- Email: [email protected]
The Vulvar Pain Assessment Questionnaire: Factor Structure, Preliminary Norms, Internal Consistency, and Test-Retest Reliability
ORIGINAL RESEARCH
PAIN
The Vulvar Pain Assessment Questionnaire: Factor Structure, Preliminary Norms, Internal Consistency, and Test-Retest Reliability Emma Dargie, PhD, Ronald R. Holden, PhD, CPsych, and Caroline F. Pukall, PhD, CPsych
ABSTRACT
Background: The Vulvar Pain Assessment Questionnaire (VPAQ) was developed to assist in the assessment and diagnosis of chronic vulvar pain (vulvodynia). Aim: To further establish the psychometric properties of the VPAQ by examining factor structure, test-retest reliability, internal consistency, and scale normative data, and to gather feedback from those with vulvar pain about the usefulness and accessibility of the questionnaire. Methods: 182 participants completed a confidential online study and 70 participated again at time 2 (4 weeks later). Outcomes: Participants were asked to complete the full VPAQ, which assesses pain characteristics, effects on various parts of their lives, coping strategies used, and romantic partner factors. Additional questions captured sociodemographics and feedback about the instrument. Results: Exploratory structural equation modeling indicated that the previously established subscales, except the coping scale, had adequate model fit, and all items loaded significantly onto relevant factors. Pearson product moment correlations (r ¼ 0.57e0.96) established strong 4-week test-retest reliability for most subscale scores, and Cronbach a indicated overall acceptable to high internal consistency (a ¼ 0.56e0.95). Preliminary norms for the scales are supplied. Approximately half the participants reported an increase in their comfort level in discussing a range of topics after completing the VPAQ. Most participants reported that the length, readability, and range of VPAQ questions were “good” or “excellent.” Clinical Implications: The results of this study provide further justification for using the VPAQ scales in clinical and research settings, preliminary norms for a vulvar pain population, and suggestions for interpretation. Strengths and Limitations: This study established the psychometric properties of the VPAQ scales using multiple methods at 2 time points and gathered feedback from participants. However, data were collected online so diagnoses could not be confirmed and more than half the initial sample did not complete the survey at time 2. Conclusion: The results of this study suggest that most VPAQ subscales (except the coping subscale) have moderate to strong psychometric properties and that the VPAQ is user friendly. Dargie E, Holden RR, Pukall CF. The Vulvar Pain Assessment Questionnaire: Factor Structure, Preliminary Norms, Internal Consistency, and Test-Retest Reliability. J Sex Med 2017;14:1585e1596. Copyright 2017, International Society for Sexual Medicine. Published by Elsevier Inc. All rights reserved.
Key Words: Vulvodynia; Dyspareunia; Pain Assessment; Factor Analysis; Reliability
INTRODUCTION Millions suffer from vulvodynia (persistent vulvar pain lacking a clear identifiable cause), with prevalence estimates ranging from 8% to 15%.1e3 However, many have difficulty obtaining an accurate diagnosis and effective treatment and need to seek help Received May 22, 2017. Accepted October 31, 2017. Department of Psychology, Queen’s University, Kingston, ON, Canada Copyright ª 2017, International Society for Sexual Medicine. Published by Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jsxm.2017.10.072
J Sex Med 2017;14:1585e1596
from more than 3 providers before securing diagnoses and/or finding some relief.2,4e6 Although multiple factors play a role in this process, 1 primary concern is assessment and diagnosis by health care providers. Unfortunately, limited time is devoted to chronic pain or sexual health during medical training, perhaps leading to discomfort when working with patients with pain during sexual activity.7e13 It is not surprising that many women with vulvodynia consult at least 3 health care providers before obtaining an accurate diagnosis or pain relief.2,4e6,14 A possible issue contributing to this process is the lack of a comprehensive tool to assess and diagnose vulvodynia symptoms. 1585
1586
Thus, our group created the Vulvar Pain Assessment Questionnaire (VPAQ).15 We administered a large pool of questions to women with chronic vulvar pain and used the construct validation approach16 to select salient, relevant items for VPAQ subscales through a series of factor analyses. Although the initial motivation behind the creation of this questionnaire was to assess those with idiopathic chronic vulvar pain (ie, vulvodynia), the questionnaire domains are applicable to chronic vulvar pain resulting from a range of causes (eg, lichen sclerosus, interstitial cystitis or painful bladder syndrome, or chronic yeast infections). Indeed, the primary questionnaire assesses pain characteristics, pain severity, cognitive and emotional factors, and interference with life, sexual function, and self-stimulation and penetration. A brief version also was created to assist in expedited pain assessment. Further scales were created to capture pain characteristics, coping strategies, and romantic partner factors. Clinicians and researchers could select scales to administer based on the needs of the vulvar pain population being targeted. The VPAQ can be used to gather information on symptoms and diagnosis and can be used to identify possible treatment targets.15 When correlated with other established questionnaires, evidence of convergent and discriminant validity was observed. For example, the sexual functioning subscale of the VPAQ was strongly related to scores on the Female Sexual Function Index but only mildly or moderately related to scores on other instruments.15 For each subscale, average scores can be computed to determine which areas are of greatest concern, and we hypothesize that scores can be tracked over time to help measure treatment progress. The results of this scale construction study were quite promising, although further research is required to replicate past results, examine psychometric properties in more detail, and gather feedback on the newly constructed VPAQ scales from those with chronic vulvar pain. The goal of the present study was to further test the VPAQ by (i) confirming its factor structure, (ii) investigating test-retest reliability, (iii) replicating internal consistency findings, (iv) providing normative data for all scales, and (v) gathering feedback from participants on the usefulness and accessibility of the inventory.
METHODS This online study was approved by the university’s general research ethics board, and participants provided informed consent. Participants were recruited through word of mouth, online advertisements, and postings to relevant listservs and groups. Similar to the original study on scale construction,15 anyone older than 18 years, with access to the Internet, and who reported experiencing chronic vulvar pain was invited to participate. Once participants reached the secure survey website, they read a letter of information; if they consented to participate, then they completed the survey consisting of sociodemographic characteristics, the VPAQ, and feedback about the VPAQ items.
Dargie et al
At the end of this survey (time 1), participants were invited to sign up for time 2 by providing their e-mail address in a separate survey, in addition to a unique identifier that would enable us to link their data at the 2 time points without knowing their identity. Approximately 4 weeks after initial completion, participants were invited to complete the VPAQ a second time (time 2). Each survey took approximately 15 to 30 minutes. To thank them for their time, participants were invited to enter a draw for 1 of 4 prizes valued at 50 CAD that took place at the end of the study. Before analysis, each participant’s responses were examined for eligibility, random responding, and excessive missing data. 182 participants provided complete data for time 1, and 70 participants provided complete data for time 2 (Figure 1 presents information about those who did not complete the study).
Main Outcome Measures The VPAQ15 is composed of a comprehensive questionnaire that covers a wide range of symptoms related to chronic vulvar pain, a brief version of this broad measure, a scale devoted to pain descriptors, a scale for coping strategies, and a scale for romantic partner factors. Information about the development of the following scales can be found in the article on scale construction.15 In addition to the VPAQ scales, participants were asked to report how answering the questions changed their comfort in talking about various topics with health care professionals (eg, asking for help about their vulvar pain, talking about what the pain feels like, and sharing how being in pain affected their emotional health). Change in comfort in discussing each topic was rated on a scale from 0 (no change in comfort) to 4 (a lot more comfort). Participants also were asked to rate the VPAQ questions for length, readability, and range of questions. These characteristics were rated on a scale from 1 (poor) to 4 (excellent). Full Version (VPAQfull) The full version of the questionnaire consists of 55 items rated on 5-point scales with anchors tailored to the type of question and 8 categorical questions that assess the onset, location, temporal pattern, degree of burning pain, and associated symptoms (eg, itching) of vulvar pain. The domains encompassed by the questionnaire include cognitive and emotional reactions to the pain, pain intensity, pain unpleasantness and discomfort, life interference, sexual function interference, and self-penetration interference. In the original study on scale construction, internal consistency was good for the VPAQfull subscale scores (a > 0.77). Due to a programming error in the present study, 1 item of the emotional response subscale was not included (ie, the item assessing how much the participant felt “like giving up”). Thus, analyses were computed using a 14-item version of the emotional response scale, rather than the original 15-item scale. Screening (VPAQscreen) This brief version of the questionnaire consists of 30 items rated on 5-point scales with anchors tailored to the type of J Sex Med 2017;14:1585e1596
1587
Vulvar Pain Assessment Questionnaire Viewed study Letter of Infor mation (n = 1146) Did not consent to par ticipate (n = 859) Consented to par ticipate (n = 287)
Repeat r esponses (n = 3) Incomplete Data (n = 91) Pain-fr ee (n = 8) Male (n = 2) Under 18 year s (n = 1)
Completed Time 1 (n = 182) Did not complete Time 2 (n = 112) Completed Time 2 (n = 70)
Figure 1. Flowchart illustrating the number of participants who completed each stage of the study, and reasons for exclusion.
questions and the same 8 categorical questions in the VPAQfull that assess the temporal nature and quality of the pain. The domains addressed by the VPAQscreen are the same as those addressed by the VPAQfull, although the cognitive and emotional aspects of vulvar pain are collapsed into 1 factor. In the original study on scale construction, the internal consistency for VPAQscreen subscale scores was good (a > 0.77), similar to the internal consistency when the screening tool was implemented in a clinical context (a > 0.80; manuscript in preparation). Pain Descriptors (VPAQdesc) The VPAQdesc captures common words used to describe chronic vulvar pain. The 10 items, rated on a scale from 0 (not at all) to 4 (very much), allow participants to indicate the degree to which each descriptor applies to their pain. 3 subscales compose the VPAQdesc: burning and stinging pain, incisive pain, and sensitivity. In the original study on scale construction, the internal consistency of the VPAQdesc subscale scores was adequate to good (a > 0.63).
J Sex Med 2017;14:1585e1596
Coping Strategies (VPAQcope) This scale addresses coping strategies that are commonly used by those with vulvar pain. It consists of 12 items rated on a scale from 0 (never) to 4 (always), allowing participants to indicate the frequency with which they use such strategies. Items are grouped into 2 categories: distraction- and relaxation-based strategies and active problem-solving strategies. In the original study on scale construction, the internal consistency for the VPAQcope subscale scores was adequate to good (a > 0.69). Partner Factors (VPAQpartner) This 24-item scale assesses how romantic partners and spouses might be affected by or respond to the vulvar pain experienced by their partner. Each question is rated on a 5-point scale with anchors based on question type. 4 subscales can be calculated: how the partner or spouse responds to the pain, support-seeking behaviors, the impact of the pain on aspects of the relationship, and communication about sexual matters when experiencing pain. In the original study on scale construction, the internal consistency for the subscale scores was good (a > 0.77).
