Psychometric characteristics

Reliability

The internal consistency reliability of the QRS – measured with Cronbach’s Alpha- was 0.8 for the total scale. Internal consistency of about 0.8 is generally accepted as good reliability for the measurement of intra-individual changes pre-/post therapy, for example. It remained consistently high if certain items were deleted from the analysis ⁹.

The internal consistency was also acceptable for the four subscales with one exception. The values were 0.90, 0.82, 0.75, and 0.57 for the subscales “psycho-somatic QoL”, “sexual activity”, “sexual dysfunction-self perception”, and “sexual dysfunction-partner’s view”, respectively.

The alpha of the latter subscale will be re-analysed when data from patients before/after therapeutic intervention become available. Altogether, the currently available reliability results are promising.

Further reliability testing is pending.  

Validity

In general terms, validation is a multi-step process. The first step of validation is usually to demonstrate a plausible internal structure through factor analysis. The four domains of the QRS fit theoretical expectations and thereby the conceptual framework; sexual function has a significant impact on quality of life and consists of satisfactory activity and good functioning of the two partners. Concluding from this consideration, the content and face-validity seems to be good.

Total versus domain correlation can also used to test validity. Ideally, there should be significant and high correlations between the total scale and all subscales forming the total scale. In contrast, the correlations among the individual subscales should be smaller, because the subscales are supposed to be “independent” according the factor-analytic model used. The correlations with the total scale were highest, as expected, ranging from 0.30 to 0.77. The correlations among subscales oscillate in the majority well under 0.2 and many were not significant despite the large sample size ⁹.

Cross-validation with other relevant scales was performed to see whether the total score, and the four domain scores measure what the name of the domain suggests (concurrent validity). An explorative validation survey was performed using a sample in Italy (n=47) and later confirmed in a population-based sample in Germany (n= 108).

The following scales were concurrently completed by all participating women in Italy and Germany under certain assumptions (sequence of the administration of the five scales changing among participants):

• The Menopause Rating scale (MRS) because it intends to measure HRQoL and complaints of women in the menopausal transition but some of the psychosomatic symptoms are also prevalent in younger women;
• The Hospital Anxiety and Depression Scale (HADS) because it is an established scale, particularly measuring depressive mood. There might be similarities with the psychosomatic domain of the QSF scale;
• The scale devoted to “Short-term hormonal effects (SHE) was used because this scale intends to measure HRQoL and contains many psychosomatic and complaints related to sexual and symptoms related to reproductive issues. The QSF scale has also domains devoted to psychosomatic and sexual problems.

The following observations were made (unpublished results):

Total score of QSF: The total summary score of the QSF scale shows a weak correlation with the total score of SF12 (both about r= -0.25). This applies also for Pearson’s correlation coefficients of total and depression score of the HADS (r= 0.24 and r= 0.35), the total score of the SHE scale (r= 0.34), and the total score of the MRS scale (r=0.32) as well as the domain scores of the MRS. The results were similar in the German survey. This seems indicative for a property of the total score of the QSF scale to measure QoL- at least to some extend.

Psychosomatic domain of QSF: This domain was correlated with SF 12 (r= -0.61), specifically the mental domain (r= -0.45), the total score of HADS and domain scores (ranging from r= 0.35 to r= 0.62), and also associated with the total score and psychological domain score of the SHE scale (r= 0.22 to r= 0.38), and the total as well as psychological and psychosomatic subscale of the MRS scale (r= 0.53 to r= 0.55). Obviously, the psychosomatic domain of the QSF has many similarities with relevant domains of the SF 12, HADS, SHE, and MRS. This confirms that the psychosomatic domain of the QSF measures psychological & somatic complaints and also HRQoL. Similar observations presented the German validation survey.

Sexual domains of QSF: We have not yet concurrently applied a scale that can be considered as external validation criterion for the 3 sexually related domains of the QSF scale, particularly not for the self-rating of sexual problems and perceived assessment by the sexual partner. However, we found a few weak correlations with the sexual domain of the SHE scale (ranging between r= 0.22 and r= 0.30). We found no correlation with the sexual (urogenital) domain of the MRS scale. This might be explained by very different content (more oriented towards urogenital than sexual issues in the MRS) and also the different characteristics of clients for the MRS, e.g. higher age. This was similar in the German survey. Thus, the validation of the sexual domains of the QSF with a proper external scale is needed as a next step.

Altogether, the QSF scale with its total and domain scores seem to be valid, although little information about the validity of the 3 sexual domains is currently available.

Ability to detect changes

Since the QSF scale was not yet applied in treatment-related observational or the randomized clinical studies, there are no data to describe responsiveness or MID.

---