Spanish cross-cultural adaptation and validation of the Australian Pelvic Floor Questionnaire in running women

Pelvic floor dysfunctions are a wide range of disorders in the gynaecological, lower urinary and gastrointestinal tracts that affect the structure and/or function of the pelvic organs. The objective of this study was to carry out a cross-cultural adaptation and a psychometric analysis of the Spanish version of the Australian Pelvic Floor Questionnaire. Observational study divided into two main phases: (1) translation and cross-cultural adaptation and (2) psychometric tests. Women runners from all over the Spanish territory, from different federations, clubs and levels were recruited. Participants: 424 female runners, native Spanish, over 18 years of age and who had been practicing running for more than 6 months. The instruments used in this study were the Australian Pelvic Floor Questionnaire, Female Sexual Function Index, King Health Questionnaire, Quality of Life SF-12 and EuroQoL 5-D. The Spanish version of Australian Pelvic Floor Questionnaire has proven to be an understandable and easy-to-use tool. The general internal consistency of the questionnaire was 0.972 and the intraclass correlation coefficient ranged between ICC 0.596–0.960. The Spanish version of Australian Pelvic Floor Questionnaire is a valid and reliable measure that can be used clinically to assess pelvic floor dysfunctions among the female Spanish population.

Participants. Women runners from all over the Spanish territory, from different federations, clubs and levels were recruited. The inclusion criteria were: (1) Spanish natives over 18 years of age, (2) female runners and (3) more than 6 months practicing running sports. On the other hand, the study excluded: (1) those participants who abandoned the study without answering any of the questions of a questionnaire were excluded, and (2) those who presented a cognitive impairment that did not allow them to understand and/or answer the forms.
Ethical considerations. This study was developed following the recommendations of the Declaration of Helsinki in accordance with the ethical principles for research in human beings, and the data were used in accordance with Organic Law 3/2018, of December 5, on the Protection of Personal Data and guarantee of digital rights. All participants signed an informed consent to be part of the study. In addition, the Ethics Committee of a Spanish University approved the realisation of this study, with protocol number UVIC-CCC 81/2019. Translation and cross-cultural adaptation. To ensure terminological and conceptual equivalence, in the questions that make up the APFQ, the recommendations of the International Test Commission Guidelines for test translation and adaptation were followed 26 , as well as those of the World Health Organisation (WHO) 27 .
The process of the adaptation of the Spanish version of the APFQ from it English version can be broken down into a 5-step protocol: English to Spanish translation of the APFQ, performed by two independent and blinded native Spanish speakers; the two independent versions of the APFQ-Sp were compared and an agreement was reached to develop the preliminary version of the APFQ-Sp; subsequent back-translations (from Spanish to English) were done independently by two native English translators. Any discrepancies in the translation were discussed and resolved by a committee of 5 experts, obtaining a preliminary version. The preliminary version of the APFQ-Sp was subjected to a pilot test, with a sample of 25 participants (Fig. 1).  17,18,19). The subscales range from 0 (or 1) to 5, and the sum of the scores of these six subscales yields an overall female sexual function score (with higher scores indicating better sexual function), with a Cronbach's alpha of 0.97 for the total score 28 . In a Spanish version tested in menopausal women, Cronbach's alpha for the total score was 0.964 29 , and in another Spanish version, Cronbach's alpha was 0.85 for the total score 16 . www.nature.com/scientificreports/ mental health). Both additional domains and summary scores were calculated using algorithms where each item response has an individual weight in the total score. Higher scores indicate better perceived HRQoL 32 .
EuroQoL 5-D. EuroQol-5D is a questionnaire to measure people's quality of life. It is made up of 5 domains (mobility, self-care, regular activities, pain/discomfort, and anxiety/depression) divided into three severity levels (no problems, some problems or moderate problems, and serious problems). This system also includes a visual analog scale (EQ-5D VAS) defined by a vertical scale of 10 cm at each end, which are the extreme expressions of self-perceived health status ranging from 0 (worst health) to 100 (best health). Its reliability obtained a Cronbach's α = 0.53 33 .
Data collection. All participants completed the following questionnaires: The Spanish version of the Australian Pelvic Floor Questionnaire (APFQ-Sp), Female Sexual Function Index (FSFI), King Health Questionnaire (KHQ), EuroQoL Quality of Life Questionnaire (5D and VAS), the questionnaire on the general state of health SF-12 (Short form-12) and sociodemographic information. Following the results published in previous studies 34,35 , in which higher levels of internal consistency and reliability are observed in periods of less than 7 days between the two measurements, the APFQ-Sp was filled twice with a difference of 3-5 days between measurements . The FSFI, KHQ and QoLSF-12 questionnaires were used to analyse the construct validity of the APFQ-Sp. The data were obtained between October 2020 and July 2021. Two blinded researchers external to the study performed the data collection, as well as the data analysis.

