assessment employees, patient organizations). Before the first Delphi survey was sent out, a separate anonymous survey asking for additional background information (ie, age, gender, sexual orientation, relationship status, level of education) was sent to those who had registered for participation in the study to document the representativeness of the sample regarding its sociodemographic information.

Delphi surveys

Two rounds of Delphi surveys were conducted. Prior to the surveys, participants who registered on the SBU website for the study received an email with detailed information about the aims of the project, a description of the development of a COS, how the surveys were constructed, and explanations on how to complete them. The surveys were developed with Defgo, an online survey software program (defgo.com/uk/). Survey 1 was sent out October 2022 and was open for approximately 1 month. Up to 4 reminders were sent if no reply was received. In the surveys, participants were asked to rate each outcome on a 9-point Likert scale regarding the importance of the outcome to be included in clinical trials for PVD: 1 to 3 (the outcome is not important), 4 to 6 (the outcome is important but does not need to always be measured), 7 to 9 (the outcome is critical and should always be measured), and unable to score (do not understand the meaning or explanation of the outcome). Participants were also given the opportunity to add outcomes in the first survey round, and these were added in survey 2 if they were determined by the project team to be additional outcomes and not background information (eg, demographics and history of depression).

The results of survey 1 were analyzed, and outcomes were marked as consensus in or consensus out (Table 1). Survey 2 was sent out November 2022 to everyone who had registered on the SBU website, regardless of participation in survey 1. The second survey was open for 3 weeks, and up to 4 reminders were sent if responses were not received.

Consensus meeting

Participants were asked if they would be interested in taking part in the consensus meeting after the 2 rounds of Delphi surveys, in which the final outcomes in the COS would be determined. The consensus meeting took place via Zoom on January 18, 2023. English language was used, and the meeting opened with an initial briefing on its purpose and scope.

For the discussions during the consensus meeting, a modified nominal group technique was used.^22,23 Participants were divided in 3 subgroups lead by 1 or 2 members of the project group. All subgroups consisted of a balanced mix of patient representatives, health care professionals, and researchers from various countries. Each subgroup discussed the outcomes based on the following criteria: (1) outcomes that all agreed should be in the final COS, (2) outcomes that all agreed should not be in the final COS, and (3) outcomes that the group could not agree on.

Results

Identification of outcomes

In total, 40 scientific articles and 92 study protocols on ClinicalTrials.gov were reviewed for outcomes. Of those, 36 articles and 25 protocols were considered relevant for data extraction. Included and excluded studies are listed in Tables S2 and S3 with reasons for exclusion. The mean number of outcomes in the included studies was 6.5. It was therefore decided that no more than 6 outcomes should be in the final COS (Table 2). A flowchart of the process is illustrated in Figure 2.

From the studies and protocols, 402 outcomes were extracted and listed. Outcomes relating to methodological aspects and composite outcomes were excluded. Duplicates were removed and similar outcomes were combined before being categorized according to the IMMPACT recommendations,^21,24 resulting in 63 unique outcomes in survey 1 (Appendix S2).

Participants for the Delphi surveys and the consensus meeting

In total, 463 persons from 22 countries applied on the website for participation in the Delphi process. Of those, 6 were removed for various reasons. The majority came from North America and Europe, with only a few participants from the rest of the world (Table S4). All perspectives (ie, stakeholder statuses) that were assessed as being critical to include were represented (Table 3).

Sixty-nine percent of the participants answered the anonymous questionnaire on additional background information (Table S4). Data showed that the dominant gender in most perspective groups was “woman” and the mean age of participants was 40 years. Most participants were in a relationship and the majority had a high level of education. Regarding sexual identity, all categories were represented.

Seventy participants declared an interest in taking part in the consensus meeting. Of those, 20 were invited and 18 attended. The selection of the participants for the consensus group was based on gender, country, and perspective, with the aim of ensuring as much diversity of these variables within the group (Table S5).