1588
Statistical Analysis Analyses were conducted using SPSS 22.0 (IBM Corp, Armonk, NY, USA) and MPlus 7.4 (http://www.statmodel.com/). After examining the data for appropriate assumptions, exploratory structural equation modeling was conducted to examine the quality of the factor structure for each part of the VPAQ. To evaluate each subscale, 3 techniques were used.17 (i) Model fit indices were examined to determine how well the proposed factor structure fit the data collected during the current study. (ii) Significance of the factor loadings was examined to ensure that each item loaded significantly onto the appropriate factor (t > 1.96, P < .05). (iii) Each factor loading parameter was examined for interpretability. To evaluate model fit, we examined patterns across the root mean square error of approximation, the standardized root mean square residual, the comparative fit index, and the Tucker-Lewis index. Several sources were consulted to determine appropriate criteria for evaluating these tests of model fit,17e19 with the conclusion that acceptable fit could be established with a root mean square error of approximation and a standardized root mean square residual less than 0.10 (good fit < 0.05) and a comparative fit index or Tucker-Lewis index greater than 0.90 (good fit > 0.95). Test-retest reliability was computed for each subscale by examining the correlation between corresponding VPAQ scores at the 2 time points (ie, Pearson r was computed). Preliminary norms were established from this sample based on mean subscale scores. This approach was used instead of creating percentile ranks because the entire sample consisted of participants with vulvar pain and did not include those without vulvar pain. Typically, percentile ranks are used to indicate when the presence of symptoms is atypical for the general population, rather than characterizing variations in those who report such symptoms. Furthermore, use of mean scores is more accessible and easy to interpret for those less familiar with the use of percentiles. In addition, descriptive statistics were examined for feedback participants gave about its content. According to published guidelines,20 the sample size in the present study was adequate to obtain 0.80 power for the VPAQfull, VPAQscreen, and VPAQpartner model fit analyses. However, the VPAQcope and VPAQdesc model fit analyses were underpowered. For test-retest reliability, the size of the present sample had adequate power (0.80) to detect large effects when a was set at 0.01.21
RESULTS The age of participants who completed time 1 ranged from 18 to 85 years, and the average age was 36.13 years (SD ¼ 15.89; median ¼ 29.00). The age of participants who completed time 2 ranged from 20 to 78 years, and the average age was 36.59 years (SD ¼ 15.42; median ¼ 30.50). Other sociodemographics and self-reported diagnoses are presented in Table 1. No differences were found when comparing those who completed time points 1
Dargie et al
and 2 with those who completed only time 1 on these sociodemographic characteristics (P > .23 by Fisher exact test for all comparisons).
Exploratory Structural Equation Modeling Although a confirmatory factor analysis (CFA) approach is common for scale development and validation, some have argued that a traditional CFA can be overly conservative when examining scales that have factors related to one another.22 Indeed, traditional CFA requires that items be permitted to load onto only 1 factor, and cross-loadings are restricted to 0. The results we reported in the original study on scale construction indicated weak to moderate correlations among some factors,15 suggesting that a CFA approach would not be appropriate for further examining model fit in the present study. Thus, exploratory structural equation modeling procedures were examined using MPlus,23 which allows for indices of model fit to be computed with non-0 cross-loadings. Based on this approach, most VPAQ subscales had good or acceptable model fit (Table 2). In addition, most items loaded significantly onto the proposed factors in the hypothesized direction (Tables 3e7). For the VPAQfull, the most problematic item captured worries that the participant’s pelvic muscles would be too tight. This item did not load significantly onto the cognitive responses to pain factor, as hypothesized. Rather, it loaded significantly onto the emotional responses to pain factor. There also were 6 items with significant cross-loadings onto other factors (>0.30), although the largest factor loading was on the hypothesized factor. For the VPAQscreen, all items loaded significantly onto the hypothesized factor. However, 4 items had significant cross-loadings onto other factors (>0.30), although only 1 (painful tampon use) loaded more strongly onto another factor (cognitive and emotional response) than the hypothesized factor (self-penetration interference). For the VPAQdesc, there was only 1 item with a significant cross-loading (>0.30); the item capturing stinging pain loaded on the burning and stinging pain subscale in addition to the incisive pain subscale, although the loading was stronger on the former. The VPAQcope was the most problematic of the VPAQ scales. 3 items originally part of the distraction or relaxation subscale (going to one’s “happy place,” doing something to take one’s mind off the pain, and focusing on staying optimistic) loaded significantly onto the second factor, the active problemsolving subscale. Further, there were no significant crossloadings (>0.30) for the VPAQpartner, and all items loaded onto the hypothesized factors.
Internal Consistency, Norms, and Test-Retest Reliability Internal consistency for the time 1 data was assessed using Cronbach a. Only the burning and stinging subscale of the VPAQdesc had scores for which reliability was not adequate, good, or excellent (Table 3). Of note, scores on the burning and J Sex Med 2017;14:1585e1596
1589
Vulvar Pain Assessment Questionnaire
Table 1. Sociodemographic characteristics and self-reported diagnoses at time 1 and time 2 (w4 weeks later)* Time 1, n (%) Sexual orientation Heterosexual Gay or lesbian Bisexual Queer, asexual, other Gender identity Woman Transwoman, androgynous, other Decline response Relationship status Married or common law Dating Not dating Other or declined response Children 0 1 2 3e5 Declined response Employment Full time Part time Retired Student Disability, unemployed, medical leave Other (eg, housewife, home parent) Education [Some] high school [Some] trade school [Some] college or undergraduate [Some] graduate or professional Diagnosis Vulvar or genital pain but no diagnosis Vulvodynia Vulvar vestibulitis syndrome Provoked vestibulodynia Generalized vulvodynia Vaginismus Genito-pelvic pain or penetration disorder Pelvic floor muscle dysfunction Interstitial cystitis Chronic pelvic pain Irritable bowel syndrome
152 2 22 6
(83.5) (1.1) (12.1) (3.3)
177 (97.3) 4 (2.1) 1 (0.5)
Time 1, n (%)
Time 2, n (%) 58 1 8 3
(82.9) (1.4) (11.4) (4.3)
67 (95.7) 3 (4.2) 0 (0.0)
74 75 30 3
(40.7) (41.2) (16.5) (1.6)
27 25 17 1
(38.6) (35.6) (24.3) (1.4)
137 12 17 15 1
(75.3) (6.6) (9.3) (8.2) (0.5)
58 8 5 4 0
(75.7) (11.4) (7.1) (5.7) (0.0)
81 27 15 28 20
(44.5) (14.8) (8.2) (15.4) (11.0)
32 10 4 10 9
(45.7) (14.3) (5.7) (14.3) (12.9)
11 (6.0)
5 (7.1)
12 (6.5) 4 (2.1) 91 (50.0)
4 (5.7) 1 (1.4) 32 (45.7)
75 (41.2)
33 (47.1)
22 (12.1)
11 (15.7)
96 47 34 15 29 3
(52.7) (25.8) (18.7) (8.2) (15.9) (1.6)
36 24 15 8 13 2
(51.4) (34.3) (21.4) (11.4) (18.6) (2.9)
33 10 19 23
(18.1) (5.5) (9.9) (12.6)
15 4 11 11
(21.4) (5.7) (15.7) (15.7)
(continued)
J Sex Med 2017;14:1585e1596
Table 1. Continued
Endometriosis Lichen sclerosus Pelvic inflammatory disease Pudendal nerve entrapment Other (eg, vulvar dermatitis, postherpetic pain, polycystic ovarian syndrome)
8 7 2 10 13
Time 2, n (%)
(4.4) (3.8) (1.1) (5.5) (7.1)
1 4 0 1 6
(1.4) (5.7) (0.0) (1.4) (8.6)
Time 1 ¼ 1st completion of the Vulvar Pain Assessment Questionnaire; Time 2 ¼ 2nd completion of the Vulvar Pain Assessment Questionnaire. *Participants were invited to select as many diagnoses as applied to them.
stinging items were only moderately correlated (r ¼ 0.39) compared with a strong correlation between the 2 items that compose the incisive subscale (r ¼ 0.72). Means and SDs were computed for each time 1 VPAQ subscale (possible range ¼ 0e4), and independent-samples t-tests were used to determine whether those who also completed time 2 differed significantly from those who did not; there were no significant differences (P > .05 for all comparisons). Next, Pearson product-moment correlations were computed between subscale scores at times 1 and 2. Most of these relations were strong, except for the burning and stinging and incisive subscales of the VPAQdesc and the relationship impact and communication subscales of the VPAQpartner, which were moderate (Table 8).
Participant Feedback After completing the VPAQ during time 1, participants were asked whether doing so had changed their level of comfort in speaking to medical professionals about various topics. Many participants reported an increase in comfort when it came to asking for help about their vulvar pain (n ¼ 72; 40.0%), talking about what their pain feels like (n ¼ 89; 48.9%), being able to explain their overall experience of pain (n ¼ 91; 50.0%), and sharing how being in pain affects their emotional health (n ¼ 90; 49.5%), physical health (n ¼ 77; 42.3%), and sexual life (n ¼ 95; 52.2%). Table 2. Model fit indices for the VPAQ derived from exploratory structural equation modeling procedures using a geomin rotation VPAQ
df
RMSEA
CFI
TLI
SRMR
Full version Screening Pain descriptors Coping strategies Partner factors
1122 295 18 43 186
0.07 0.07 0.07 0.13 0.11
0.85 0.92 0.97 0.78 0.85
0.81 0.88 0.93 0.67 0.78
0.04 0.04 0.03 0.08 0.05
CFI ¼ comparative fit index; df ¼ degrees of freedom; RMSEA ¼ root mean square error of approximation; SRMR ¼ standardized root mean square residual; TLI ¼ Tucker-Lewis index; VPAQ ¼ Vulvar Pain Assessment Questionnaire.
