Background
Multidimensional self-report measures of sexual function for women do not include the assessment of vulvar discomfort, limiting our understanding of its prevalence. In an effort to improve the measurement of patient-reported health, the National Institutes of Health funded the creation of the Patient Reported Outcomes Measurement Information System (PROMIS). This included the development of the PROMIS Sexual Function and Satisfaction measure, and version 2.0 of the Sexual Function and Satisfaction measure included 2 scales to measure vulvar discomfort with sexual activity.
Objectives
The objectives of the study were to describe the development of 2 self-reported measures of vulvar discomfort with sexual activity, describe the relationships between these scales and scales for lubrication and vaginal discomfort, and report the prevalence of vulvar discomfort with sexual activity in a large, nationally representative sample of US women.
Study Design
We followed PROMIS measure development standards, including qualitative development work with patients and clinicians and psychometric evaluation of candidate items based on item response theory, in a probability sample of 1686 English-speaking US adult women. We tested 16 candidate items on vulvar discomfort. We present descriptive statistics for these items, correlation coefficients among the vulvar and vaginal scales, and mean PROMIS scores with 95% confidence intervals separately by menopausal status for the 1046 women who reported sexual activity in the past 30 days.
Results
Based on the psychometric evaluation of the candidate items, we created 2 separate 4 item scales, one to measure labial discomfort and pain and one to measure clitoral discomfort and pain. Additional items not included in the scales assess pain quality, numbness, and bleeding. The correlations between the lubrication, vaginal discomfort, and the 2 vulvar discomfort measures ranged from 0.46 to 0.77, suggesting that these measures represent related yet distinct concepts. In our nationally representative sample, 1 in 5 US women endorsed some degree of vulvar discomfort with sexual activity in the past 30 days. Menopausal status was associated with lower lubrication and higher vaginal discomfort but not with vulvar discomfort.
Conclusion
The PROMIS Vulvar Discomfort with Sexual Activity–Labial and Vulvar Discomfort with Sexual Activity–Clitoral scales are publicly available for use in research and clinical settings. There is limited overlap between vulvar discomfort and lubrication or vaginal discomfort. The importance of measuring vulvar discomfort as part of a comprehensive assessment of sexual function is underscored by its prevalence.
Diagnosis and treatment of dyspareunia require the gynecologist or other health care provider to localize the patient’s discomfort and to determine its etiology. Whereas dyspareunia typically involves introital and deeper vaginal pain, clinical insight suggests that some women also complain of vulvar pain or discomfort.
Vulvar pain terminology has evolved over time, and an updated expert consensus on classification of persistent pain (vulvodynia) was recently published. Research has explored correlates and causes of vulvar pain, for example, dermatoses, and, in women with cancer, iatrogenic conditions including radiation changes, lymphedema, and estrogen suppression. Regardless of the cause, it is important to consider the role that vulvar discomfort plays in women’s experiences of sexual activity.
The most commonly used multidimensional patient-reported outcome measures of sexual function for women do not include specific references to the vulva. This gap in assessment limits the ability to understand women’s experiences with vulvar discomfort or pain during sexual activity. Likewise, the prevalence of vulvar discomfort with sexual activity in the US general population is unknown.
In this manuscript, first, we provide details about 2 new measures of vulvar discomfort and pain that are part of the Patient-Reported Outcomes Measurement Information System (PROMIS) Sexual Function and Satisfaction measure (SexFS) version 2.0 (v2). Second, we describe the relationships between these vulvar discomfort scales and the PROMIS scales for lubrication and vaginal discomfort. Finally, we report the prevalence of vulvar discomfort or pain with sexual activity in a large, nationally representative sample of US women.
Materials and Methods
This study was approved by the Institutional Review Board of the Duke University School of Medicine. All patient participants provided informed consent.