Delphi surveys 1 and 2

All outcomes and the full results of the surveys are presented in Table S6. To be able to remove some outcomes after survey 1, an adjustment of the consensus criteria was made as it was deemed necessary (this was the only protocol deviation undertaken). According to the protocol, if ≥70% from all perspective groups had prioritized an outcome of 1 to 6, it should be removed from further prioritizing. None of the outcomes met those criteria in survey 1, and the cutoff was changed to ≥50%, at which point 10 outcomes could be removed (Table 1). The respondents of the first survey added 6 more outcomes to be prioritized in survey 2. The analyses of surveys 1 and 2 resulted in 19 outcomes with the highest ranking to be moved forward to the consensus meeting (Table 4).

Consensus meeting

Before the meeting, the participants were asked to rank the 19 outcomes from 1 to 19, with a rank of 1 being most important. The result of this ranking constituted the starting point of the discussions during the meeting. There were 3 rounds of 45-minute discussions in the subgroups. In between the rounds, the results from each subgroup were presented and discussed in the whole group. Last, there was a final discussion in the whole group with a voting procedure before the final COS was decided. The 6 outcomes in the COS are presented in Table 5.

Overall, outcomes specific to PVD were selected. All groups agreed that 3 outcomes for pain intensity should be in the COS (outcomes 1, 2, and 4). The importance of including a variety of outcomes related to pain was emphasized by patient representatives, pointing out that pain is the major symptom in PVD. During the discussions, it was highlighted that there is a difference in measuring insertional pain depending on whether it is in a sexual context or not. As an example, the insertion of a speculum during an appointment with a gynecologist and the insertion of a tampon were proposed to be different from sexual intercourse and sex toy use. Provoked vulvar pain by pressure/contact was also considered important, as it can be measured by clinicians in a standardized way or self-reported in response to noninsertional activities.

There was agreement on sexual function as a critical outcome. However, there was a discussion on whether “pain interference on sexual life” and “pain interference on one’s life” were overlapping outcomes. After voting in the whole group, both outcomes were selected with the comment that “pain interference on one’s life” has a broader meaning, including quality-of-life perspectives.

“Pain anxiety” and “pelvic floor function” were prioritized by at least 1 subgroup. After general discussion and voting, it was decided to omit these outcomes. Nevertheless, there was an overall recommendation that these outcomes should be measured as often as possible, depending on the intervention. Arguments supporting the inclusion of pain anxiety were that it is important to have an outcome that captures the psychological aspect of the condition and that treatments can focus on alleviating anxiety and teaching the patient to cope with the pain. Arguments against including this outcome were based primarily on the reasoning that it is not relevant to all interventions. Patient representatives stated that it was of greater importance to measure outcomes connected to pain, in the hopes of overcoming the pain and its impact on life activities rather than learning to manage the pain.

Participants in favor for “pelvic floor function” argued the importance of addressing the role of pelvic floor dysfunction in the treatment of PVD. Those against inclusion of this outcome raised the practical issue of measuring pelvic floor function, for instance, in psychological interventions and in internet-based trials.

The 3 outcomes involving treatment outcome, treatment satisfaction, and adverse events were left outside the COS, since these outcomes are usually measured in all treatment studies regardless of condition or disease (Table 5).

Following the consensus meeting, the results were sent to all participants who took part in the Delphi surveys, with the opportunity to give feedback. Only 1 general comment was received from 1 participant, expressing gratitude for the mission to create the COS. No further changes were made.

Discussion

In the present study, we used robust methods to agree on what to measure in a clinical trial on PVD, the first part in the development of a COS. After 2 rounds of Delphi surveys and a consensus meeting, the final COS consisted of 6 outcomes that should be measured in future clinical trials to improve comparisons of the effects of different interventions: insertional pain (nonsexual), insertional pain (sexual), provoked vulvar pain by pressure/contact, pain-related interference on one’s life, pain interference on sexual life, and sexual function.