1590
Dargie et al
Table 3. Geomin-rotated factor loadings for the full version of the Vulvar Pain Assessment Questionnaire* Factor 1 Pain severity 1. Average pain intensity 2. Worst pain intensity 3. Average unpleasantness 4. Worst unpleasantness 5. Average distress 6. Worst distress Feeling 7. Sad 8. Unable to make changes in my life 9. Bad about myself because of the pain 10. Emotionally exhausted because of the pain 11. Anger toward my pain 12. That the pain will never stop 13. Depressed 14. Like my body has let me down 15. Physically tense 16. That I am not a worthwhile person 17. Distracted 18. Hateful things about myself as a person 19. Stressed about the pain 20. That it is unfair that I have pain Thinking and worrying 21. That people might think I’m a bad sexual partner 22. That my partners might think I am frigid 23. That my partners will leave me 24. That people [would] think less of me because of my pain 25. That other people are better sexual partners than me 26. That I am a bad sexual partner 27. That I will not be able to find [a] future partner[s] 28. That my pelvic muscles will be too tight Interference with 29. Sitting 30. Walking 31. Wearing tight-fitting clothing 32. Taking part in recreational activities 33. Ability to go to work 34. Going out with friends 35. Fulfilling responsibilities to your family 36. Ability to perform tasks at work 37. Activities involving direct or indirect pressure to the vulva 38. Using sanitary pads 39. Ability to fall asleep Interference with 40. My response to sexual advances made by my partner 41. Desire for sexual activity 42. Feeling sexual pleasure 43. Orgasm frequency 44. Taking part in non-penetrative sexual activity 45. Taking part in penetrative sexual activity 46. Worrying about sexual satisfaction no longer being possible 47. Worrying that any sensation in my genitals will lead to pain
0.53 0.85 0.56 0.85 0.48 0.66 0.15
0.23
0.20
Factor 2
Factor 3
Factor 4
Factor 5
Factor 6
0.24 0.21 0.36 0.37 0.65 0.63 0.66 0.76 0.74 0.60 0.75 0.75 0.56 0.62 0.52 0.61 0.73 0.59
0.16 0.17 0.19 0.16 0.14 0.15
0.28 0.33 0.22
0.80 0.66 0.62 0.57 0.84 0.75 0.59
0.25
0.36
0.21 0.21
0.15
0.26 0.14
0.23
0.28 0.72 0.84 0.77 0.85 0.85 0.83 0.80 0.65 0.72 0.49 0.56
0.26 0.19
0.23
0.16
0.15
0.78 0.83 0.82 0.61 0.65 0.65 0.55 0.54
0.15 0.21
(continued)
J Sex Med 2017;14:1585e1596
1591
Vulvar Pain Assessment Questionnaire Table 3. Continued Factor 1 48. Taking off my clothes around my partner 49. Worrying about the next time my partner(s) will want sexual activity How often the following cause pain 50. Using tampons 51. Solitary sexual stimulation of my vulva (ie, masturbation) 52. Masturbation with partner present 53. Penetrating myself with fingers (partner absent) 54. Penetrating myself with a sex toy (partner absent)
Factor 2
Factor 3
0.23 0.33
Factor 4
Factor 5
0.27 0.30
0.35 0.60
0.28
0.19
0.29
Factor 6
0.33 0.47 0.45 0.76 0.80
*All non-significant factor loadings are not listed; all factor loadings listed are significant at P < .05.
Participants also were asked to provide feedback on the length, readability, and range of questions asked during the VPAQscreen administration. Most indicated that the VPAQ
was “good” or “excellent” when it came to length (n ¼ 159; 87.4%), readability (n ¼ 160; 87.9%), and the range of questions (n ¼ 154; 84.6%).
Table 4. Geomin-rotated factor loadings for the screening of the Vulvar Pain Assessment Questionnaire* Pain severity 1. Average pain intensity 2. Average unpleasantness 3. Average pain distress Thinking/Feeling 4. Bad about myself because of the pain 5. That my partners might think I am frigid 6. That my partners will leave me 7. That people [would] think less of me because of my pain 8. Sad 9. Unable to make changes in my life 10. That people might think I’m a bad sexual partner 11. Emotionally exhausted because of the pain 12. Anger toward my pain 13. That the pain will never stop Interference with 14. Sitting 15. Walking 16. Wearing tight-fitting clothing 17. Taking part in recreational activities 18. Ability to go to work 19. Ability to fall asleep Interference with 20. My response to sexual advances made by my partner 21. Desire for sexual activity 22. Feeling sexual pleasure 23. Orgasm frequency 24. Taking part in non-penetrative sexual activity 25. Taking part in penetrative sexual activity How often the following cause pain 26. Using tampons 27. Solitary sexual stimulation of my vulva (ie, masturbation) 28. Masturbation with partner present 29. Penetrating myself with fingers (partner absent) 30. Penetrating myself with a sex toy (partner absent)
Factor 1
Factor 2
0.76 0.92 0.70
0.27
0.18
0.22
0.70 0.63 0.65 0.73 0.49 0.51 0.65 0.56 0.50 0.48
0.38 0.18 0.28
0.19
0.82 0.87 0.80 0.88 0.70 0.53
0.26 0.37
0.12 0.15 0.21 0.17 0.36
*All non-significant factor loadings are not listed; all factor loadings listed are significant at P < .05. J Sex Med 2017;14:1585e1596
Factor 3
0.29 0.42 0.20 0.20
Factor 4
Factor 5
0.26
0.28
0.15
0.76 0.85 0.84 0.66 0.67 0.65
0.27
0.31 0.52 0.47 0.78 0.80
1592
Dargie et al
Table 5. Geomin-rotated factor loadings for pain descriptors of the Vulvar Pain Assessment Questionnaire* Factor 1 1. Burning 2. Stinging 3. Sharp 4. Stabbing 5. Aching 6. Irritating 7. Raw 8. Sensitive 9. Tender 10. Sore
0.69 0.49
Factor 2
Factor 3
0.40 0.95 0.76 0.17
0.31 0.46 0.55 0.82 0.92 0.76
0.26 0.22 0.19
*All non-significant factor loadings are not listed; all factor loadings listed are significant at P < .05.
Participants were given an opportunity to provide open-ended feedback on the VPAQ scales. Common suggestions included requests for the ability to provide more detailed responses (eg, more precise pain locations on the vulva diagram, where pain is felt during different sexual activities, additional pain descriptors, whether the pain affects caring for children), to convey how the pain has changed or affected them over time rather than focusing on the past 4 weeks, and whether they were currently engaged in other treatment for their vulvar pain (eg, medication, physical therapy). Other participants noted frustration with the process of finding medical professionals to treat their pain and expressed a need to capture those experiences alongside reports of their vulvar pain symptoms. Several also expressed gratitude for the creation of the VPAQ and the fact that researchers were studying this poorly understood condition.
DISCUSSION The purpose of this study was to evaluate the psychometric characteristics of each part of the VPAQ, to provide preliminary Table 6. Geomin-rotated factor loadings for the coping mechanisms scale of the Vulvar Pain Assessment Questionnaire* Factor 1 1. Relax my body 2. Breathe deeply 3. Go to my “happy place” 4. Practice yoga or stretching 5. Do something that takes my mind off the pain 6. Focus on staying optimistic 7. Visit my doctor(s) 8. Look for information on my pain 9. Use prescription medication 10. Talk to people in my social network 11. Talk to others with similar pain 12. Avoid anything that might cause pain
0.84 0.84 0.28 0.40
0.24
Factor 2
0.45 0.22 0.58 0.44 0.64 0.59 0.50 0.58 0.57 0.42
*All non-significant factor loadings are not listed; all factor loadings listed are significant at P < .05.
normative data for the VPAQ subscales, and to gather feedback from participants on its helpfulness and accessibility. The psychometric properties examined in this study provide further justification for use of the VPAQ in research studies, and the feedback gathered from participants shed light on the usefulness and acceptability of the scales and provide guidance on how it could be expanded in clinical settings. The results of the exploratory structural equation modeling23 confirmed that all subscales—other than the VPAQcope (see below)—had adequate to good factor structure.17e19 Although most scales met the standards for acceptable model fit, most did not exceed the more rigorous standards suggested by some.17e19 Furthermore, most items in the VPAQ subscales significantly loaded onto the same factors proposed in the original study15 in the hypothesized direction. Several significantly cross-loading items likely contributed to difficulties with model fit. However, the VPAQcope was the only scale that was consistently problematic when examining model fit and factor loadings. Our analyses did not support the 2-factor solution from the original study on scale construction, although this analysis was underpowered for the present sample.15 Future research with larger samples could determine whether these patterns are common across all samples and whether these items should be moved, modified, or deleted. Perhaps the factor structure of the VPAQ would be stronger within specific populations of people experiencing vulvar pain. For example, the questionnaire was developed with those suffering from vulvodynia in mind (ie, idiopathic chronic vulvar pain),1 but participants were not excluded from this online study based on other diagnoses (ie, they could have had another disorder associated with the experience of vulvar pain, such as lichen sclerosus, interstitial cystitis or painful bladder syndrome, or chronic yeast infections). Future studies should recruit large numbers of participants in numerous diagnostic groups to determine the validity of the questionnaire across specific diagnoses. However, the combination of reasonable fit indices and statistically significant factor loading parameters provides preliminary evidence of a psychometrically sound factor structure for all but the VPAQcope. Measures of internal consistency were strong for most of the VPAQ, similar to the original study on scale construction.15 Indeed, only the burning and stinging subscale of the VPAQdesc had scores that did not have adequate, good, or excellent internal consistency. This finding indicates that experiencing burning pain might not always correspond with the experience of stinging pain. As such, perhaps these 2 items should be considered separately instead of creating an average score of the 2 adjectives. Indeed, scores on these items were only moderately correlated compared with a strong correlation between the 2 items that compose the incisive subscale. The low internal consistency also could be accounted for by the small number of items composing this subscale.24 Moreover, it should be noted that our measure of internal consistency, Cronbach a, provided a J Sex Med 2017;14:1585e1596
1593
Vulvar Pain Assessment Questionnaire
Table 7. Geomin-rotated factor loadings for the impact of vulvar pain on romantic relationships in the Vulvar Pain Assessment Questionnaire* Factor 1 Supportive response 1. Asks what s/he can do 2. Wants to talk about it 3. Tries to acknowledge my pain 4. Seek emotional support 5. Seek physical comfort 6. Share your feelings 7. Problem solve Negative partner response 8. Gets angry 9. Blames me 10. Appears frustrated 11. Is visibly upset 12. Looks sad How has your pain affected 13. Physical intimacy 14. Emotional intimacy 15. Sexual intimacy 16. Relationship quality 17. General communication 18. Sexual communication Comfort communicating about 19. Sexual desire 20. Frequency of activity 21. Amount of “foreplay” 22. Duration of activity 23. Sexual position 24. Technique
0.55 0.52 0.52 0.87 0.84 0.85 0.56
Factor 2
Factor 3
Factor 4
0.26
0.74 0.73 0.76 0.83 0.52
0.17 0.26
0.2
0.46 0.82 0.45 0.76 0.82 0.71 0.84 0.78 0.92 0.91 0.84 0.89
*All non-significant factor loadings are not listed; all factor loadings listed are significant at P < .05.
lower-bound estimate of reliability and might underestimate the true reliability of a scale. Similar patterns were observed when examining test-retest reliability: there were strong relationships among scores at time 1 and 4 weeks later at time 2 for most subscales, although reliability was only moderate for 4 subscales. Once again, the burning and stinging subscale was in the moderate range, as was the incisive pain subscale of the VPAQdesc. Similar patterns were observed for relationship impact and impact on communications within romantic relationships, for which moderate test-retest reliability was observed. Perhaps the characteristics captured by these subscales fluctuate more frequently than other, more stable, symptoms captured by this questionnaire, such as how severe the pain is and how women with vulvar pain interact with their romantic partner when experiencing pain. It would be useful to track these variables over time to understand which factors vary, which remain stable, and how the experience of the pain is affected. Such investigations also would provide guidance on whether the subscales would need to be modified to improve reliability.
