A cross-sectional study with repeated measures assessment, which started with the transcultural translation and adaptation of the PIKQ to the Portuguese language according to Beaton et al.,16 and proceeded with the validation. This study collected data from December 2020 to December 2021 with 130 women. All volunteers who had scheduled a medical appointment at the public climacteric outpatient clinic in the city of Recife, Pernambuco, Brazil, and who attended on the data collection days were included. Women with a medical diagnosis of UI, aged over 31 years, were included. Volunteers who did not understand or speak Portuguese were excluded.

The PIKQ is a reliable and validated instrument that examines knowledge about UI and POP in general gynecology and urogynecology women's patients. It is composed of two scales with 12 questions each. The item receives a score of one if answered correctly and zero if answered incorrectly or as “I don't know.” The average score is calculated for each scale individually through the sum of the correct answers. The scores range from 0 to 12, and higher scores indicate better knowledge.11

The patients for the PIKQ study11 were recruited from the gynecology and urogynecology clinics. Both scales had excellent internal consistency (Cronbach's alpha > 0.8). Mean test scores were slightly higher upon retest (UI, 0.46; POP, 0.33). The Pearson's correlation between initial and repeat scores was high for UI (0.67) and POP (0.94).

The translation and cross-cultural adaptation followed five stages: 1. initial translation into Portuguese, 2. synthesis of translations, 3. back-translation, 4. the Committee of Experts, and 5. testing of the pre-final version of the questionnaire.17-19

The English version was submitted to the initial translation into Brazilian-Portuguese by four translators (T): one of whom was a layperson on the subject (T4) and three experts on the topic (T1; T2; T3). The translators were invited to fill in the “additional comments” when uncertainties arised. An impartial person was added to mediate discussions. A synthesis was produced from working with the original quiz and the versions of all translators (T1, T2, T3, and T4), resulting in a single document (T-1234).

The T-1234 version was translated back into the original language. This is a validity process to ensure that the translated version accurately reflects the item content of the original version.17 The back-translations were produced by two bilingual people with English as their native language. To avoid bias, the translators were not informed of the concepts and did not have access to the original version.

The expert committee included translators, healthcare professionals, and experts in women's health. They reviewed all translations, discussed possible differences, and developed the final version.

The translators were asked to achieve equivalence of the source (English) and target (Portuguese) words and phrases, taking into account four areas: 1. Semantic equivalence (i.e., does the words mean the same thing?); 2. Idiomatic equivalence: colloquialisms are difficult to translate. The committee was asked to formulate an equivalent expression in the target version; 3. Experiential equivalence: the items seek to experience everyday life; however, often in a different culture, a certain attitude may not be experienced. The questionnaire item would have to be replaced with a similar item that is experienced in the target culture; and 4. Conceptual equivalence: words often have different conceptual meanings between cultures. The committee must examine the source and back-translated questionnaires for all such equivalences.17

The pre-test is the final stage of the process and uses the pre-final version with between 30 and 40 people.17 In our pre-test, 43 women aged over 31 years, with UI, answered the last version of the quiz, and expressed their impressions about answers, instructions, and layout. The interviewer had to be aware of many identical responses to different items, which may lead to exclusion from the sample. This stage has aspects of content and face validation.16

We followed the COnsensus-based Standards for the selection of health status Measurement Instrument's (COSMIN) to assess the validity (content and face, structural, and hypothesis testing) and reliability (internal consistency, test-retest, and measurement error) of the PIKQ.17

A sample of 130 women was used to verify the internal consistency.19-21 There was a random selection among patients with a stable UI condition to select 30 participants to complete the instrument again 7–14 days after to verify the reliability. This period was used to avoid recalling the previous answers rather than giving an independent response, in scales with few items.22

Statistical analyzes were performed by JAMOVI version 2.2.5 and SPSS AMOS version 26, with a significance level of 5 %. The absolute and relative frequencies of the missing values were evaluated for content validity analysis.

Permission to translate the PIQK was obtained through email with the authors. This study is based on the Resolution of the National Health Council and was approved by the Ethics and Research Committee of the Universidade de Pernambuco, CAAE: 39,871,920.4.0000.5195. All participants signed an informed consent form.

The sample consisted of 130 women with a median age of 56 years (range from 31 to 83) and complaints of UI. Table 1 shows the profile and clinical characteristics of the sample.

A total of 43 participants, with a median age of 63 years (31–83) (Table 1), were interviewed. All participants reported that the format of the questions of the PIKQ 12 item-version was adequate. There were no suggestions for changes. No missing items were observed. But, four women (9.1 %) reported having difficulty understanding question 4 on the PIKQ-UI, while five women (11.6 %) reported having difficulty understanding question 7 on the PIKQ-POP. None of the questions presented a misunderstanding rate greater than 20 %.

