Study Design
Measure development
Prolapse-specific BI themes that were identified in our previous work were used as a framework for the development of potential questions for use in our measure. Five individuals (4 urogynecologists, 1 qualitative researcher) wrote candidate items based on the general themes of sexuality and intimacy, sense of femininity and attractiveness, and concealment. The items were written to assess the 4 previously described components of BI: (1) perception (self-rated severity), (2) affect (satisfaction or dissatisfaction, bother, self-consciousness), (3) behavior (avoidance, concealment), and (4) cognition (beliefs or thoughts associated with the condition). Candidate items were written for an 8th-grade reading level with both affirmative and negative versions. To evaluate the initial pool of 75 candidate items, we assessed face validity and used modified methods that had been described by Walsh et al to evaluate item clarity, assumptions, and knowledge; we then eliminated items that did not meet these criteria. We chose the affirmative versions of the questions to avoid reverse scoring issues and subject burden with positive and negative question versions. We reviewed the remaining items for theme redundancy and selected or eliminated items by consensus decision. We chose 21 items for the initial version of the Body Image in Pelvic Organ Prolapse questionnaire (BIPOP), and all items used the same 5 Likert scale responses (strongly agree, agree, neither agree nor disagree, disagree, strongly disagree). We created 2 versions of the instrument for partnered and nonpartnered women, because themes of sexuality and intimacy and some aspects of attractiveness could seem less relevant to women without partners. We modified phrasing of questions in the nonpartnered version to reflect how a woman without a partner might feel in the situation described. As an example, the nonpartnered version was phrased “Because of my prolapse, I would worry that my partner might avoid being intimate with me,” compared with the partnered version “Because of my prolapse, I worry that my partner might avoid being intimate with me.” Instructions for administration and scoring were incorporated into the questionnaire. A higher BIPOP score indicated better BI and less impact of prolapse on BI. In the initial version of the BIPOP, items were scored from 1-5, and 2 items were reverse-coded. The BIPOP is scored as a mean score because the generation of a mean value automatically accounts for missing data by excluding values from missing items in both the numerator and the denominator when the mean is calculated. Scoring options for both a mean and total are available.
Initial measurement evaluation
We approached women who were seeking care at the University of Pittsburgh and the Medical University of South Carolina after the institutional review boards at both sites approved the study. Women who were ≥18 years old with stage ≥II symptomatic prolapse by POPQ examination were offered participation. Symptomatic prolapse was defined as answering “Yes” to 1 or both of the following questions from the Pelvic Floor Distress Inventory: (1) “Do you usually have a sensation of bulging or protrusion from the vaginal area?” and (2) “Do you usually have a bulge or something falling out that you can see or feel in the vaginal area?” Women were excluded if they were asymptomatic or unable to complete informed consent and study forms, which was determined by clinician judgment. Participants signed informed consent at enrollment. The South Carolina site was involved later in the study to help complete recruitment.
Cognitive interviews were performed at the study outset to determine whether any candidate items needed modification, whether the items were not understood as intended, and to further confirm face and content validity. At the Pittsburgh site only, participants were offered participation in the cognitive interviews until 5 partnered and 5 nonpartnered participants were enrolled. These 10 women completed the study questionnaires and then were interviewed about the BIPOP by a research assistant who used an interview guide to assess (1) comprehension, (2) decision processes, such as motivation and social desirability, (3) response processes, (4) relevance and importance, and (5) comprehensiveness.
An exploratory factor analysis (EFA) that used principle components with Varimax rotation was performed on the baseline BIPOP data to determine the number of underlying factors and the candidate items that load on or associate with those factors. Factor loadings range from –1.0 to +1.0 and represent both the strength and direction of the association of a particular item with the underlying factor. The factors themselves represent groups of conceptually related items that were identified through EFA on the basis of the intercorrelations among those items. The factor analysis was also used to eliminate items based on cross-loading (ie, items that were associated at least moderately with >1 factor) or on failure to load strongly on at least 1 factor (loading, > .30). The EFA was an iterative process and was repeated until a parsimonious model was created. Next, separate EFAs were performed on the dataset by partner status (partnered or nonpartnered). Cronbach’s α was calculated for the individual subscales and the total scale. When Cronbach’s α is interpreted, values >0.90 are considered “excellent”; values 0.80-0.89 are considered “good”; values 0.70-0.79 are considered “acceptable,” and values <0.69 are considered “questionable.”
Further evaluation of reliability and validity
We performed several types of validity assessments throughout the development process, which included assessment of face validity, content validity, construct validity, tests of association and difference, and test-retest analyses. During the initial review of questions by the investigators and the cognitive interviews with participants, we assessed face validity (the extent to which questions appeared to measure the BIPOP construct) and content validity (the extent to which questions represented the construct). Construct validity was based on examination of the degree of association between BIPOP scores and other key measures (convergent construct validity). Tests of association and difference were performed between the BIPOP subscales and study measure, with the use of Pearson’s correlations, the Student t test, and 1-way analysis of variance. Using intraclass correlation, we performed test-retest analyses of the BIPOP with data from 100 participants who completed the BIPOP once at enrollment and again 2 weeks later.