J Sex Med 2017;14:1585e1596
There were no difficulties identified with the internal consistency or test-retest reliability of the VPAQfull or the VPAQscreen, the 2 clusters hypothesized to be most important for the assessment and diagnosis of chronic vulvar pain. Furthermore, no major concerns were raised based on the model fit indices for these clusters, although it would be appropriate to revisit these analyses with different populations to determine whether the questionnaire requires revising to attain strong model fit. Thus, although the psychometric properties of the supplementary scales are not as consistently strong, the results of the present study support the use of the VPAQscreen and the VPAQfull to characterize vulvar pain symptoms. When examining the normative data for the present sample, average ratings on each scale generally indicated mild to moderate difficulties in the domains captured by the VPAQ. We suggest that a rating of at least 2 on a given subscale indicates clinical significance. Scores below 2 indicate that the person experiences mild to no pain severity, lower intensity cognitive and emotional responses, minimal interference in various life domains, and so on. Conversely, scores of at least 2 likely
1594
Dargie et al
Table 8. Internal consistency, normative data, and test-retest reliability for the VPAQ subscales VPAQ subscales* Full version Severity Emotional response Cognitive response Life interference Sexual function interference Self-penetration Screening Severity Cognitive or emotional Life interference Sexual function interference Self-penetration Pain descriptors Burning and stinging Incisive Sensitivity Coping strategies Distraction or relaxation Problem solving Partner factors Support seeking Negative response Relationship impact Communication
Time 1, Cronbach a†
Time 1, mean (SD)†
4-wk test-retest, reliability (r)‡
0.91 0.94 0.89 0.93 0.91 0.75
2.38 2.09 1.68 1.28 2.45 1.87
(0.80) (1.00) (1.13) (1.12) (1.12) (1.19)
0.74§ 0.77§ 0.78§ 0.96§ 0.87§ 0.73§
0.87 0.89 0.91 0.89 0.75
1.95 2.00 1.36 2.66 1.87
(0.85) (0.99) (1.20) (1.20) (1.19)
0.75§ 0.77§ 0.93§ 0.83§ 0.73§
0.56 0.84 0.81
2.47 (1.17) 1.91 (1.37) 2.58 (0.94)
0.66§ 0.57§ 0.78§
0.75 0.71
1.74 (0.82) 1.63 (0.76)
0.80§ 0.83§
0.88 0.81 0.85 0.95
2.11 0.87 1.50 2.01
(0.90) (0.79) (0.65) (1.15)
0.71§ 0.80§ 0.49§ 0.50§
VPAQ ¼ Vulvar Pain Assessment Questionnaire. *Mean subscale scores are based on means across items and can range from 0 to 4. † For partner subscales, n ¼ 143; for all other subscales, n ¼ 182 based on time 1 data. ‡ For partner subscales, n ¼ 50; for all other subscales, n ¼ 66 based on time 2 data. § P < .001.
represent symptoms of greater significance. However, further research is required to test and refine this hypothesis. Furthermore, there might be certain pain symptom “profiles” that could be determined for various vulvar pain subtypes (eg, provoked vestibulodynia vs generalized vulvodynia), which might further aid in the diagnostic process. However, the preliminary norms generated by this study should be interpreted with caution because of the convenience sample (ie, participants were not randomly selected from the general population) and because the diagnoses of participants were heterogeneous and could not be confirmed. Participants provided positive feedback related to their involvement in this study. Indeed, approximately half the participants indicated that simply answering a range of questions that captured their pain experiences improved their comfort levels in speaking with medical professionals about various topics relevant to their experience of pain. In addition, most participants in the present study provided positive feedback about the length of the questionnaire, the readability of the questions, and the range of characteristics captured by the items. Moreover, some participants commented that they would have liked the
opportunity to share additional details about their pain experiences, describe the pain and its impact over time, and convey their experiences seeking help and treatment methods they had attempted. Overall, these findings suggest that the VPAQ might aid in communication about various symptoms between patients and providers, and patients or participants should be given the opportunity to provide additional pertinent information or tailor the questionnaires based on the information they want to share with their providers. This study has several limitations. Although the sample size was adequate for many of our analyses, some were underpowered. The sample size also limited our ability to compare various vulvar pain diagnostic groups to determine how their different characteristics might have affected our results, particularly model fit. Furthermore, all data were collected with an online survey representing a non-random sample of convenience, and we could not confirm diagnoses. Although a strong link between selfreport and clinical diagnosis has been established,25e27 it would strengthen the findings of this study if diagnoses were validated through a gynecologic examination. Furthermore, 1 item was missing from the emotional response subscale of the J Sex Med 2017;14:1585e1596
1595
Vulvar Pain Assessment Questionnaire
VPAQfull, which could affect conclusions drawn about this subscale based on this study. Indeed, the results relating to the VPAQfull should be interpreted with caution until further studies can be conducted with all items included. We also did not include any attention checks in our online survey, which limited our ability to verify the validity of survey responses. Further, more than half the sample did not complete the survey at time 2. Although further analyses indicated no differences between those who did and those who did not complete the time 2 survey, it would be ideal to replicate these results in a sample with a less pronounced attrition rate.
CONCLUSION The results of the present study provide preliminary norms for and promising evidence of the psychometric strength of the VPAQ inventory and its usefulness from the perspective of those with chronic vulvar pain. Although model fit was not strong for some VPAQ scales, adequate model fit paired with largely strong internal consistency and test-retest reliability and strong positive feedback from participants show the overall strength of the measures. Although the VPAQcope requires revisiting in future studies, the results of this study indicate that the other VPAQ scales are ready for use in clinical and research settings. Ideally, the VPAQ will enable health care providers to more efficiently and accurately diagnose women with vulvar pain, and this study will provide further justification for its use in clinical and research settings.
REFERENCES 1. Bornstein J, Goldstein AT, Stockdale CK, et al. 2015 ISSVD, ISSWSH, and IPPS consensus terminology and classification of persistent vulvar pain and vulvodynia. J Low Genit Tract Dis 2016;20:126-130. 2. Harlow BL, Stewart E. A population-based assessment of chronic unexplained vulvar pain: have we underestimated the prevalence of vulvodynia? J Am Med Womens Assoc 2003; 58:82-88. 3. Harlow BL, Kunitz CG, Nguyen RHN, et al. Prevalence of symptoms consistent with a diagnosis of vulvodynia: population-based estimates from 2 geographic regions. Am J Obstet Gynecol 2014;210:40.e1-40.e8. 4. Arnold LD, Bachmann GA, Rosen R, et al. Vulvodynia: characteristics and associations with comorbidities and quality of life. Obstet Gynecol 2006;107:617-624. 5. Baggish MS. Diagnosis and management of vulvar vestibulitis syndrome in 559 women (1991e2011). J Gynecol Surg 2012; 28:75-82. 6. Dargie E, Pukall CF. Women in “sexual” pain: exploring the manifestations of vulvodynia. J Sex Marital Ther 2016; 42:309-323. 7. Glajchen M. Chronic pain: treatment barriers and strategies for clinical practice. J Am Board Fam Pract Board Fam Pract 2001;14:211-218. 8. Green CR, Wheeler JRC, Laporte F, et al. How well is chronic pain managed? Who does it well? Pain Med 2002;3:56-65. 9. Nicolaou A. Conference report: The Change Pain Physician Survey. Curr Med Res Opin 2011;27:2073-2074.
Corresponding Author: Caroline F. Pukall, PhD, CPsych, Department of Psychology, Queen’s University, Kingston, ON K7L 3N6, Canada; E-mail: [email protected]
10. Pancholy AB, Goldenhar L, Fellner AN, et al. Resident education and training in female sexuality: results of a national survey. J Sex Med 2011;8:361-366.
Conflicts of Interest: The authors report no conflicts of interest.
11. Shindel AW, Ando KA, Nelson CJ, et al. Medical student sexuality: how sexual experience and sexuality training impact U.S. and Canadian medical students’ comfort in dealing with patients’ sexuality in clinical practice. Acad Med 2010; 85:1321-1330.
Funding: None.
STATEMENT OF AUTHORSHIP Category 1 (a) Conception and Design Emma Dargie; Caroline F. Pukall (b) Acquisition of Data Emma Dargie; Caroline F. Pukall (c) Analysis and Interpretation of Data Emma Dargie; Ronald R. Holden; Caroline F. Pukall Category 2 (a) Drafting the Article Emma Dargie; Ronald R. Holden; Caroline F. Pukall (b) Revising It for Intellectual Content Emma Dargie; Ronald R. Holden; Caroline F. Pukall Category 3 (a) Final Approval of the Completed Article Emma Dargie; Ronald R. Holden; Caroline F. Pukall J Sex Med 2017;14:1585e1596
12. Solursh DS, Ernst JL, Lewis RW, et al. The human sexuality education of physicians in North American medical schools. Int J Impot Res 2003;15(Suppl 5):S41-S45. 13. Upshur CC, Luckmann RS, Savageau JA. Primary care provider concerns about management of chronic pain in community clinic populations. J Gen Intern Med 2006; 21:652-655. 14. Reed BD, Harlow SD, Sen A, et al. Prevalence and demographic characteristics of vulvodynia in a population-based sample. Am J Obstet Gynecol 2012;206:170.e1-170.e9. 15. Dargie E, Holden RR, Pukall CF. The Vulvar Pain Assessment Questionnaire Inventory. Pain 2016;157:2672-2686. 16. Simms LJ, Watson D. The construct validation approach to personality scale construction. In: Robins RW, Frayley RC, Knager RF, eds. Handbook of research methods in personality psychology. New York: Guilford; 2007. p. 240-258.
1596 17. Schumacker RE, Lomax RG. A beginner’s guide to structural equation modeling. 3rd ed. New York: Taylor & Francis Group; 2010. 18. Hu L, Bentler PM. Fit indices in covariance structure modeling: sensitivity to underparameterized model misspecification. Psychol Methods 1998;3:424-453. 19. Browne MW, Cudek R. Alternative ways of assessing model fit. Sociol Methods Res 1992;21:230-258. 20. Maccallum RC, Browne MW, Sugawara HM. Power analysis and determination of sample size for covariance structure modeling of fit involving a particular measure of model. Psychol Methods 1996;13:130-149. 21. Cohen J. A power primer. Psychol Bull 1992;112:155-159. 22. Vazsonyi AT, Ksinan A, Mikuska J, et al. The Big Five and adolescent adjustment: an empirical test across six cultures. Pers Individ Dif 2015;83:234-244.
Dargie et al 23. Asparouhov T, Muthén BO. Exploratory structural equation modeling, Volume 16. York: Taylor & Francis Group; 2009. 24. Nunnally J, Bernstein IH. Psychometric theory. 3rd ed. New York: McGraw-Hill; 1994. 25. Harlow BL, Vazquez G, MacLehose RF, et al. Self-reported vulvar pain characteristics and their association with clinically confirmed vestibulodynia. J Womens Health 2009; 18:1333-1339. 26. Reed BD, Haefner HK, Harlow SD, et al. Reliability and validity of self-reported symptoms for predicting vulvodynia. Obstet Gynecol 2006;108:906-913. 27. Dargie E, Chamberlain SM, Pukall CF. Provoked vestibulodynia: diagnosis, self-reported pain, and presentation during gynaecological examinations. J Obstet Gynaecol Can 2017; 39:145-151.