Data analysis.
A frequency analysis of some of the characteristics of the sample was performed, as well as a descriptive analysis of the sociodemographic variables, including the outcome measures used (APFQ-Sp, IFSF, KHQ, SF-12 and EuroQoL5D), calculating the mean and the standard deviation. To analyse the distribution and normality of the sample, the Kolmogorov-Smirnov test was used. Floor and ceiling effect were analysed. The Cronbach's α coefficients were calculated to analyse the internal consistency of the measures. In addition, the response to the item was analysed using the Intraclass Correlation Index (ICC-2:1). The reliability values were classified according to the following scale: Poor: ≤ 0.40; Moderate: 0.40-0.60; Good: 0.60-0.80; Excellent: ≥ 0.80 36 .
The formula SEM = s √ 1 − r was used to calculate the standard error of measurement (SEM). For both measures (APFQ-Sp1 and APFQ-Sp2) the test score's standard deviation was "s", and "r" was Pearson's correlation coefficient. Following the analysis described by Stratford 37 , to measure the sensitivity of the tool, the minimal detectable change 90 (MDC90) was used. The formula used to calculate the MDC90 was as follows: MDC90 = SEM × √2 × 1.65. The floor or ceiling effect was considered to be present if more than 15% of the participants reached the lowest or highest score, respectively.
The structure and validity of the construct was analysed from the extraction by maximum likelihood (EMV). To maintain the original structure of the APFQ, a 4-factor forced model was performed. In addition, to perform the EMV, the requirement of a minimum of 10 subjects per item was satisfied (minimum number 420 -subjects measured 424) 38 .
To perform the statistical analysis of this study, the SPSS statistical treatment programme (V.23.0) was used.

Results
Translation and cross-cultural adaptation. The translated and culturally adapted version of the APFQ into Spanish (APFQ-Sp) is presented in Supplementary File S1. The Table 1 shows the anthropometric characteristics of the participants. The total of 424 women who participated in this study had a mean age of 38.56 (± 9.064) years. More than 75% of the participants had a university level of education (bachelor's, master's or doctorate). Almost half of the participants had not had a previous pregnancy. In addition to this, the type of delivery, number of previous abortions, etc., can be analysed in depth. On the other hand, the sports and federative data of the participants are also presented. Table 2 shows the mean, minimum, maximum and standard deviation values of all the assessment tools used in this study, that is, the Australian Pelvic Floor Questionnaire (APFQ-Sp), the Female Sexual Function Index (FSFI), King Health Questionnaire (KHQ), EuroQoL Quality of Life Questionnaire (5D and VAS), the questionnaire on the general state of health SF-12 (Short form-12). Moreover, the values of the different sub-indices or sections in which the different tools are divided are presented. When performing the floor effect and ceiling effect analysis, it was observed that 23 (0.54%) and 16 (0.37%) participants reached the minimum and maximum APFQ-Sp score, respectively. Given these results, the floor/ceiling effects were considered not relevant.
The minimum and maximum values of the APFQ-Sp questionnaire were reached by 1.65% and 4.48% of the participants, respectively, completing the questionnaire in an average time of 18 min. The general internal consistency of the questionnaire was 0.972 and the intraclass correlation coefficient ranged between ICC: 0.596 -0.960) ( Table 3). On the other hand, the SEM and MDC90 values were 0.04 and 0.009, respectively.
In construct validity, the maximum likelihood extraction method presented a value of 0.858 in the Kaiser-Meyer-Olkin test, with a significant value in the Bartlett sphericity test (p < 0.001) and a Chi-square value of 10,432.61 in the Bartlett sphericity test and in the Kaiser-Meyer-Olkin test (0.833). The APFQ-Sp presents a solution of two factors, since they are the only two factors that explain more than 10% of the variance each (18.737% and 12.521%, respectively); however, there is a wide distribution of the variance explained in the APFQ-Sp, since up to question 14 the explained variance exceeds 2% (Table 4). Figure 2 shows the sedimentation graph, while Table 5 shows the load of each of the questions in the two factors that met the established criteria; When analysing the correlation between the total value of the APFQ-Sp and its sub-scales, it is observed that the levels of correlation oscillate between r = 0.103 (Defecation Section -Sexual Activity Section) and r = 0.752 (Total APFQ -Sexual Activity Section) ( Table 6). In addition, in the calculation of the criterion validity (Table 6) Table 6.

Discussion
This study aimed to carry out a cross-cultural adaptation and a validation study of the tool for the assessment and monitoring of pelvic floor dysfunction APFQ into Spanish. The translation and cross-cultural adaptation of the APFQ-Sp was carried out following the recommendations of the literature, which ensures the conceptual equivalence between the translated version and the original version and it is essential to facilitate the use of the APFQ-Sp among Spanish speakers, while favouring the comparison of potential results with versions of the APFQ published in other languages. Based on the cross-cultural adaptation process carried out, the APFQ-Sp proved to be an understandable and easy-to-use tool.