Study measures
The study measures included self-reported demographic items (age, race, ethnicity, and partner status), pertinent past medical and surgical history, current medications, and completion of validated questionnaires that assess general BI, symptoms and distress related to pelvic floor disorders, sexual function, depressive symptoms, and self-esteem. Expected associations between our BIPOP and the study measures are shown in Table 1 . Two validated general BI measurements were used in this study: the Body Exposure during Sexual Activity Questionnaire, which was chosen because it assesses BI in intimate situations and thus serves as a proxy measure for prolapse BI, and the Body Image Quality of Life Inventory, which is a general BI measure that is designed to assess the impact of BI on quality of life.
| Measure | Predicted association with Body Image in Pelvic Organ Prolapse questionnaire | Measure characteristic | Baseline study scores |
|---|---|---|---|
| Body Exposure during Sexual Activity Questionnaire | – | Validated assesses anxious/avoidant body focus during sexual activity; higher score reflects more self-conscious focus and avoidance; score range: 0–4 | 1.7 ± 1.0 a |
| Body Image Quality of Life Inventory | + | Validated general body image measure; assesses impact of body image on quality of life; higher score represents a more positive effect of body image on one’s life and a lower scores reflects a more negative effect; score range: −3 to +3 | 1.3 ± 1.0 a |
| Pelvic Floor Distress Inventory-20 | – | Validated; contains 3 subscales (prolapse, urinary, bowel); assesses symptoms of distress related to pelvic floor disorders; higher score represents more symptom bother; score range: 0–300 | 104.4 ± 63.0 a |
| Pelvic Organ Prolapse Distress Inventory | 40.8 ± 24.3 a | ||
| Urinary Distress Inventory | 36.6 ± 27.0 a | ||
| Colorectal-Anal Distress Inventory | 27.1 ± 23.2 a | ||
| Pelvic Floor Impact Questionnaire-7 | – | Validated; contains 3 subscales (prolapse, urinary, bladder); assesses life impact of pelvic floor disorders; higher scores represent greater impact on quality of life; score range: 0–300 | 47.6 (19.1–130.9) b |
| Pelvic Organ Prolapse Impact Questionnaire | 14.3 (4.8–47.6) b | ||
| Urinary Impact Questionnaire | 19.1 (4.9–52.3) b | ||
| Colorectal-Anal Impact Questionnaire | 4.8 (0.0–33.3) b | ||
| Prolapse and urinary incontinence impact of sexual function questionnaire-12 | – | Validated; assesses sexual function as impacted by pelvic floor disorders; higher scores represent better sexual function; score range: 0–48 | 30.0 ± 6.2 a |
| Patient Health Questionnaire-9 | – | Validated; assesses depressive symptoms; higher score represents more depressive symptoms; score range: 0–27 | 4.0 (0–39) b |
| Rosenberg Self-esteem questionnaire | + | Validated; assesses global feelings of self-worth; higher score represents greater levels of self-esteem; score range: 0–30 | 27.4 ± 6.1 a |
Pelvic Organ Prolapse Quantification examination
The Pelvic Organ Prolapse Quantification examination was performed as part of clinical care as described by Bump et al, was abstracted from the clinical record, and was used to define prolapse stage.
Sample size considerations
We estimated that enrollment of 240 women would meet our goal of 210 participants who completed the baseline study questionnaires. Our goal of 200 participants was based on a general rule of having approximately 10 participants per item for each of the 21 BIPOP items for EFA. We anticipated that approximately 20% of the participants would not return a complete baseline questionnaire dataset and thus planned to enroll up to 240 women. We planned to recruit 10 women, 5 partnered and 5 nonpartnered, to participate in the cognitive interviews. For test-retest analyses, we planned a convenience sample of one-half the cohort (approximately 100) to repeat the BIPOP measure.
Results
Two hundred eleven women were enrolled in this study, and 201 women had complete study questionnaires (completion rate, 95%). Mean age of study participants was 60.2 ± 10.5 years, and 83% identified themselves as postmenopausal. Participants had a median Pelvic Organ Prolapse Quantification stage of III (range, II–IV), mean body mass index of 28 ± 6 kg/m 2 , and 60% nonsmokers. They described their race/ethnicity as 97.6% white, 1.9% black, and 0.5% Hispanic. More than 80% of participants had a partner; 73% were married; 14% were divorced; 7% were widowed; 4% were single, and 2% were classified as “other.” Forty percent of the women reported having had a hysterectomy; 22% had previous prolapse or urinary incontinence surgery; 17% had previous breast or plastic surgery, and 2% had previous bariatric surgery. The recruitment rate between partnered and nonpartnered women was not significantly different. Two hundred one participants were enrolled at the University of Pittsburgh site between January 2011 and July 2012, and 10 participants were recruited at the Medical University of South Carolina site between February 2012 and July 2012.
Cognitive interview and baseline scale results
Review of the cognitive interview transcripts confirmed participants’ understanding of each measure item (in comparison with the measure designers’ intent) and measured comprehension, relevance, and clarity of questions and acceptability of length and subject matter. No significant changes in the candidate questions or Likert responses were made based on the cognitive interviews. Interview responses from participants who completed the questionnaires supported the face and content validity of study questions. Baseline scores for each of the study measures are listed in Table 2 and were similar to those reported previously in comparable populations.
|
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