PROMIS SexFS v2 vulvar discomfort scales
In an effort to improve the measurement of patient-reported health, the National Institutes of Health funded the creation of the PROMIS, which included the development of the SexFS measures, version 1 and v2. In the next 2 subsections, we summarize the qualitative and psychometric development of the new measures of Vulvar Discomfort with Sexual Activity–Labial and Vulvar Discomfort with Sexual Activity–Clitoral.
Qualitative development
After a literature review to examine extant patient-reported outcome measures, we conducted 27 patient and 10 clinician focus groups. Initial patient focus groups focused on experiences of sexual function broadly, and general comments about pain and discomfort with sexual activity were common. However, we did not include a specific question about vulvar discomfort in these groups, and few patients spontaneously distinguished vulvar discomfort from vaginal discomfort in their comments.
Subsequent focus group discussions with clinicians (especially those specializing in sexual medicine) revealed a need to emphasize distinctions between vaginal and vulvar discomfort as well as the distinctions between labial and clitoral discomfort. These clinical experts noted that the different functions and sensitivities of these tissues and anatomic locations were important and advised that these distinctions were essential for the appropriate identification of symptoms, formulation of treatment needs, and assessment of clinical outcomes.
Accordingly, we generated 16 candidate questionnaire items about vulvar discomfort distinguishing between the labia and clitoris, with item form modeled on existing items about vaginal discomfort. We tested these items in cognitive interviews with patients to evaluate comprehension of the question stems and response options.
From these patient interviews, it was clear that modifications to define labia and clitoris within each item were necessary. Based on patient feedback, our approach was to add “(lips around the opening of your vagina)” whenever we used the word labia, to add “(clit)” after using the word clitoris, and to further distinguish these items from vaginal discomfort when possible by specifying “inside your vagina” for questions related to vaginal discomfort.
After making these changes, we tested the revised items with new patients to reevaluate comprehension and found that patients generally understood the revised items. The 16 revised items on vulvar discomfort were then included in item testing for the PROMIS SexFS v2.
Psychometric evaluation of items
We wanted to collect a representative sample on which to base our psychometric evaluation of the PROMIS SexFS items. In June 2013, English-speaking US adult women were recruited through the KnowledgePanel of the Gesellschaft für Konsumforschung (GfK), an online panel that uses a probability sample of US mailing addresses to select its participants.
For those individuals who are selected to the panel who do not already have Internet access, a laptop and Internet access are provided by GfK. For those individuals who are selected to the panel and already have Internet access, GfK provides a small incentive for survey completion (∼$6.00).
After a pilot test of the survey, 5039 female GfK panel members were invited to participate. Of those, 2219 viewed the informed consent form and 1498 completed the survey. The full sample available for psychometric evaluation included this general US sample of 1498 women plus an oversample of 188 women recruited after the general sample who reported recent oral, vaginal, or vulvar discomfort with sexual activity, based on responses to screener questions. There was no explicit refusal conversion process; instead, enrollment ended when the targeted sample size was met, based on preplanned psychometric analyses for measure development.
To keep the overall SexFS v2 measure as streamlined as possible, our measurement goal was to develop a single measure (scale) of vulvar discomfort that covered both labial and clitoral discomfort. However, we recognized from our initial qualitative work that this might not be a unidimensional construct; thus, we prespecified analyses to compare confirmatory models with labial and clitoral items on separate factors vs all items loading on a single factor. Following methods outlined by Reeve et al, we used a variety of approaches to assess unidimensionality and model fit. The detailed psychometric results have been published elsewhere.
Scoring the PROMIS SexFS
The PROMIS SexFS v2 measures use the T-score metric, on which a score of 50 on each scale corresponds to the average for sexually active adults in the US general population with an SD of 10. For each domain, higher scores represent more of that domain, so a higher score on the vulvar or vaginal discomfort scales indicates more discomfort, whereas a higher score on the lubrication scale indicates more lubrication.
