Consistency of three different questionnaires for evaluating sexual function in healthy young women

Most studies on female sexual dysfunction (FSD) are performed in population inventories and under specific clinical conditions such as postpartum, urinary incontinence, post-breast cancer surgery, obesity, and diabetes, among others [1‐5]. These approaches are performed using validated psychometric scales, with the most frequently used being FSFI [6], GRISS [7] and, also, the SQ-F in the Brazilian population [8].

Female sexual dysfunction is a complex and multi-faceted disorder that has a wide spectrum of symptoms and severity and during their life cycle, women are frequently incapable or disinterested in participating in a sexual relationship. The causes are often multifactorial, requiring multidisciplinary assessment that addresses biological, psychological, sociocultural and relational aspects [9]. The concept of female sexual response has changed since Basson et al. [10] proposed the inclusion of an intimacy-based sex response cycle, in a nonlinear model that incorporates the value of emotional intimacy, sexual stimulus, and satisfaction with the relationship to replace the current linear model. In this context, in addition to a reproductive repertoire, sexuality comprises both physical and psychological feelings during part of a woman’s lifetime both for their own feeling as well as in relation to their partner [11]. In this perspective, scales which have included domains of interaction between sexual partners could have a higher chance to better diagnose sexual dysfunction.

According to the Fifth Edition of the Diagnostic and Statistical Manual [12] three disorders are cited: Female Orgasmic Disorder (FOD); Female Sexual Interest/Arousal Disorder (FSAD), and Genito-Pelvic Pain/Penetration Disorder (GPPPD). These disorders include four main domains for desire, arousal, orgasm, and pain (dyspareunia and vaginism), in which both physiological and psychological aspects of the female sexual experience influence the response [13].

Sexual dysfunctions are highly prevalent in the sexual disorders of both sexes according to Laumann, Paik & Rosen [14] who analyzed 1749 women and 1410 men, aged 18–59 years, living in the USA. The authors found that sexual dysfunction is more prevalent in women (43 to 34%), where the main complaints range from 27 to 32% for lack of interest in sex, 22–28% for inability to reach orgasm, 8–21% for pain during sex, and 17–27% for unpleasant sex.

For the Brazilian population, a study of 1219 women with a mean age of 35.6 + 12.31, observed that 49.0% exhibit at least one sexual dysfunction, 26.7% for desire, 23% dyspareunia and 21% for orgasm. In this study, educational level was inversely correlated with the risk of sexual desire, orgasm and pain dysfunctions during sexual intercourse [15]. In another cross-sectional study of 201 sexually active Brazilian women aged from 18 to 45 years, 90 from a public health service and 111 from private services, dysfunction occurred in 30.2% without a significant difference in the prevalence of sexual dysfunction or in sexual domain scores between income or educational status [16]. A prevalence of at least one type of sexual dysfunction in 36% of a sample of 100 women aged between 20 and 39 years old was also found [17]. Orgasm dysfunction occurred in 18% of the participants, and dyspareunia was reported by 13% of the women during the month prior to the study. Sexual appetite dysfunction occurred in 11% of the women, 8% had excitation dysfunction and only 1% reported vaginismus.

Even though there is high prevalence of sexual dysfunction in women, both physicians and patients are frequently hesitant to discuss this topic [18]. Moreover, according to reported sexual difficulties collected in the Global Study of Sexual Attitudes and Behaviors study 27,500 people (men and women aged 40–80 years) only around 19% (18.0% of men and 18.8% of women) sought medical help [19]. Sexual activity, sexual dysfunctions and related help-seeking behavior were investigated in adults (aged 40–80 years) using a telephone survey (5998 individuals; 2992 men and 3006 women) in 5 countries (United States, Canada, the United Kingdom, Australia, and New Zealand). Seventy-two percent (72%) of the respondents positive for dysfunction had no action in regards to it, and supposedly had a lack of perception or lack of discomfort about the problem. In addition, they did not know they needed to seek a doctor in such cases. Another cited cause was embarrassment (23%) [20]. Thus, this reflects lacking a favorable environment in both which men and women would feel not comfortable asking for information, and poses a challenge to physicians and other health professionals. This also indicates the importance of using questionnaires that have high precision resulting in accurate diagnoses and for effective monitoring treatments to counterbalance this gap [21, 22].

Sexual function has been evaluated by different surveys, but they may or not include domains of interaction with partners [21]. They also show a large variation in number of items that characterize each domain and they consequently vary in terms of psychometric properties. For example, FSFI is one of the most used questionnaires, but no domain related to the couple’s relationship is present. [6] On the other hand two others frequently used scales are the GRISS and SF-Q, with the latter being the most for the Brazilian population, covering issues where the quality of the couple’s relationships are addressed [7, 8]. Thus, the objective of this study was to compare the precision between scales which only quantify properties of female-specific sexual dysfunction and those which also score for quality of the couples’ relationship for diagnosing sexual dysfunction therefore directing the improvement of primary diagnosis of sexual disorders.

In regard to relationship status, 60% of the participants reported being single (n = 30), 30.0% married (n = 15) and five divorced (10.0%). Fifty-six percent (56%) identified themselves as white (n = 28), 28.0% brown and 16.0% black. In relation to life style, 35.4% (n = 17) of the sample were social drinkers, 100% (n = 48) non-smokers, 70.8% sedentary. There were 60.4% (n = 29) who did not use birth control pills and 20.8% (n = 10) had children.

