What is the "true" prevalence of female sexual dysfunctions and does the way we assess these conditions have an impact?

<b>Introduction.</b> A wide range of prevalence estimates of female sexual dysfunctions (FSD) have been reported.<br /><b>Aim</b>. Compare instruments used to assess FSD to determine if differences between instruments contribute to variation in reported prevalence.<br /><b>Main Outcome Measures. </b>Sexual Function Questionnaire combined with Female Sexual Distress Scale (SFQ-FSDS) was our gold standard, validated instrument for assessing FSD. Alternatives were SFQ alone and two sets of simple questions adapted from Laumann et al. 1994.<br /><b>Methods.</b> A postal survey was administered to a random sample of 356 Australian women aged 20 to 70 years.<br /><b>Results.</b> When assessed by SFQ-FSDS, prevalence estimates (95% confidence intervals) of hypoactive sexual desire disorder, sexual arousal disorder (lubrication), orgasmic disorder, and dyspareunia were 16% (12% to 20%), 7% (5% to 11%), 8% (6% to 12%), and 1% (0.5% to 3%), respectively. Prevalence estimates varied across alternative instruments for these disorders: 32% to 58%, 16% to 32%, 16% to 33%, and 3% to 23%, respectively. Compared with SFQ-FSDS alternative instruments produced higher estimates of desire, arousal and orgasm disorders and displayed a range of sensitivities (0.25 to 1.0), specificities (0.48 to 0.99), positive predictive values (0.01 to 0.56), and negative predictive values (0.95 to 1.0) across the disorders investigated. Kappa statistics comparing SFQ-FSDS and alternative instruments ranged from 0 to 0.71 but were predominantly 0.44 or less. Changing recall from previous month to 1 month or more in the previous year produced higher estimates for all disorders investigated. Including sexual distress produced lower estimates for desire, arousal, and orgasm disorders.<br /><b>Conclusions</b>. Prevalence estimates of FSD varied substantially across instruments. Relatively low positive predictive values and kappa statistics combined with a broad range of sensitivities and specificities indicated that different instruments identified different subgroups. Consequently, the instruments researchers choose when assessing FSD may affect prevalence estimates and risk factors they report.<br />