Construct validity.
To assess the construct validity, the structure of the original version of the APFQ was taken into account, where 4 factors are identified. In this sense, two factors explain a level of variance greater than 10% and, in addition, in the scree plot they show a change in proportion in the level of explained variance. In this sense, if all the criteria that are usually considered for factor extraction had been taken into account (> 10% of the variance, eigenvalue > 1.0, and scree plot inflection point), two factors of APFQ-Sp would have been extracted. In this sense, the only version whose construct validity has been analysed is the Arabic version, which shows values of KMO = 0.806 and Bartlett sphericity test = 4150.46. It would be interesting to carry out studies to analyse the construct validity of the rest of the versions and determine whether they behave in a similar way as the versions that have performed such analysis (Spanish and Arabic).
Internal consistency and test-retest. The internal consistency in the APFQ-Sp showed a Cronbach's α of 0.972, and, in the sub-scales, it ranged between Cronbach's α = 0.828 (sexual activity section) and 0.935 (urinary tract section) ( Table 3). These values are slightly higher than those observed in the Arabic (0.877) 21 , Chinese (0.83-0.89) 20,40 , Serbian (0.822-0.846) 19 , Turkish (0.733-0.858) 22 and original version (0.74-1.00) 18 , although all except one dimension from the Turkish version and one from the original version are considered to have excellent internal consistency 36 .  (Table 3), it is observed that they are consistent with the Arabic version 21 , which presents ICC values: 0.500-0.833. However, these values are slightly lower than those observed in the Serbian (ICC: 0.896-0.944) 19 , Turkish (ICC: 0.876-0.954) 22    www.nature.com/scientificreports/ that the total value of the questionnaire tends to correlate better with all the reference questionnaires in comparison with the dimensions of the APFQ-Sp. In this sense, when the results are compared with other versions, it is observed that, with the exception of the original version, the APFQ-Sp is the only one that evaluated this psychometric aspect. The original version performs a convergence analysis with the short version of the Urogenital Distress Inventory (SUDI), showing correlation levels of r = 0.80, while the level of correlations between pelvic organ prolapse and prolapse symptoms quantification measurements (measured in 106 patients) showed a range of r = 0.25-0.68. In this sense, it is observed that the APFQ complements very well with other questionnaires or diagnostic instruments for patients with pelvic floor problems, although it would be interesting to know the level of correlation of the other versions to have a much more complete perspective.
Implications for future research. There is a need for developing valid and reliable instruments to measure pelvic floor dysfunctions in order to provide accurate and consistent measurements over time 42,43 . These instruments must be concise, valid, reliable, evidence-based and developed using best practices 42,43 . In this context, the APFQ is a measure that was proposed to evaluate pelvic floor dysfunctions in women 18 . The APFQ was developed based on the most valid and reliable questions to focus on the main pelvic floor dysfunctions in women through a systematic review of the literature to identify measures with the best psychometric properties. This study provides evidence for the validity of the APFQ-Sp. In this sense, the APFQ-Sp is a concise, valid, reliable and evidence-based document and, at the same time, it is an instrument developed using best practices. Therefore, the APFQ-Sp is a measure that can be recommended for the assessment of pelvic floor dysfunction in Spanish female runners.  www.nature.com/scientificreports/ The current study demonstrates that the APFQ-Sp is a valid measure to assess pelvic floor dysfunction in the Spanish population, which allows researchers and clinicians to use this tool within both clinical and research settings. In this sense, research on the pelvic floor is an area of special interest, since it has implications for the development of interventions for both the prevention and treatment of pelvic floor dysfunctions in the female population. In addition, future research should study the APFQ in different clinical populations, such as cancer related to the pelvic floor, or analyse some psychometric variables that have not been taken into account in this validation study, and that are linked to longitudinal studies, such as the sensitivity to change.
Strengths and weaknesses. This study presents some strengths that show the appropriateness of its execution. The main strength is that it allows the APFQ to be adapted into Spanish, the second most widely spoken language in the world and one of the five official languages of the UN. On the one hand, this cross-cultural  20,40 versions. In addition, the cross-cultural adaptation and the subsequent validation study were carried out respecting the minimum number of subjects recommended in the literature 38 . In this sense, there would be 420 minimum necessary subjects, and it was carried out with 424 participants. However, there are some weaknesses that must be taken into account when interpreting the results of this study. Specifically, this study did not perform the psychometric analysis of longitudinal variables, such as error scores, responsiveness or sensitivity to change. In this sense, future studies should be designed and executed in order to assess these psychometric variables in the APFQ-SP.
The cross-cultural adaptation and validation of the APFQ-SP has been carried out in a specific group of women with a higher level of incidence of PFDs. However, there are other population groups that also have a higher incidence of PFDs, such as multiparous women or women over 65 years of age, so future studies should be designed to validate the APFQ in these specific population groups.

Conclusions
The Spanish version of the APFQ is a valid and reliable measure that can be used clinically to assess pelvic floor dysfunctions in the Spanish female population. This instrument is complete and includes different dimensions on the most relevant aspects and symptoms of female pelvic floor dysfunctions, allowing its use by both researchers and clinical professionals, who speak Spanish, for the evaluation and follow-up of patients with pelvic floor dysfunctions. Table 6. Correlation matrix between the APFQ-Sp, its different dimensions and the IFSF, KHQ, SF-12 and EuroQoL-5D questionnaires.