J Sex Med 2017;14:1585e1596
PAIN
The Vulvar Pain Assessment Questionnaire: Factor Structure, Preliminary Norms, Internal Consistency, and Test-Retest Reliability Emma Dargie, PhD, Ronald R. Holden, PhD, CPsych, and Caroline F. Pukall, PhD, CPsych
ABSTRACT
Background: The Vulvar Pain Assessment Questionnaire (VPAQ) was developed to assist in the assessment and diagnosis of chronic vulvar pain (vulvodynia). Aim: To further establish the psychometric properties of the VPAQ by examining factor structure, test-retest reliability, internal consistency, and scale normative data, and to gather feedback from those with vulvar pain about the usefulness and accessibility of the questionnaire. Methods: 182 participants completed a confidential online study and 70 participated again at time 2 (4 weeks later). Outcomes: Participants were asked to complete the full VPAQ, which assesses pain characteristics, effects on various parts of their lives, coping strategies used, and romantic partner factors. Additional questions captured sociodemographics and feedback about the instrument. Results: Exploratory structural equation modeling indicated that the previously established subscales, except the coping scale, had adequate model fit, and all items loaded significantly onto relevant factors. Pearson product moment correlations (r ¼ 0.57e0.96) established strong 4-week test-retest reliability for most subscale scores, and Cronbach a indicated overall acceptable to high internal consistency (a ¼ 0.56e0.95). Preliminary norms for the scales are supplied. Approximately half the participants reported an increase in their comfort level in discussing a range of topics after completing the VPAQ. Most participants reported that the length, readability, and range of VPAQ questions were “good” or “excellent.” Clinical Implications: The results of this study provide further justification for using the VPAQ scales in clinical and research settings, preliminary norms for a vulvar pain population, and suggestions for interpretation. Strengths and Limitations: This study established the psychometric properties of the VPAQ scales using multiple methods at 2 time points and gathered feedback from participants. However, data were collected online so diagnoses could not be confirmed and more than half the initial sample did not complete the survey at time 2. Conclusion: The results of this study suggest that most VPAQ subscales (except the coping subscale) have moderate to strong psychometric properties and that the VPAQ is user friendly. Dargie E, Holden RR, Pukall CF. The Vulvar Pain Assessment Questionnaire: Factor Structure, Preliminary Norms, Internal Consistency, and Test-Retest Reliability. J Sex Med 2017;14:1585e1596. Copyright 2017, International Society for Sexual Medicine. Published by Elsevier Inc. All rights reserved.
Key Words: Vulvodynia; Dyspareunia; Pain Assessment; Factor Analysis; Reliability
INTRODUCTION Millions suffer from vulvodynia (persistent vulvar pain lacking a clear identifiable cause), with prevalence estimates ranging from 8% to 15%.1e3 However, many have difficulty obtaining an accurate diagnosis and effective treatment and need to seek help Received May 22, 2017. Accepted October 31, 2017. Department of Psychology, Queen’s University, Kingston, ON, Canada Copyright ª 2017, International Society for Sexual Medicine. Published by Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jsxm.2017.10.072
J Sex Med 2017;14:1585e1596
from more than 3 providers before securing diagnoses and/or finding some relief.2,4e6 Although multiple factors play a role in this process, 1 primary concern is assessment and diagnosis by health care providers. Unfortunately, limited time is devoted to chronic pain or sexual health during medical training, perhaps leading to discomfort when working with patients with pain during sexual activity.7e13 It is not surprising that many women with vulvodynia consult at least 3 health care providers before obtaining an accurate diagnosis or pain relief.2,4e6,14 A possible issue contributing to this process is the lack of a comprehensive tool to assess and diagnose vulvodynia symptoms. 1585
1586
Thus, our group created the Vulvar Pain Assessment Questionnaire (VPAQ).15 We administered a large pool of questions to women with chronic vulvar pain and used the construct validation approach16 to select salient, relevant items for VPAQ subscales through a series of factor analyses. Although the initial motivation behind the creation of this questionnaire was to assess those with idiopathic chronic vulvar pain (ie, vulvodynia), the questionnaire domains are applicable to chronic vulvar pain resulting from a range of causes (eg, lichen sclerosus, interstitial cystitis or painful bladder syndrome, or chronic yeast infections). Indeed, the primary questionnaire assesses pain characteristics, pain severity, cognitive and emotional factors, and interference with life, sexual function, and self-stimulation and penetration. A brief version also was created to assist in expedited pain assessment. Further scales were created to capture pain characteristics, coping strategies, and romantic partner factors. Clinicians and researchers could select scales to administer based on the needs of the vulvar pain population being targeted. The VPAQ can be used to gather information on symptoms and diagnosis and can be used to identify possible treatment targets.15 When correlated with other established questionnaires, evidence of convergent and discriminant validity was observed. For example, the sexual functioning subscale of the VPAQ was strongly related to scores on the Female Sexual Function Index but only mildly or moderately related to scores on other instruments.15 For each subscale, average scores can be computed to determine which areas are of greatest concern, and we hypothesize that scores can be tracked over time to help measure treatment progress. The results of this scale construction study were quite promising, although further research is required to replicate past results, examine psychometric properties in more detail, and gather feedback on the newly constructed VPAQ scales from those with chronic vulvar pain. The goal of the present study was to further test the VPAQ by (i) confirming its factor structure, (ii) investigating test-retest reliability, (iii) replicating internal consistency findings, (iv) providing normative data for all scales, and (v) gathering feedback from participants on the usefulness and accessibility of the inventory.
METHODS This online study was approved by the university’s general research ethics board, and participants provided informed consent. Participants were recruited through word of mouth, online advertisements, and postings to relevant listservs and groups. Similar to the original study on scale construction,15 anyone older than 18 years, with access to the Internet, and who reported experiencing chronic vulvar pain was invited to participate. Once participants reached the secure survey website, they read a letter of information; if they consented to participate, then they completed the survey consisting of sociodemographic characteristics, the VPAQ, and feedback about the VPAQ items.
Dargie et al
At the end of this survey (time 1), participants were invited to sign up for time 2 by providing their e-mail address in a separate survey, in addition to a unique identifier that would enable us to link their data at the 2 time points without knowing their identity. Approximately 4 weeks after initial completion, participants were invited to complete the VPAQ a second time (time 2). Each survey took approximately 15 to 30 minutes. To thank them for their time, participants were invited to enter a draw for 1 of 4 prizes valued at 50 CAD that took place at the end of the study. Before analysis, each participant’s responses were examined for eligibility, random responding, and excessive missing data. 182 participants provided complete data for time 1, and 70 participants provided complete data for time 2 (Figure 1 presents information about those who did not complete the study).
Main Outcome Measures The VPAQ15 is composed of a comprehensive questionnaire that covers a wide range of symptoms related to chronic vulvar pain, a brief version of this broad measure, a scale devoted to pain descriptors, a scale for coping strategies, and a scale for romantic partner factors. Information about the development of the following scales can be found in the article on scale construction.15 In addition to the VPAQ scales, participants were asked to report how answering the questions changed their comfort in talking about various topics with health care professionals (eg, asking for help about their vulvar pain, talking about what the pain feels like, and sharing how being in pain affected their emotional health). Change in comfort in discussing each topic was rated on a scale from 0 (no change in comfort) to 4 (a lot more comfort). Participants also were asked to rate the VPAQ questions for length, readability, and range of questions. These characteristics were rated on a scale from 1 (poor) to 4 (excellent). Full Version (VPAQfull) The full version of the questionnaire consists of 55 items rated on 5-point scales with anchors tailored to the type of question and 8 categorical questions that assess the onset, location, temporal pattern, degree of burning pain, and associated symptoms (eg, itching) of vulvar pain. The domains encompassed by the questionnaire include cognitive and emotional reactions to the pain, pain intensity, pain unpleasantness and discomfort, life interference, sexual function interference, and self-penetration interference. In the original study on scale construction, internal consistency was good for the VPAQfull subscale scores (a > 0.77). Due to a programming error in the present study, 1 item of the emotional response subscale was not included (ie, the item assessing how much the participant felt “like giving up”). Thus, analyses were computed using a 14-item version of the emotional response scale, rather than the original 15-item scale. Screening (VPAQscreen) This brief version of the questionnaire consists of 30 items rated on 5-point scales with anchors tailored to the type of J Sex Med 2017;14:1585e1596
1587
Vulvar Pain Assessment Questionnaire Viewed study Letter of Infor mation (n = 1146) Did not consent to par ticipate (n = 859) Consented to par ticipate (n = 287)
Repeat r esponses (n = 3) Incomplete Data (n = 91) Pain-fr ee (n = 8) Male (n = 2) Under 18 year s (n = 1)
Completed Time 1 (n = 182) Did not complete Time 2 (n = 112) Completed Time 2 (n = 70)
Figure 1. Flowchart illustrating the number of participants who completed each stage of the study, and reasons for exclusion.
questions and the same 8 categorical questions in the VPAQfull that assess the temporal nature and quality of the pain. The domains addressed by the VPAQscreen are the same as those addressed by the VPAQfull, although the cognitive and emotional aspects of vulvar pain are collapsed into 1 factor. In the original study on scale construction, the internal consistency for VPAQscreen subscale scores was good (a > 0.77), similar to the internal consistency when the screening tool was implemented in a clinical context (a > 0.80; manuscript in preparation). Pain Descriptors (VPAQdesc) The VPAQdesc captures common words used to describe chronic vulvar pain. The 10 items, rated on a scale from 0 (not at all) to 4 (very much), allow participants to indicate the degree to which each descriptor applies to their pain. 3 subscales compose the VPAQdesc: burning and stinging pain, incisive pain, and sensitivity. In the original study on scale construction, the internal consistency of the VPAQdesc subscale scores was adequate to good (a > 0.63).
J Sex Med 2017;14:1585e1596
Coping Strategies (VPAQcope) This scale addresses coping strategies that are commonly used by those with vulvar pain. It consists of 12 items rated on a scale from 0 (never) to 4 (always), allowing participants to indicate the frequency with which they use such strategies. Items are grouped into 2 categories: distraction- and relaxation-based strategies and active problem-solving strategies. In the original study on scale construction, the internal consistency for the VPAQcope subscale scores was adequate to good (a > 0.69). Partner Factors (VPAQpartner) This 24-item scale assesses how romantic partners and spouses might be affected by or respond to the vulvar pain experienced by their partner. Each question is rated on a 5-point scale with anchors based on question type. 4 subscales can be calculated: how the partner or spouse responds to the pain, support-seeking behaviors, the impact of the pain on aspects of the relationship, and communication about sexual matters when experiencing pain. In the original study on scale construction, the internal consistency for the subscale scores was good (a > 0.77).