The EFA results demonstrated that the PIKQ-UI with 12 questions and one factor presented a cumulative explained variance of 17.90 %, with a KMO index of 0.60 and Bartlett's index (p < 0.001). However, only four questions had adequate factorial loading (> 0.30), and two questions had negative loadings (Table 2), evidencing the inadequate factorability of the single-factor model.

The PIKQ-UI was tested with two factors and 10 questions (questions 7 and 9 were excluded because they had negative factor loadings and/or <0.10 in all factors) and showed adequate values for KMO (0.68), Bartlett's index (p < 0.001). All questions had adequate factor loading (> 0.30) for one of the tested factors, and the percentage of cumulative explained variance was 42.50 %.

The EFA results for the PIKQ-POP showed that the model with 12 questions and one factor had an inadequate KMO value (0.55), and the percentage of cumulative explained variance was 9.57 %. Only five questions had factor loadings greater than 0.30, and two questions had negative factor loadings. The model was tested with two factors and 10 questions (questions 6 and 9 were excluded because they had factor loadings lower than 0.10 in all factors) and showed adequate values for KMO (0.60) and Bartlett's index (p < 0.001). All questions had factor loadings greater than 0.30 in at least one factor, except questions numbers 1, 7, and 8.

Given the inadequate results for the factorability and adjustments of the PIKQ-UI and PIKQ-POP with a single factor, we performed the CFA for the alternative models with two factors and 10 questions. After completing the bootstrap ML method, the Bollen–Stine p values for the PIKQ-UI 10 items (p = 0.47) and PIKQ-POP 10 items (p = 0.59) were acceptable. The statistical adjustment for the final two-factor model with covariance parameters of PIKQ-UI and PIKQ-POP 10 items respectively showed the values: CAIC 165.02 and 154.81, CFI 0.95 and 1.00, TLI 0.95 and 1.07 and, RMSEA(CI 90%) 0.04 (0.00 – 0.08) and 0.00 (0.00 – 0.06). All models demonstrated proper fit. Fig. 1 shows the CFA results, with the factor loading of the PIKQ-Br with 10 questions.

Fig. 1.

Path diagram demonstrating the factor structure of PIKQ-IU (A) and PIKQ-POP (B) and describing the factor load of each question. Q = question; e = error.

(0.24MB).

Positive and moderate correlations for hypotheses were observed between the scores of the PIKQ-UI (rho = 0.430) and PIKQ-UI (10 questions) (rho = 0.439) and the NRS-UI. Positive and weak correlations regarding POP knowledge were observed between the PIKQ-POP (rho = 0.241) and PIKQ-POP (10 questions) (rho = 0.278) scores with NRS-POP. The final version of the PIKQ-Br with 10 questions can be seen in Fig. 2. The scores followed the same used in the original version, with the total score for each scale, however, ranging from 0 to 10. The item receives a score of one if answered correctly and zero if answered incorrectly or as “I don't know ”. Higher scores indicate better knowledge.

Fig. 2.

Final version of the PIKQ-Br.

(1.02MB).

Cronbach's alpha coefficient revealed inadequate indicators for the internal consistency analysis of the models with a single factor and 12 questions of both scales. Cronbach's alpha coefficient was 0.571 for PIKQ-UI and 0.472 for PIKQ-POP. The correlation values between the questions and the PIKQ-UI score ranged between 0.35 and 0.54, except for items 7 and 9, which showed a negative correlation. The correlation values for the PIKQ-POP between the questions and the total score ranged between 0.23 and 0.54, except for question 6, which showed a negative correlation.

Cronbach's alpha coefficient to assess the internal consistency of the models with two factors and 10 questions was 0.719 for the PIKQ-UI and 0.676 for the PIKQ-POP, suggesting a good and minimally acceptable level, respectively. The correlations between the questions and the total scores varied between 0.30 and 0.54, demonstrating an adequate correlation of each question with the total score. Cronbach's alpha coefficients for PIKQ-UI (10 questions) with one and two factors were 0.674 and 0.718, respectively, and 0.613 and 0.710 for the PIKQ-POP.

Table 3 shows the test-retest reliability and measurement error indicators. All versions of both scales showed excellent reliability (ICC > 0.93). The 12 and 10-item versions of the PIKQ-UI had SEM values of 0.55 and 0.24 and SDC values of 1.51 and 0.66, respectively. The 12-item version of the PIKQ-POP had an SEM of 0.52 and an SDC of 1.44, and the version with 10 items presented values of 0.31 and 0.85, respectively.