Of the 6 core IMMPACT-recommended domains to measure in chronic pain clinical trials—pain, physical functioning, emotional functioning, participant ratings of improvement and satisfaction with treatment, symptoms and adverse events, and participant disposition in terms of adherence and reasons for premature withdrawal from the trial²¹—the latter 3 were not considered specific to PVD clinical trials as they should be and are often measured in all clinical trials. Three of the 6 domains determined in the present study related to pain (a core IMMPACT domain)²¹—specifically, pain intensity: insertional pain (sexual), insertional pain (nonsexual), and provoked vulvar pain by pressure/contact. Two outcomes related to physical functioning (pain interference on sexual life, sexual function), another core IMMPACT domain.²¹ IMMPACT recommends measuring physical functioning in a generic or disease-specific manner.²¹ Included in the Delphi surveys of the present study were generic domains (eg, health-related quality of life) and specific domains (eg, sexual function) related to physical functioning, and the 2 that emerged as critical in the present study related to sexuality (pain interference on sexual life, sexual function). This finding emphasizes the importance of assessing sexuality-specific outcomes in PVD clinical trials.¹¹

The domain of interpersonal functioning does not represent an IMMPACT core domain, but this domain was suggested as a supplementary one for vulvodynia clinical trials,¹¹ given the importance of interpersonal consequences of and contributors to PVD.²⁴ Within this domain, the outcome of pain-related interference on one’s life, defined as the degree to which the pain interferes with one’s life, emerged as critical in the present study. This construct may seem similar to that of a generic measure of health-related quality of life within the physical function domain as proposed by IMMPACT.²¹ However, the category of interpersonal functioning in the present study focused on relationships and activities with family, friends, and others; thus, this core outcome relates more to interpersonal connections than physical function. This finding underlines the importance of capturing the interpersonal experiences of those with PVD.

The heavy emphasis on pain intensity in the present COS underscored the importance of this outcome to all stakeholders, especially the patient representatives, and this emphasis is reflected in the treatment outcome literature in vulvodynia. For example, Sadownik et al,¹⁰ using the core IMMPACT domains as a guideline, reported numerous outcomes related to self-reported and clinical examination-recorded pain intensity: self-reported outcomes included pain severity with intercourse and pain with tampon insertion, and clinically based outcomes included pain sensitivity of the vestibule measured in a variety of ways (eg, cotton swab test, speculum insertion). In addition, aspects of physical functioning that have been measured in vulvodynia clinical trials featured in the systematic review by Sadownik et al emerged as critical in the present study (ie, pain interference on sexual life, sexual function). Given that measures of interpersonal functioning were not a part of their systematic review, as they are not considered a core IMMPACT domain, an examination of the systematic review by Davenport et al regarding outcomes in 25 vulvodynia trials, which did not use the IMMPACT domains as a framework, revealed that a few used measures of relationship adjustment,¹³ a specific domain within interpersonal functioning.

In the present study, a thorough overview was done to identify relevant outcomes in predominantly randomized controlled trials and registered protocols on ClinicalTrials.gov to be included in the first Delphi round. All outcomes in the final COS were previously measured with various frequencies. In Figure 3, the most frequently used outcomes of the COS are “insertional pain (sexual),” “provoked vulvar pain by pressure/contact,” and “sexual function” but only in approximately 50% of the studies. Another observation is that some of the selected outcomes are rarely used. In fact, “pain-related interference on one’s life” and “pain interference on sexual life” were measured in <7% of the studies. One explanation for the sparse use of these 2 outcomes in previous studies might be the current lack of relevant instruments. If this is the case, it will be addressed in the next part of the project, evaluating measurement instruments that have been used in PVD treatment studies. However, these findings speak to the importance of utilizing a systematic nonbiased approach and incorporating multiple stakeholders in the process of determining what to measure in the COS.

Two outcomes that were not prioritized for the consensus meeting, “pain catastrophizing” and “sexual distress,” were used in approximately 25% of the studies. This result likely reflects the inconsistency in outcome measures in previous trials, which further supports the importance of the present study. As for all COSs, it is important to emphasize that the selected outcomes constitute the basic set of outcomes to be measured in clinical trials and that there might be other critical or relevant outcomes to add to the COS depending on factors such as the goals of the study and the type of intervention. Therefore, if variables such as pain catastrophizing and sexual distress are important to measure in a specific trial, they can be added to the COS as outcome variables.

It is critical to note that “pelvic floor function” caused an extended discussion during the consensus meeting and was ruled out by voting. Pelvic floor dysfunction in patients with PVD has been demonstrated in several studies, and reduction of this dysfunction constitutes an important part of treatment according to expertise-based guidelines.^6,7,25 In fact, multimodal physical therapy for pelvic floor dysfunction was the only intervention with some support of evidence in the recently published systematic review.^6,26 Despite these arguments, there was concern that “pelvic floor function” in the COS could hamper the utilization of the COS in all clinical trials for practical reasons. For instance, measuring pelvic floor function in studies featuring predominantly psychological interventions could be problematic, since measures of pelvic floor function may not be part of the objectives in those studies. As this outcome will probably be in many future trials, we will assess measures of pelvic floor function in the how to measure step.