1588
Statistical Analysis Analyses were conducted using SPSS 22.0 (IBM Corp, Armonk, NY, USA) and MPlus 7.4 (http://www.statmodel.com/). After examining the data for appropriate assumptions, exploratory structural equation modeling was conducted to examine the quality of the factor structure for each part of the VPAQ. To evaluate each subscale, 3 techniques were used.17 (i) Model fit indices were examined to determine how well the proposed factor structure fit the data collected during the current study. (ii) Significance of the factor loadings was examined to ensure that each item loaded significantly onto the appropriate factor (t > 1.96, P < .05). (iii) Each factor loading parameter was examined for interpretability. To evaluate model fit, we examined patterns across the root mean square error of approximation, the standardized root mean square residual, the comparative fit index, and the Tucker-Lewis index. Several sources were consulted to determine appropriate criteria for evaluating these tests of model fit,17e19 with the conclusion that acceptable fit could be established with a root mean square error of approximation and a standardized root mean square residual less than 0.10 (good fit < 0.05) and a comparative fit index or Tucker-Lewis index greater than 0.90 (good fit > 0.95). Test-retest reliability was computed for each subscale by examining the correlation between corresponding VPAQ scores at the 2 time points (ie, Pearson r was computed). Preliminary norms were established from this sample based on mean subscale scores. This approach was used instead of creating percentile ranks because the entire sample consisted of participants with vulvar pain and did not include those without vulvar pain. Typically, percentile ranks are used to indicate when the presence of symptoms is atypical for the general population, rather than characterizing variations in those who report such symptoms. Furthermore, use of mean scores is more accessible and easy to interpret for those less familiar with the use of percentiles. In addition, descriptive statistics were examined for feedback participants gave about its content. According to published guidelines,20 the sample size in the present study was adequate to obtain 0.80 power for the VPAQfull, VPAQscreen, and VPAQpartner model fit analyses. However, the VPAQcope and VPAQdesc model fit analyses were underpowered. For test-retest reliability, the size of the present sample had adequate power (0.80) to detect large effects when a was set at 0.01.21
RESULTS The age of participants who completed time 1 ranged from 18 to 85 years, and the average age was 36.13 years (SD ¼ 15.89; median ¼ 29.00). The age of participants who completed time 2 ranged from 20 to 78 years, and the average age was 36.59 years (SD ¼ 15.42; median ¼ 30.50). Other sociodemographics and self-reported diagnoses are presented in Table 1. No differences were found when comparing those who completed time points 1
Dargie et al
and 2 with those who completed only time 1 on these sociodemographic characteristics (P > .23 by Fisher exact test for all comparisons).
Exploratory Structural Equation Modeling Although a confirmatory factor analysis (CFA) approach is common for scale development and validation, some have argued that a traditional CFA can be overly conservative when examining scales that have factors related to one another.22 Indeed, traditional CFA requires that items be permitted to load onto only 1 factor, and cross-loadings are restricted to 0. The results we reported in the original study on scale construction indicated weak to moderate correlations among some factors,15 suggesting that a CFA approach would not be appropriate for further examining model fit in the present study. Thus, exploratory structural equation modeling procedures were examined using MPlus,23 which allows for indices of model fit to be computed with non-0 cross-loadings. Based on this approach, most VPAQ subscales had good or acceptable model fit (Table 2). In addition, most items loaded significantly onto the proposed factors in the hypothesized direction (Tables 3e7). For the VPAQfull, the most problematic item captured worries that the participant’s pelvic muscles would be too tight. This item did not load significantly onto the cognitive responses to pain factor, as hypothesized. Rather, it loaded significantly onto the emotional responses to pain factor. There also were 6 items with significant cross-loadings onto other factors (>0.30), although the largest factor loading was on the hypothesized factor. For the VPAQscreen, all items loaded significantly onto the hypothesized factor. However, 4 items had significant cross-loadings onto other factors (>0.30), although only 1 (painful tampon use) loaded more strongly onto another factor (cognitive and emotional response) than the hypothesized factor (self-penetration interference). For the VPAQdesc, there was only 1 item with a significant cross-loading (>0.30); the item capturing stinging pain loaded on the burning and stinging pain subscale in addition to the incisive pain subscale, although the loading was stronger on the former. The VPAQcope was the most problematic of the VPAQ scales. 3 items originally part of the distraction or relaxation subscale (going to one’s “happy place,” doing something to take one’s mind off the pain, and focusing on staying optimistic) loaded significantly onto the second factor, the active problemsolving subscale. Further, there were no significant crossloadings (>0.30) for the VPAQpartner, and all items loaded onto the hypothesized factors.
Internal Consistency, Norms, and Test-Retest Reliability Internal consistency for the time 1 data was assessed using Cronbach a. Only the burning and stinging subscale of the VPAQdesc had scores for which reliability was not adequate, good, or excellent (Table 3). Of note, scores on the burning and J Sex Med 2017;14:1585e1596
1589
Vulvar Pain Assessment Questionnaire
Table 1. Sociodemographic characteristics and self-reported diagnoses at time 1 and time 2 (w4 weeks later)* Time 1, n (%) Sexual orientation Heterosexual Gay or lesbian Bisexual Queer, asexual, other Gender identity Woman Transwoman, androgynous, other Decline response Relationship status Married or common law Dating Not dating Other or declined response Children 0 1 2 3e5 Declined response Employment Full time Part time Retired Student Disability, unemployed, medical leave Other (eg, housewife, home parent) Education [Some] high school [Some] trade school [Some] college or undergraduate [Some] graduate or professional Diagnosis Vulvar or genital pain but no diagnosis Vulvodynia Vulvar vestibulitis syndrome Provoked vestibulodynia Generalized vulvodynia Vaginismus Genito-pelvic pain or penetration disorder Pelvic floor muscle dysfunction Interstitial cystitis Chronic pelvic pain Irritable bowel syndrome
152 2 22 6
(83.5) (1.1) (12.1) (3.3)
177 (97.3) 4 (2.1) 1 (0.5)
Time 1, n (%)
Time 2, n (%) 58 1 8 3
(82.9) (1.4) (11.4) (4.3)
67 (95.7) 3 (4.2) 0 (0.0)
74 75 30 3
(40.7) (41.2) (16.5) (1.6)
27 25 17 1
(38.6) (35.6) (24.3) (1.4)
137 12 17 15 1
(75.3) (6.6) (9.3) (8.2) (0.5)
58 8 5 4 0
(75.7) (11.4) (7.1) (5.7) (0.0)
81 27 15 28 20
(44.5) (14.8) (8.2) (15.4) (11.0)
32 10 4 10 9
(45.7) (14.3) (5.7) (14.3) (12.9)
11 (6.0)
5 (7.1)
12 (6.5) 4 (2.1) 91 (50.0)
4 (5.7) 1 (1.4) 32 (45.7)
75 (41.2)
33 (47.1)
22 (12.1)
11 (15.7)
96 47 34 15 29 3
(52.7) (25.8) (18.7) (8.2) (15.9) (1.6)
36 24 15 8 13 2
(51.4) (34.3) (21.4) (11.4) (18.6) (2.9)
33 10 19 23
(18.1) (5.5) (9.9) (12.6)
15 4 11 11
(21.4) (5.7) (15.7) (15.7)
(continued)
J Sex Med 2017;14:1585e1596
Table 1. Continued
Endometriosis Lichen sclerosus Pelvic inflammatory disease Pudendal nerve entrapment Other (eg, vulvar dermatitis, postherpetic pain, polycystic ovarian syndrome)
8 7 2 10 13
Time 2, n (%)
(4.4) (3.8) (1.1) (5.5) (7.1)
1 4 0 1 6
(1.4) (5.7) (0.0) (1.4) (8.6)
Time 1 ¼ 1st completion of the Vulvar Pain Assessment Questionnaire; Time 2 ¼ 2nd completion of the Vulvar Pain Assessment Questionnaire. *Participants were invited to select as many diagnoses as applied to them.
stinging items were only moderately correlated (r ¼ 0.39) compared with a strong correlation between the 2 items that compose the incisive subscale (r ¼ 0.72). Means and SDs were computed for each time 1 VPAQ subscale (possible range ¼ 0e4), and independent-samples t-tests were used to determine whether those who also completed time 2 differed significantly from those who did not; there were no significant differences (P > .05 for all comparisons). Next, Pearson product-moment correlations were computed between subscale scores at times 1 and 2. Most of these relations were strong, except for the burning and stinging and incisive subscales of the VPAQdesc and the relationship impact and communication subscales of the VPAQpartner, which were moderate (Table 8).
Participant Feedback After completing the VPAQ during time 1, participants were asked whether doing so had changed their level of comfort in speaking to medical professionals about various topics. Many participants reported an increase in comfort when it came to asking for help about their vulvar pain (n ¼ 72; 40.0%), talking about what their pain feels like (n ¼ 89; 48.9%), being able to explain their overall experience of pain (n ¼ 91; 50.0%), and sharing how being in pain affects their emotional health (n ¼ 90; 49.5%), physical health (n ¼ 77; 42.3%), and sexual life (n ¼ 95; 52.2%). Table 2. Model fit indices for the VPAQ derived from exploratory structural equation modeling procedures using a geomin rotation VPAQ
df
RMSEA
CFI
TLI
SRMR
Full version Screening Pain descriptors Coping strategies Partner factors
1122 295 18 43 186
0.07 0.07 0.07 0.13 0.11
0.85 0.92 0.97 0.78 0.85
0.81 0.88 0.93 0.67 0.78
0.04 0.04 0.03 0.08 0.05
CFI ¼ comparative fit index; df ¼ degrees of freedom; RMSEA ¼ root mean square error of approximation; SRMR ¼ standardized root mean square residual; TLI ¼ Tucker-Lewis index; VPAQ ¼ Vulvar Pain Assessment Questionnaire.