Statistical analysis
For analyses used in measure development and in reporting the prevalence of responses to individual items, the full, unweighted sample was used (n = 1686). Where the sample is noted as representative, it refers to the weighted sample (n = 1757, of whom 1046 were sexually active and provided data on the vulvar discomfort items). This included weighting (by sex) to the October 2012 Current Population Survey based on age, race/ethnicity, education, income, Census region, metropolitan status, and Internet access. Statistical analyses of this group adjusted for the sample design.
We present mean PROMIS scores with 95% confidence intervals based on the mean estimate and SEM for each of the 4 domains, separately by self-reported menopausal status and compared using a Wald F test in survey regression analysis. To examine overlap among the vulvar and vaginal measures, we calculated Pearson product moment correlations among the 4 scale scores. We used SAS version 9.4 (SAS Institute, Cary, NC) and a 2-tailed significance level of P = .05 for all assessments.
Results
Sample
Sociodemographic and health characteristics were generally similar between the full unweighted sample and the representative weighted samples ( Table 1 ). However, notable differences included age and menopausal status. The mean age was 49 years in the unweighted sample compared with 47 years in the weighted total sample and 43 years in the weighted sexually active sample. Sixty-two percent of women were menopausal or perimenopausal in the unweighted sample compared with 56% in the weighted total sample and 50% in the weighted sexually active sample.
| Characteristic | Full sample (unweighted), n, % b | Representative sample (weighted), n, % c | Representative sample (weighted), sexually active only, n, % c |
|---|---|---|---|
| Total | 1686 | 1757 | 1046 |
| Age, y | |||
| 18–29 | 284 (17%) | 352 (20%) | 244 (23%) |
| 30–44 | 368 (22%) | 434 (25%) | 325 (31%) |
| 45–59 | 534 (32%) | 495 (28%) | 319 (31%) |
| 60 or older | 500 (30%) | 477 (27%) | 158 (15%) |
| Race | |||
| White | 1324 (83%) | 1371 (78%) | 804 (77%) |
| Black/African-American | 169 (11%) | 234 (13%) | 148 (14%) |
| Asian | 27 (2%) | 58 (3%) | 41 (4%) |
| Multiple races or other | 80 (5%) | 95 (5%) | 42 (4%) |
| Hispanic or Latino ethnicity | 141 (8%) | 207 (12%) | 139 (13%) |
| Sexual identity | |||
| Heterosexual or straight | 1596 (95%) | 1652 (94%) | 972 (94%) |
| Gay/lesbian | 23 (1%) | 29 (1%) | 20 (2%) |
| Bisexual | 49 (3%) | 55 (3%) | 43 (4%) |
| Other | 9 (1%) | 7 (0%) | 4 (0%) |
| Educational attainment | |||
| Less than high school | 110 (7%) | 186 (11%) | 105 (10%) |
| High school graduate/GED | 525 (31%) | 529 (30%) | 274 (26%) |
| Some college | 534 (32%) | 527 (30%) | 329 (32%) |
| Bachelor’s degree or higher | 517 (31%) | 516 (29%) | 339 (32%) |
| Relationship status | |||
| Married or in a civil union or domestic partnership | 954 (57%) | 871 (50%) | 617 (59%) |
| Living with a partner | 146 (9%) | 182 (10%) | 149 (14%) |
| In a relationship but not living together | 129 (8%) | 144 (8%) | 110 (11%) |
| Dating | 68 (4%) | 89 (5%) | 73 (7%) |
| Single | 309 (18%) | 380 (22%) | 79 (8%) |
| Other | 73 (4%) | 81 (5%) | 11 (1%) |
| Employment status | |||
| Working (employee or self-employed) | 836 (50%) | 857 (49%) | 564 (54%) |
| On temporary layoff or looking for work | 166 (10%) | 215 (12%) | 153 (15%) |
| Retired | 359 (21%) | 342 (19%) | 119 (11%) |
| Disabled | 119 (7%) | 122 (7%) | 58 (6%) |
| Other | 207 (12%) | 221 (13%) | 152 (15%) |
| Household income | |||
| <$25,000 | 298 (18%) | 364 (21%) | 180 (17%) |