During the project, effort was made to adhere to the methodological recommendations,^14,16,17 which is a major strength of the study. In all stages of the process, the importance of maintaining diversity of participating stakeholders was acknowledged. We included participants with diverse sexual identities and adjusted the language to ensure inclusion so that the final outcomes can be broadly applicable to various sexual activities. Effort was also made to utilize terminology so that the COS can apply to all regardless of relationship status. Another important aspect is that patient representatives took part in the development of the COS, not only as participants in all steps of the process, but also in the project group. In discussions during the consensus meeting, the impact of each participant’s experience was evident in the process of deciding the final outcomes. However, despite the international representation and the diversity among the participants, it is impossible to rule out that a different group of participants would have decided on another set of outcomes.

Although careful planning of the study was undertaken, it has inevitable shortcomings. The response rate of the first Delphi survey was 69% of those who had registered on the SBU website. For the second survey, the response rate was 50% of the initial applicants, or 72% of those from the first survey. This result was somewhat disappointing, but it is within the range of response rates reported in similar studies of reproductive health.^20,27–29 Unfortunately, it was not possible to conduct dropout analyses, and the reasons for participants who did not take part in the surveys are therefore unclear.

The treatment of PVD usually involves several professions, and one of the biggest challenges in the consensus process was to reach out to relevant stakeholders from diversity and global perspectives (patients, professionals, researchers, countries). There was international participation in the Delphi rounds, but the representation was mainly from North America and Europe. Information on the SBU website and the surveys were written in Swedish and English, and participation could have been hampered due to language barriers. In the consensus group, there was a lack of relatives/partners of patients with PVD and an overrepresentation of physicians, which could have biased some of the results. To minimize a potential bias, the final COS was sent to all who took part in the Delphi surveys, inviting them to comment on the process and the results.

Another problematic issue was the rating of the outcomes in the surveys. In general, there was a reluctance to score outcomes as not critical (1-6 on the Likert scale). After survey 1, we had to amend the protocol to be able to remove 10 outcomes after the first survey. In future COS development studies, a modification of outcome prioritization could simplify this process.

The next steps to fulfill the development of a complete set of critical outcomes are to decide how and preferably when the outcomes should be measured. Part of this process was included in the protocol, and we are currently reviewing the literature and creating a list of studies investigating the psychometric properties for instruments and measurement tools related to PVD. However, these results will be presented in full in a separate report. This procedure is essential for the implementation of the COS, thus enabling unbiased comparisons of the effects of different interventions. Meanwhile, researchers can choose to use the decided-on outcomes using the instruments that they find suitable while the next steps of deciding how and when to measure are in progress.

Conclusions

In this international project, patients, researchers, and health care professionals have decided on what critical outcomes to be used in future clinical trials for PVD. The process adhered to recommendations for creating a COS, and the following outcomes were decided: insertional pain (sexual), insertional pain (nonsexual), provoked vulvar pain by pressure/contact, pain-related interference on one’s life, pain interference on sexual life, and sexual function. We encourage all investigators undertaking intervention research in this field to use these outcomes as a minimum to facilitate comparison among studies and to increase the potential for evidence synthesis across clinical studies. We expect that this work will help bring consistency and uniformity to outcome selection and reporting in treatment trials for PVD. However, before the COS can be fully implemented, there is a need to decide on how and preferably when the outcomes should be measured.

Acknowledgments

We thank all the stakeholders who participated in the Delphi study and consensus meeting.

Funding

The project was conducted within assignment of the Swedish Agency for Health Technology Assessment and Assessment of Social Services, and external funding was not sought or used.

Conflicts of interests

The authors report no conflicts of interest. All authors, as well as those included in the COS development consensus meeting, filed conflicts-of-interest forms used by Swedish governmental agencies before engagement. These are available upon request.

Data availability

Data are available on request.

References