1590
Dargie et al
Table 3. Geomin-rotated factor loadings for the full version of the Vulvar Pain Assessment Questionnaire* Factor 1 Pain severity 1. Average pain intensity 2. Worst pain intensity 3. Average unpleasantness 4. Worst unpleasantness 5. Average distress 6. Worst distress Feeling 7. Sad 8. Unable to make changes in my life 9. Bad about myself because of the pain 10. Emotionally exhausted because of the pain 11. Anger toward my pain 12. That the pain will never stop 13. Depressed 14. Like my body has let me down 15. Physically tense 16. That I am not a worthwhile person 17. Distracted 18. Hateful things about myself as a person 19. Stressed about the pain 20. That it is unfair that I have pain Thinking and worrying 21. That people might think I’m a bad sexual partner 22. That my partners might think I am frigid 23. That my partners will leave me 24. That people [would] think less of me because of my pain 25. That other people are better sexual partners than me 26. That I am a bad sexual partner 27. That I will not be able to find [a] future partner[s] 28. That my pelvic muscles will be too tight Interference with 29. Sitting 30. Walking 31. Wearing tight-fitting clothing 32. Taking part in recreational activities 33. Ability to go to work 34. Going out with friends 35. Fulfilling responsibilities to your family 36. Ability to perform tasks at work 37. Activities involving direct or indirect pressure to the vulva 38. Using sanitary pads 39. Ability to fall asleep Interference with 40. My response to sexual advances made by my partner 41. Desire for sexual activity 42. Feeling sexual pleasure 43. Orgasm frequency 44. Taking part in non-penetrative sexual activity 45. Taking part in penetrative sexual activity 46. Worrying about sexual satisfaction no longer being possible 47. Worrying that any sensation in my genitals will lead to pain
0.53 0.85 0.56 0.85 0.48 0.66 0.15
0.23
0.20
Factor 2
Factor 3
Factor 4
Factor 5
Factor 6
0.24 0.21 0.36 0.37 0.65 0.63 0.66 0.76 0.74 0.60 0.75 0.75 0.56 0.62 0.52 0.61 0.73 0.59
0.16 0.17 0.19 0.16 0.14 0.15
0.28 0.33 0.22
0.80 0.66 0.62 0.57 0.84 0.75 0.59
0.25
0.36
0.21 0.21
0.15
0.26 0.14
0.23
0.28 0.72 0.84 0.77 0.85 0.85 0.83 0.80 0.65 0.72 0.49 0.56
0.26 0.19
0.23
0.16
0.15
0.78 0.83 0.82 0.61 0.65 0.65 0.55 0.54
0.15 0.21
(continued)
J Sex Med 2017;14:1585e1596
1591
Vulvar Pain Assessment Questionnaire Table 3. Continued Factor 1 48. Taking off my clothes around my partner 49. Worrying about the next time my partner(s) will want sexual activity How often the following cause pain 50. Using tampons 51. Solitary sexual stimulation of my vulva (ie, masturbation) 52. Masturbation with partner present 53. Penetrating myself with fingers (partner absent) 54. Penetrating myself with a sex toy (partner absent)
Factor 2
Factor 3
0.23 0.33
Factor 4
Factor 5
0.27 0.30
0.35 0.60
0.28
0.19
0.29
Factor 6
0.33 0.47 0.45 0.76 0.80
*All non-significant factor loadings are not listed; all factor loadings listed are significant at P < .05.
Participants also were asked to provide feedback on the length, readability, and range of questions asked during the VPAQscreen administration. Most indicated that the VPAQ
was “good” or “excellent” when it came to length (n ¼ 159; 87.4%), readability (n ¼ 160; 87.9%), and the range of questions (n ¼ 154; 84.6%).
Table 4. Geomin-rotated factor loadings for the screening of the Vulvar Pain Assessment Questionnaire* Pain severity 1. Average pain intensity 2. Average unpleasantness 3. Average pain distress Thinking/Feeling 4. Bad about myself because of the pain 5. That my partners might think I am frigid 6. That my partners will leave me 7. That people [would] think less of me because of my pain 8. Sad 9. Unable to make changes in my life 10. That people might think I’m a bad sexual partner 11. Emotionally exhausted because of the pain 12. Anger toward my pain 13. That the pain will never stop Interference with 14. Sitting 15. Walking 16. Wearing tight-fitting clothing 17. Taking part in recreational activities 18. Ability to go to work 19. Ability to fall asleep Interference with 20. My response to sexual advances made by my partner 21. Desire for sexual activity 22. Feeling sexual pleasure 23. Orgasm frequency 24. Taking part in non-penetrative sexual activity 25. Taking part in penetrative sexual activity How often the following cause pain 26. Using tampons 27. Solitary sexual stimulation of my vulva (ie, masturbation) 28. Masturbation with partner present 29. Penetrating myself with fingers (partner absent) 30. Penetrating myself with a sex toy (partner absent)
Factor 1
Factor 2
0.76 0.92 0.70
0.27
0.18
0.22
0.70 0.63 0.65 0.73 0.49 0.51 0.65 0.56 0.50 0.48
0.38 0.18 0.28
0.19
0.82 0.87 0.80 0.88 0.70 0.53
0.26 0.37
0.12 0.15 0.21 0.17 0.36
*All non-significant factor loadings are not listed; all factor loadings listed are significant at P < .05. J Sex Med 2017;14:1585e1596
Factor 3
0.29 0.42 0.20 0.20
Factor 4
Factor 5
0.26
0.28
0.15
0.76 0.85 0.84 0.66 0.67 0.65
0.27
0.31 0.52 0.47 0.78 0.80
1592
Dargie et al
Table 5. Geomin-rotated factor loadings for pain descriptors of the Vulvar Pain Assessment Questionnaire* Factor 1 1. Burning 2. Stinging 3. Sharp 4. Stabbing 5. Aching 6. Irritating 7. Raw 8. Sensitive 9. Tender 10. Sore
0.69 0.49
Factor 2
Factor 3
0.40 0.95 0.76 0.17
0.31 0.46 0.55 0.82 0.92 0.76
0.26 0.22 0.19
*All non-significant factor loadings are not listed; all factor loadings listed are significant at P < .05.
Participants were given an opportunity to provide open-ended feedback on the VPAQ scales. Common suggestions included requests for the ability to provide more detailed responses (eg, more precise pain locations on the vulva diagram, where pain is felt during different sexual activities, additional pain descriptors, whether the pain affects caring for children), to convey how the pain has changed or affected them over time rather than focusing on the past 4 weeks, and whether they were currently engaged in other treatment for their vulvar pain (eg, medication, physical therapy). Other participants noted frustration with the process of finding medical professionals to treat their pain and expressed a need to capture those experiences alongside reports of their vulvar pain symptoms. Several also expressed gratitude for the creation of the VPAQ and the fact that researchers were studying this poorly understood condition.
DISCUSSION The purpose of this study was to evaluate the psychometric characteristics of each part of the VPAQ, to provide preliminary Table 6. Geomin-rotated factor loadings for the coping mechanisms scale of the Vulvar Pain Assessment Questionnaire* Factor 1 1. Relax my body 2. Breathe deeply 3. Go to my “happy place” 4. Practice yoga or stretching 5. Do something that takes my mind off the pain 6. Focus on staying optimistic 7. Visit my doctor(s) 8. Look for information on my pain 9. Use prescription medication 10. Talk to people in my social network 11. Talk to others with similar pain 12. Avoid anything that might cause pain
0.84 0.84 0.28 0.40
0.24
Factor 2
0.45 0.22 0.58 0.44 0.64 0.59 0.50 0.58 0.57 0.42
*All non-significant factor loadings are not listed; all factor loadings listed are significant at P < .05.
normative data for the VPAQ subscales, and to gather feedback from participants on its helpfulness and accessibility. The psychometric properties examined in this study provide further justification for use of the VPAQ in research studies, and the feedback gathered from participants shed light on the usefulness and acceptability of the scales and provide guidance on how it could be expanded in clinical settings. The results of the exploratory structural equation modeling23 confirmed that all subscales—other than the VPAQcope (see below)—had adequate to good factor structure.17e19 Although most scales met the standards for acceptable model fit, most did not exceed the more rigorous standards suggested by some.17e19 Furthermore, most items in the VPAQ subscales significantly loaded onto the same factors proposed in the original study15 in the hypothesized direction. Several significantly cross-loading items likely contributed to difficulties with model fit. However, the VPAQcope was the only scale that was consistently problematic when examining model fit and factor loadings. Our analyses did not support the 2-factor solution from the original study on scale construction, although this analysis was underpowered for the present sample.15 Future research with larger samples could determine whether these patterns are common across all samples and whether these items should be moved, modified, or deleted. Perhaps the factor structure of the VPAQ would be stronger within specific populations of people experiencing vulvar pain. For example, the questionnaire was developed with those suffering from vulvodynia in mind (ie, idiopathic chronic vulvar pain),1 but participants were not excluded from this online study based on other diagnoses (ie, they could have had another disorder associated with the experience of vulvar pain, such as lichen sclerosus, interstitial cystitis or painful bladder syndrome, or chronic yeast infections). Future studies should recruit large numbers of participants in numerous diagnostic groups to determine the validity of the questionnaire across specific diagnoses. However, the combination of reasonable fit indices and statistically significant factor loading parameters provides preliminary evidence of a psychometrically sound factor structure for all but the VPAQcope. Measures of internal consistency were strong for most of the VPAQ, similar to the original study on scale construction.15 Indeed, only the burning and stinging subscale of the VPAQdesc had scores that did not have adequate, good, or excellent internal consistency. This finding indicates that experiencing burning pain might not always correspond with the experience of stinging pain. As such, perhaps these 2 items should be considered separately instead of creating an average score of the 2 adjectives. Indeed, scores on these items were only moderately correlated compared with a strong correlation between the 2 items that compose the incisive subscale. The low internal consistency also could be accounted for by the small number of items composing this subscale.24 Moreover, it should be noted that our measure of internal consistency, Cronbach a, provided a J Sex Med 2017;14:1585e1596
1593
Vulvar Pain Assessment Questionnaire
Table 7. Geomin-rotated factor loadings for the impact of vulvar pain on romantic relationships in the Vulvar Pain Assessment Questionnaire* Factor 1 Supportive response 1. Asks what s/he can do 2. Wants to talk about it 3. Tries to acknowledge my pain 4. Seek emotional support 5. Seek physical comfort 6. Share your feelings 7. Problem solve Negative partner response 8. Gets angry 9. Blames me 10. Appears frustrated 11. Is visibly upset 12. Looks sad How has your pain affected 13. Physical intimacy 14. Emotional intimacy 15. Sexual intimacy 16. Relationship quality 17. General communication 18. Sexual communication Comfort communicating about 19. Sexual desire 20. Frequency of activity 21. Amount of “foreplay” 22. Duration of activity 23. Sexual position 24. Technique
0.55 0.52 0.52 0.87 0.84 0.85 0.56
Factor 2
Factor 3
Factor 4
0.26
0.74 0.73 0.76 0.83 0.52
0.17 0.26
0.2
0.46 0.82 0.45 0.76 0.82 0.71 0.84 0.78 0.92 0.91 0.84 0.89
*All non-significant factor loadings are not listed; all factor loadings listed are significant at P < .05.
lower-bound estimate of reliability and might underestimate the true reliability of a scale. Similar patterns were observed when examining test-retest reliability: there were strong relationships among scores at time 1 and 4 weeks later at time 2 for most subscales, although reliability was only moderate for 4 subscales. Once again, the burning and stinging subscale was in the moderate range, as was the incisive pain subscale of the VPAQdesc. Similar patterns were observed for relationship impact and impact on communications within romantic relationships, for which moderate test-retest reliability was observed. Perhaps the characteristics captured by these subscales fluctuate more frequently than other, more stable, symptoms captured by this questionnaire, such as how severe the pain is and how women with vulvar pain interact with their romantic partner when experiencing pain. It would be useful to track these variables over time to understand which factors vary, which remain stable, and how the experience of the pain is affected. Such investigations also would provide guidance on whether the subscales would need to be modified to improve reliability.
