Synopsis of the Derogatis Interview for Sexual FunctioningDISF/DISF-SR
Derogatis, Leonard R.
Derogatis Interview for Sexual Functioning (DISF and DISF-SR)
Clinical Psychometric Research Inc.,
1228 Wine Spring Lane,
Towson, MD 21204
Year Developed: 1987, 1989
Primary Measurement Constructs: Principal foundation constructs underlying effective sexual functioning.
The Derogatis Interview for Sexual Functioning (DISF) is a brief semi-structured interview designed to provide an estimate of the quality of an individual’s current sexual functioning in quantitative terms. The DISF represents quality of current sexual functioning in a multidomain format, which to some degree parallels the phases of the sexual response cycle (Masters & Johnson, 1966). The 25 interview items of the DISF are arranged into five domains of sexual functioning: I. Sexual Cognition /Fantasy, II. Sexual Arousal, III. Sexual Behavior/Experience, IV. Orgasm, V. Sexual Drive/Relationship. In addition, the DISF Total Score is computed, which summarizes quality of sexual functioning across the five primary DISF domains. The DISF interview requires between 15 to 20 minutes to complete, and there are distinct gender-keyed versions for men and women.
In addition to the DISF interview, there is a distinct self-report version of the test known as the DISF-SR. The DISF-SR is also composed of 26 items, and was designed to be as comparable to the DISF interview as possible. Time requirements for the DISF-SR are similar to the DISF; however, in most contexts the self-report version typically requires a few minutes less than the interview. Time requirements drop noticeably for both versions on successive administrations, such as in clinical efficacy or effectiveness trials, where the test is administered sequentially. With slight modifications in format, the DISF-SR may also be utilized to gain evaluations of the patient’s sexual performance by the patient’s spouse.
The DISF and DISF-SR were developed to address the unmet need for a set of brief, gender-keyed, multidimensional outcome measures that would represent the status of an individual’s current sexual functioning, and do so at multiple levels of interpretation. The DISF/DISF-SR are designed to be interpreted at three distinct levels: the discrete item level (e.g., “A full erection upon awakening”, “Your ability to have an orgasm”), the functional domain level (e.g., Sexual Arousal Score), and the global summary level (e.g., DISF/DISF-SR Total Score). Since the DISF interview and the DISF-SR self-report inventory are matched on an almost item-for-item basis, clinician and patient assessments of the patient’s quality of sexual functioning may be obtained in both raw and standardized score formats. Both instruments may be used repeatedly throughout efficacy or effectiveness trials, or may
be implemented solely at pre- and post-intervention without significant “practice” effects or loss of validity.
Norms and Standardized Scores
Norms have been developed for both the DISF and the DISF/SR, based in each case, on several hundred nonpatient community respondents. The norms are gender-keyed (i.e., separate norms for men and women), and are represented as standardized scores in terms of Area T-scores. The Area standardized score possesses distinct advantages over the simple Linear transformation, in that the former provides accurate percentile equivalents (i.e., T-score of 30= 2nd centile; T-score of 40= 16th centile; T-score of 50= 50th centile; T-score of 60= 84th centile; T-score of 70= 98th centile, etc.). This important characteristic is not true of Linear T-scores except when the underlying raw score distribution is perfectly normal. In addition to enabling accurate comparisons across respondents, Area T-scores also facilitate meaningful comparisons of strengths and weaknesses within a respondent’s profile of sexual functioning. A patient may reveal a relatively unremarkable profile with the exception of a profound decrement in a single functional domain, or may show a low grade degradation of performance across multiple areas of functioning. Because DISF/DISF-SR domain scores are available in an equivalent standardized metric, such evaluations can help pinpoint the nature and extent of sexual dysfunctions.
Clinical/Research Instrument, Interview & Self-Report
English, French, German, Finnish, Polish
The DISF/DISF-SR are each comprised of 25 items. In the case of the former, items are cast in the format of a semi-structured interview structured via 4-point Likert scales. The items of the DISF-SR are also represented a 4-point Likert scales, and are designed to the extent possible, to match the items of the DISF.