Within the sample, 18.8% (n = 9) of the women exhibited sexual dysfunctions according to the FSFI. Analysis using the SQ-F detected 7 women (14.6%) with sexual dysfunction, whereas the GRISS Inventory also found that 18.2% (n = 8) of the women presented dysfunctions.

According to the results of this study, all three questionnaires detected sexual dysfunction or its absence. The results showed “excellent to good”’ domain discrimination for FSFI and GRISS, but “fair to poor” for SQ-F. However, agreement for paired-comparison between the three questionnaires showed a similar fair accuracy for the detected sexual response. These findings indicate that even those questionnaires that have an excellent to fair accuracy in most domains (FSFI and GRISS), and that which show “fair to poor” agreement in detected dysfunction (SQ-F) are fair in accordance. This might have occurred because some domains need to be reviewed in terms of their detection ability in these scales. For instance, the desire domain shows low reliability [21] on the FSFI questionnaire, as well as in the present study, where it showed a fair probability (low sensibility and specificity) when compared to its other domains which were excellent or good. A similar specificity grade (fair) for desire was also observed for SQ-F and was probably higher due to this similarity agreement between both questionnaires. Moreover, the proximity among the domains explored by them are higher than those of GRISS.

The domains explored by GRISS are more general and also included the feelings about the partners while FSFI has only specific female psychometric properties and does not explore a partnerships evaluation.

Although the prevalence of sexual dysfunction increases in middle-age and menopause, a significant percentage of sexual dysfunctions are also reported for young women. For instance, in using a sample of 201 sexually active Brazilian women aged 18 to 45 years, Prado, Mota & Lima [16] demonstrated that sexual dysfunction affected 30.2% of these individuals. Similarly, Ferreira, Souza & Amorim [17] found that 36% of women aged between 20 and 39 years exhibit at least one type of sexual dysfunction, 18% for orgasm and 13% for dyspareunia.

Around 18% of the sample of the present study showed sexual dysfunctions and the concordance between questionnaires was around 30%, but increases to 80% for characterizing typical sexual function. Therefore, the use of questionnaires to investigate sexual function in women should be re-evaluated and updated to improve their reliability for diagnoses if necessary.

In a study where the scales for diagnosing female sexual dysfunction were systematized, 27 available scales were listed, of which ten measure psychometric properties of specific domains of sexual dysfunction in women [20]. Of these, five presented reliability scores equal to or above 0.9 (Cronbach’s alpha), including FSFI. However, only the FSFI [6] contains items that contemplate the general domains to characterize the female sexual response included in the DSM-5 [12]. For example, one scale specifically addressed a cut-off point for the desire disorder domain in FSFI in order to achieve high sensitivity and specificity. Both indices increased from 75 and 84% to 92 and 89%, respectively. This scale (FSFI), as previously mentioned, has a moderate discriminative effect for this domain [28]. The other scales showed good scores for arousability, [29] impact of sexual dysfunction on women’s sexual life quality, [30] and to exclusively diagnose hypoactive sexual desire disorder (HSDD) [31] or HSDD due female distress [32]. From this evidence, application of the FSFI scale associated to a correction of the desire sexual domain is recommended [28].

The present study has limitations in estimating the sensitivity of inter-questionnaire agreement, likely due to the sample size and the fact that only young women were assessed. Also, data generalization is restricted since the sampling was performed by convenience. However, this study showed evidence of a need to improve the questionnaires that are presently being used to better characterize female sexual function Although more complete protocols have been proposed in relation to the main aspects of sexual function for constructing guidelines of the sexual history of men and women for health care professionals, [33] at the moment this protocol is not available in the public health service. Thus, the high reliability and adjustments introduced in the FSFI seem to provide a good instrument for its current use in the clinical diagnosis of the sexual health of women.

A comparison of accuracy to detect typical sexual function and sexual dysfunction was investigated using three widely used questionnaires, namely the FSFI, GRISS and SQ-F. Precision grades from “excellent to good” were detected for most FSFI and GRISS individual domains, and “fair to poor” accuracy for SQ-F. Despite these results, the comparison between each 2 scales using the ROC curve to test the accuracy in predicting female sexual response diagnoses, were similar, being measured as “fair”. The sensitivity and specificity between FSFI x SQ-F were 84.6 and 33.3%, respectively, whereas between FSFI and GRISS were 82.9 and 28.6% and GRISS x SQ-F 82.1 and 25.0%. The number of questions per domains of the scales ranged from one to four, and it was found that lower scores for a domain for the FSFI were also low for a similar domain in the SQ-F. In addition, these two questionnaires have high similar questions when compared to GRISS. As SQ-F and GRISS contain items related to the sexual partnership a better agreement between them was expected. However, they were the lowest which might be due to the low number of questions on the SQ-F to characterize similar domains. This study evidenced high agreement between scales to truth-positive symptoms and low agreement for truth-negative symptoms thereby conferring fair accuracy between the studied scales. The limited grade for discriminatory capacity (AUC) between surveys for sexual response must be considered when comparing results from these three different questionnaires and also to compare with other different scales. In addition, despite the diversity of scales, the high reliability and fit for their desire domain suggest that the FSFI scale has good accuracy for the current clinical assessment of women’s sexual health.