J Sex Med 2017;14:1585e1596
There were no difficulties identified with the internal consistency or test-retest reliability of the VPAQfull or the VPAQscreen, the 2 clusters hypothesized to be most important for the assessment and diagnosis of chronic vulvar pain. Furthermore, no major concerns were raised based on the model fit indices for these clusters, although it would be appropriate to revisit these analyses with different populations to determine whether the questionnaire requires revising to attain strong model fit. Thus, although the psychometric properties of the supplementary scales are not as consistently strong, the results of the present study support the use of the VPAQscreen and the VPAQfull to characterize vulvar pain symptoms. When examining the normative data for the present sample, average ratings on each scale generally indicated mild to moderate difficulties in the domains captured by the VPAQ. We suggest that a rating of at least 2 on a given subscale indicates clinical significance. Scores below 2 indicate that the person experiences mild to no pain severity, lower intensity cognitive and emotional responses, minimal interference in various life domains, and so on. Conversely, scores of at least 2 likely
1594
Dargie et al
Table 8. Internal consistency, normative data, and test-retest reliability for the VPAQ subscales VPAQ subscales* Full version Severity Emotional response Cognitive response Life interference Sexual function interference Self-penetration Screening Severity Cognitive or emotional Life interference Sexual function interference Self-penetration Pain descriptors Burning and stinging Incisive Sensitivity Coping strategies Distraction or relaxation Problem solving Partner factors Support seeking Negative response Relationship impact Communication
Time 1, Cronbach a†
Time 1, mean (SD)†
4-wk test-retest, reliability (r)‡
0.91 0.94 0.89 0.93 0.91 0.75
2.38 2.09 1.68 1.28 2.45 1.87
(0.80) (1.00) (1.13) (1.12) (1.12) (1.19)
0.74§ 0.77§ 0.78§ 0.96§ 0.87§ 0.73§
0.87 0.89 0.91 0.89 0.75
1.95 2.00 1.36 2.66 1.87
(0.85) (0.99) (1.20) (1.20) (1.19)
0.75§ 0.77§ 0.93§ 0.83§ 0.73§
0.56 0.84 0.81
2.47 (1.17) 1.91 (1.37) 2.58 (0.94)
0.66§ 0.57§ 0.78§
0.75 0.71
1.74 (0.82) 1.63 (0.76)
0.80§ 0.83§
0.88 0.81 0.85 0.95
2.11 0.87 1.50 2.01
(0.90) (0.79) (0.65) (1.15)
0.71§ 0.80§ 0.49§ 0.50§
VPAQ ¼ Vulvar Pain Assessment Questionnaire. *Mean subscale scores are based on means across items and can range from 0 to 4. † For partner subscales, n ¼ 143; for all other subscales, n ¼ 182 based on time 1 data. ‡ For partner subscales, n ¼ 50; for all other subscales, n ¼ 66 based on time 2 data. § P < .001.
represent symptoms of greater significance. However, further research is required to test and refine this hypothesis. Furthermore, there might be certain pain symptom “profiles” that could be determined for various vulvar pain subtypes (eg, provoked vestibulodynia vs generalized vulvodynia), which might further aid in the diagnostic process. However, the preliminary norms generated by this study should be interpreted with caution because of the convenience sample (ie, participants were not randomly selected from the general population) and because the diagnoses of participants were heterogeneous and could not be confirmed. Participants provided positive feedback related to their involvement in this study. Indeed, approximately half the participants indicated that simply answering a range of questions that captured their pain experiences improved their comfort levels in speaking with medical professionals about various topics relevant to their experience of pain. In addition, most participants in the present study provided positive feedback about the length of the questionnaire, the readability of the questions, and the range of characteristics captured by the items. Moreover, some participants commented that they would have liked the
opportunity to share additional details about their pain experiences, describe the pain and its impact over time, and convey their experiences seeking help and treatment methods they had attempted. Overall, these findings suggest that the VPAQ might aid in communication about various symptoms between patients and providers, and patients or participants should be given the opportunity to provide additional pertinent information or tailor the questionnaires based on the information they want to share with their providers. This study has several limitations. Although the sample size was adequate for many of our analyses, some were underpowered. The sample size also limited our ability to compare various vulvar pain diagnostic groups to determine how their different characteristics might have affected our results, particularly model fit. Furthermore, all data were collected with an online survey representing a non-random sample of convenience, and we could not confirm diagnoses. Although a strong link between selfreport and clinical diagnosis has been established,25e27 it would strengthen the findings of this study if diagnoses were validated through a gynecologic examination. Furthermore, 1 item was missing from the emotional response subscale of the J Sex Med 2017;14:1585e1596
1595
Vulvar Pain Assessment Questionnaire
VPAQfull, which could affect conclusions drawn about this subscale based on this study. Indeed, the results relating to the VPAQfull should be interpreted with caution until further studies can be conducted with all items included. We also did not include any attention checks in our online survey, which limited our ability to verify the validity of survey responses. Further, more than half the sample did not complete the survey at time 2. Although further analyses indicated no differences between those who did and those who did not complete the time 2 survey, it would be ideal to replicate these results in a sample with a less pronounced attrition rate.
CONCLUSION The results of the present study provide preliminary norms for and promising evidence of the psychometric strength of the VPAQ inventory and its usefulness from the perspective of those with chronic vulvar pain. Although model fit was not strong for some VPAQ scales, adequate model fit paired with largely strong internal consistency and test-retest reliability and strong positive feedback from participants show the overall strength of the measures. Although the VPAQcope requires revisiting in future studies, the results of this study indicate that the other VPAQ scales are ready for use in clinical and research settings. Ideally, the VPAQ will enable health care providers to more efficiently and accurately diagnose women with vulvar pain, and this study will provide further justification for its use in clinical and research settings.
REFERENCES 1. Bornstein J, Goldstein AT, Stockdale CK, et al. 2015 ISSVD, ISSWSH, and IPPS consensus terminology and classification of persistent vulvar pain and vulvodynia. J Low Genit Tract Dis 2016;20:126-130. 2. Harlow BL, Stewart E. A population-based assessment of chronic unexplained vulvar pain: have we underestimated the prevalence of vulvodynia? J Am Med Womens Assoc 2003; 58:82-88. 3. Harlow BL, Kunitz CG, Nguyen RHN, et al. Prevalence of symptoms consistent with a diagnosis of vulvodynia: population-based estimates from 2 geographic regions. Am J Obstet Gynecol 2014;210:40.e1-40.e8. 4. Arnold LD, Bachmann GA, Rosen R, et al. Vulvodynia: characteristics and associations with comorbidities and quality of life. Obstet Gynecol 2006;107:617-624. 5. Baggish MS. Diagnosis and management of vulvar vestibulitis syndrome in 559 women (1991e2011). J Gynecol Surg 2012; 28:75-82. 6. Dargie E, Pukall CF. Women in “sexual” pain: exploring the manifestations of vulvodynia. J Sex Marital Ther 2016; 42:309-323. 7. Glajchen M. Chronic pain: treatment barriers and strategies for clinical practice. J Am Board Fam Pract Board Fam Pract 2001;14:211-218. 8. Green CR, Wheeler JRC, Laporte F, et al. How well is chronic pain managed? Who does it well? Pain Med 2002;3:56-65. 9. Nicolaou A. Conference report: The Change Pain Physician Survey. Curr Med Res Opin 2011;27:2073-2074.
Corresponding Author: Caroline F. Pukall, PhD, CPsych, Department of Psychology, Queen’s University, Kingston, ON K7L 3N6, Canada; E-mail: [email protected]
10. Pancholy AB, Goldenhar L, Fellner AN, et al. Resident education and training in female sexuality: results of a national survey. J Sex Med 2011;8:361-366.
Conflicts of Interest: The authors report no conflicts of interest.
11. Shindel AW, Ando KA, Nelson CJ, et al. Medical student sexuality: how sexual experience and sexuality training impact U.S. and Canadian medical students’ comfort in dealing with patients’ sexuality in clinical practice. Acad Med 2010; 85:1321-1330.
Funding: None.
STATEMENT OF AUTHORSHIP Category 1 (a) Conception and Design Emma Dargie; Caroline F. Pukall (b) Acquisition of Data Emma Dargie; Caroline F. Pukall (c) Analysis and Interpretation of Data Emma Dargie; Ronald R. Holden; Caroline F. Pukall Category 2 (a) Drafting the Article Emma Dargie; Ronald R. Holden; Caroline F. Pukall (b) Revising It for Intellectual Content Emma Dargie; Ronald R. Holden; Caroline F. Pukall Category 3 (a) Final Approval of the Completed Article Emma Dargie; Ronald R. Holden; Caroline F. Pukall J Sex Med 2017;14:1585e1596
12. Solursh DS, Ernst JL, Lewis RW, et al. The human sexuality education of physicians in North American medical schools. Int J Impot Res 2003;15(Suppl 5):S41-S45. 13. Upshur CC, Luckmann RS, Savageau JA. Primary care provider concerns about management of chronic pain in community clinic populations. J Gen Intern Med 2006; 21:652-655. 14. Reed BD, Harlow SD, Sen A, et al. Prevalence and demographic characteristics of vulvodynia in a population-based sample. Am J Obstet Gynecol 2012;206:170.e1-170.e9. 15. Dargie E, Holden RR, Pukall CF. The Vulvar Pain Assessment Questionnaire Inventory. Pain 2016;157:2672-2686. 16. Simms LJ, Watson D. The construct validation approach to personality scale construction. In: Robins RW, Frayley RC, Knager RF, eds. Handbook of research methods in personality psychology. New York: Guilford; 2007. p. 240-258.
1596 17. Schumacker RE, Lomax RG. A beginner’s guide to structural equation modeling. 3rd ed. New York: Taylor & Francis Group; 2010. 18. Hu L, Bentler PM. Fit indices in covariance structure modeling: sensitivity to underparameterized model misspecification. Psychol Methods 1998;3:424-453. 19. Browne MW, Cudek R. Alternative ways of assessing model fit. Sociol Methods Res 1992;21:230-258. 20. Maccallum RC, Browne MW, Sugawara HM. Power analysis and determination of sample size for covariance structure modeling of fit involving a particular measure of model. Psychol Methods 1996;13:130-149. 21. Cohen J. A power primer. Psychol Bull 1992;112:155-159. 22. Vazsonyi AT, Ksinan A, Mikuska J, et al. The Big Five and adolescent adjustment: an empirical test across six cultures. Pers Individ Dif 2015;83:234-244.
Dargie et al 23. Asparouhov T, Muthén BO. Exploratory structural equation modeling, Volume 16. York: Taylor & Francis Group; 2009. 24. Nunnally J, Bernstein IH. Psychometric theory. 3rd ed. New York: McGraw-Hill; 1994. 25. Harlow BL, Vazquez G, MacLehose RF, et al. Self-reported vulvar pain characteristics and their association with clinically confirmed vestibulodynia. J Womens Health 2009; 18:1333-1339. 26. Reed BD, Haefner HK, Harlow SD, et al. Reliability and validity of self-reported symptoms for predicting vulvodynia. Obstet Gynecol 2006;108:906-913. 27. Dargie E, Chamberlain SM, Pukall CF. Provoked vestibulodynia: diagnosis, self-reported pain, and presentation during gynaecological examinations. J Obstet Gynaecol Can 2017; 39:145-151.
J Sex Med 2017;14:1585e1596