| $25,001–50,000 | 420 (25%) | 426 (24%) | 228 (22%) |
| $50,001–75,000 | 330 (20%) | 319 (18%) | 186 (18%) |
| $75,001–100,000 | 265 (16%) | 283 (16%) | 198 (19%) |
| >$100,000 | 373 (22%) | 364 (21%) | 254 (24%) |
| Health conditions | |||
| Arthritis or rheumatism | 470 (28%) | 429 (24%) | 203 (20%) |
| Hypertension | 448 (27%) | 442 (25%) | 191 (18%) |
| Depression | 390 (23%) | 375 (21%) | 206 (20%) |
| Anxiety | 335 (20%) | 300 (17%) | 161 (15%) |
| Migraines or severe headaches | 331 (20%) | 301 (17%) | 196 (19%) |
| Diabetes (type 1 or type 2) | 153 (9%) | 149 (8%) | 69 (7%) |
| Heart disease | 143 (9%) | 124 (7%) | 56 (5%) |
| Cancer (other than nonmelanoma skin cancer) | 116 (7%) | 113 (6%) | 46 (4%) |
| Self-rated health | |||
| Excellent | 239 (14%) | 285 (16%) | 206 (20%) |
| Very good | 605 (36%) | 622 (35%) | 408 (39%) |
| Good | 531 (32%) | 531 (30%) | 283 (27%) |
| Fair | 239 (14%) | 234 (13%) | 109 (11%) |
| Poor | 49 (3%) | 54 (3%) | 30 (3%) |
a Not all fields add to 100 because of rounding and weighting
b Full sample available for psychometric evaluation that includes the general population sample and an oversample from the general population of women with oral, vaginal, or vulvar discomfort with sexual activity based on responses to screener questions
c General population sample weighted to the October 2012 Current Population Survey.
PROMIS SexFS vulvar discomfort scales
Psychometric evaluation of the new measures indicated acceptable reliability and model fit ( Table 2 ). We found that separate dimensions (scales) for clitoral and labial discomfort were more appropriate than a unidimensional model that combined them into 1 scale. Best model fit for each of the scales was achieved by including the items on frequency and intensity of pain and discomfort with sexual activity, which is conceptually similar to many other measures of bodily pain. Items on pain quality, numbness, and bleeding can be used as individual items to provide additional context and/or detail but were not included in the final PROMIS scale scores.
| Instrument name | Vulvar discomfort, clitoral | Vulvar discomfort, labial |
|---|---|---|
| Number of items | 4 | 4 |
| Cronbach’s α a | 0.96 | 0.97 |
| Comparative fit index b | 1.00 | 1.00 |
| Tucker-Lewis index c | 0.99 | 1.00 |
| Root-mean-square error of approximation d | 0.04 | 0.06 |
a Measure of reliability; values greater than 0.70 were considered acceptable
b Measure of model fit; values above 0.95 were considered acceptable
c Measure of model fit; values above 0.95 were considered acceptable
d Measure of model fit; values below 0.06 were considered acceptable.
Correlations among vulvar and vaginal scales
The magnitudes of the correlations between the PROMIS lubrication, vaginal discomfort, and the 2 vulvar discomfort measures suggest that these measures represent related yet distinct concepts ( Table 3 ). The correlation among women who had any nonzero response (that is, anything greater than none or never) within any of the 4 measures was highest between scores on the clitoral and labial discomfort domains (r = 0.77), indicating 59% shared variance (r 2 ) between the two, which means 41% of the differences among women in clitoral discomfort are not accounted for by differences in labial discomfort and vice versa. The correlation between labial and vaginal discomfort was higher than the correlation between clitoral and vaginal discomfort, and the correlations between either vulvar discomfort scale and lubrication were notably lower than the correlation between vaginal discomfort and lubrication.