Reliability and Validity Studies
The DISF and DISF-SR have both demonstrated favorable profiles of psychometric characteristics (Tables 1-4). Table 1 provides a summary of several reliability studies on the two tests which show both versions of the DISF to be highly reliable. The interrater
reliability coefficients shown in Table 1. for the DISF were developed during preliminary training sessions for a large, multicenter drug trial in which 16 distinct clinician/ratersparticipated. These data reveal DISF domain coefficients ranged from a low of .84 for
Orgasm to a high of .92 for Sexual Cognition/Fantasy. The Interrater coefficient for theDISF Total Score was a highly satisfactory .91. Concerning the DISF-SR, both internalconsistency and test-retest reliabilities were developed from separate subgroups of the
normative sample. Coefficients _ ranged from a low of .74 for Sexual Drive/Relationship toa high of .80 for Orgasm, very acceptable values for internal consistency estimates. Similarly, test-retest reliability coefficients (based on a 1 week interval) were also good,
ranging from .80 to .90, with the stability coefficient for the DISF-SR Total Score being .86.
An important validity demonstration for multidimensional or multidomain psychological outcome measures, concerns the subtest intercorrelation matrix, and domain score-total score correlation vector. The pattern of these correlations represents a central
psychometric characteristic of the test which relates to almost all discernable aspects of construct validity (Messick, 1995). If correlations between dimension scores are high, concerns may be raised that operational definitions of the domain constructs are redundant. If domain scores do not correlate at least moderately with the Total score, then the possibility exists that the domain constructs (e.g., Sexual Arousal, Orgasm) as operationally defined by the test items, are not valid components of the higher-order, more general construct (e.g., Quality of Sexual Functioning). A theoretical optimum would find correlations between domains near zero, with each domain score showing a moderately high correlation with Total score. In such an ideal design, each domain would contribute independent true variance to the Total Score, with minimum redundancy or overlap.
Subtest intercorrelation matrices are presented for the DISF/DISF-SR based on two normal and one sexually dysfunctional sample in Tables 2-4. As is obvious from these data, in all cases the mean interdomain correlation coefficients are relatively low (i.e., .23 to .39), while the average domain-total correlations for the three samples range from .60 to .71). This pattern of subtest correlations begins to approach optimal, and strongly confirms that DISF domains are contributing relatively independent variance to the DISF Total Score.
The DISF/DISF-SR have been introduced only recently, and much of the clinical research done with the tests has primarily involved corporate-sponsored clinical drug trials. Although preliminary data from these studies indicate that the tests are highly sensitive to sexual dysfunction, and to a broad range of therapeutic agents, most of the data are proprietary and have not yet been made generally available. In two studies that have been published, (Zinreich, Derogatis, Herpst, et.al., 1990a; 1990b) the DISF was utilized with males suffering from prostate cancer about to undergo a course of radiation therapy. At time of initial cancer diagnosis, the DISF was utilized in a logistic regression model as a screen for impotence, with detailed clinical evaluation as the ultimate criterion. In this study, sensitivity was found to be 86%, specificity was 80%, and the predictive value of a positive was 86%. Subsequent to treatment, patients were assigned to three functional categories on the basis of clinical evaluation: a.)totally functional, b.)marginally functional, and c.)impotent. Scores on the five domains of the DSFI were significantly different across the three groups, with mean DISF Total Scores being 48.2, 21.5 and 14.0 respectively. In this study, with a complex sample of patients, the DISF did a superior job of identifying those individuals who were dysfunctional prior to treatment, and validly reflected differences in quality of sexual functioning subsequent to therapeutic intervention.
Currently, other studies utilizing the DISF/DISF-SR are in the process of submission and review for publication, and several new norms (e.g. geriatric, gay men) are in the process of being developed.
How to Obtain
The DISF and DISF-SR are co-distributed by Clinical Psychometric Research, Inc., 1228 Wine Spring Lane, Towson, MD 21204, Phone 410-321-6165 And MAPI Research Trust 27, rue de la Villette 6900 Lyon, France, eprovide.MAPI-trust.org
Derogatis Measurement Assessments, LLC
Derogatis, L.R. (1996) Derogatis Interview For Sexual Functioning (DISF/DISF-SR): Preliminary Scoring, Procedures & Administration Manual. Baltimore, MD, Clinical Psychometric Research.
Zinreich, E. Derogatis, L.R., Herpst, J., Auvil, G., Piantodosi, S. & Order, S.E. (1990).
Pretreatment evaluation of sexual function in patients with adenocarcinoma of the prostate.
International Journal of Radiation Oncology & Biological Physics,19, 1001-1004.
Zinreich, E. Derogatis, L.R., Herpst, J., Auvil, G., Piantodosi, S. & Order, S.E. (1990). Pre and posttreatment evaluation of sexual function in patients with adenocarcinoma /of the prostate. International Journal of Radiation Oncology & Biological Physics, 19, 729-732.