Cognitive-behavioral Bibliotherapy for Sexual Dysfunctions in Heterosexual Couples: A Randomized Waiting-list Controlled Clinical Trial in the Netherlands.

Since the innovative work of Masters and Johnson (1970), it has become clear that in many cases sexual problems can be treated successfully by means of behaviorally oriented psychotherapy (Hawton, 1995; Rosen & Leiblum, 1995). Although the enthousiasm to treat sexual dysfunctions has increased among professional helpers, the empirical basis of the effectiveness of sex therapy is still limited. Heiman and Meston (1997) concluded in a recent review that only the treatments of primary female orgasmic disorder and of male erectile disorder meet the quality standards for effective psychological treatment as proposed by the American Psychological Association's (1995) task force. Parallel to effect studies of sex therapy, dismantling studies were conducted in which the contribution of different aspects of sex therapy to its effects were examined (Hawton, 1995; Heiman & Meston, 1997; Rosen & Leiblum, 1995). Through the use of bibliotherapy the relative contribution of the nonspecific factor of therapist attention in sex therapy has been investigated. Bibliotherapy conveys the therapeutic interventions without the physical presence of a therapist. If bibliotherapy is found to produce significant and clinically relevant effects, the effects of sex therapy can more confidently be attributed to the contents of its constituent elements.

In a recent search of the literature (van Lankveld, 1998), 12 controlled studies published between 1975 and 1988, including 3 unpublished doctoral dissertations, were retrieved. In a meta-analysis of these studies an average unweighted effect size of 0.68 standard deviations measured at posttreatment was found, which is a large effect size according to Wolf's (1986) criteria. Weighted for sample size, an average effect size of 0.50 was found, which is still an effect of medium strength. Mean effect size at follow-up, however, was considerably lower (Glass' [Delta] = 0.30, with its 95% confidence interval, -0.10 to 0.72, containing zero, meaning possible deterioration).

The existing bibliotherapy literature leaves a number of questions unanswered. Bibliotherapy for sexual dysfunctions has thus far reflected the directed practice approach of Masters and Johnson (1970) or modifications of it. Cognitive-behavioral therapy has not been examined in a bibliotherapy format for sexual dysfunctions, although Everaerd and Dekker (1985) and Munjack et al. (1984) reported significant positive results of rational-emotive therapy for erectile dysfunction of men with steady relationships. The first aim of the present study is to test the hypothesis that the results of cognitive-behavioral bibliotherapy with minimal therapist support are superior to a waiting-list control condition.

Second, the basic efficacy of this approach has not yet been established for the majority of the sexual dysfunctions, since 87% of the studies to date (van Lankveld, 1998) have dealt with orgasmic disorders. Some dysfunctions, such as hypoactive sexual desire, sexual aversion, erectile disorder, female arousal disorder, and dyspareunia, have not been studied at all or have only been included in a study of mixed target problems (e.g., vaginismus: Dow, 1983). It is conceivable that differential effects of bibliotherapy may be found in these types of sexual dysfunction. Thus, the second question to be answered by the present study is whether or not cognitive-behavioral bibliotherapy with minimal therapist support has different effects on males and females with different sexual dysfunctions.

Third, the "intention-to-treat" issue has not been addressed in previous research. For future community-based use of a bibliotherapy approach, it is important to know how reliable the results of this approach are in the face of participant drop out. The third question, then, is whether the results of cognitive-behavioral bibliotherapy with minimal therapist support are sustained when adjusted for couples not completing the research project.

Fourth, the number of studies that were retrievable to date appears to be too small to detect significant differences between different bibliotherapy features and methodological characteristics (van Lankveld, 1998). The most prominent difference, though not statistically significant, was between totally self-administered bibliotherapy and minimal therapist contact bibliotherapy, the latter being associated with the larger effect size. Moreover, the role of compliance to the requirements of the bibliotherapy approach as a moderator of treatment effect has to be established. This concerns issues such as the amount of time spent in reading the bibliotherapy manual and the number of times that prescribed exercises are performed. The fourth aim of this study is to test the hypothesis that more frequent telephone contact with a therapist and a higher rate of other compliance aspects are associated with better treatment results.

METHODS

Participants

Inclusion criteria for all participants were heterosexual couples with both partners over 16 years of age seeking help for a sexual dysfunction of at least one partner. Dysfunctions were required to meet the DSM-IV criteria, including the absence of major organic causes for sexual dysfunction and of medication effects. Exclusion criteria were current psychotic disorder, major depression or abuse of alcohol or psychoactive drugs by either partner, imminent divorce, or concurrent psychological or psychiatric treatment for sexual, psychological, or marital problems, including psychopharmacological treatment.

Couples were recruited from two different sources. Forty percent were patient couples from the outpatient clinic for Psychosomatic Gynecology and Sexology of the Leiden University Medical Centre, the Netherlands. Couples were referred by general practitioners (45%), gynecologists and urologists (40%), and mental health professionals (15%). When couples were placed on the waiting list for professional treatment they received a letter of invitation to join the study along with a leaflet introducing the research procedure. Couples who wanted to participate contacted the research assistant and scheduled appointments. Sixty percent of the participating couples were recruited by advertisements in local and national Dutch newspapers. Couples responded to these advertisements by calling the aforementioned outpatient clinic from which they received information about the study by telephone. Approval of the hospital's Medical Ethics Committee was obtained for the research design and procedure.

Participant groups from both recruitment sources were collapsed. Multivariate comparisons revealed that variability of demographic characteristics and baseline scores on the outcome variables was associated with the different types of sexual dysfunction, not with recruitment route (van Lankveld, Grotjohann, van Lokven, & Everaerd, 1999).

Study Design

The study was designed as a randomized waiting-list controlled clinical trial. Participants were randomly assigned to 10 weeks of treatment with cognitive behavioral bibliotherapy and minimal therapist support by telephone followed by a 10-week follow-up period, or to a waiting-list control group. This duration of treatment and follow-up period equals the average treatment period length of previous bibliotherapy studies of sexual dysfunctions (van Lankveld, 1998). It was considered ethical to include a waiting-list control condition, because it is unknown whether cognitive-behavioral bibliotherapy is effective for sexual dysfunctions, and because participants in the control group received the bibliotherapy after the waiting period and ultimately, if deemed necessary, sex therapy. The results of our experimental intervention may, in theory, be the result of nonspecific factors of receiving any treatment at all, constituting a threat to the internal validity of the study. This concern would be more appropriately addressed by employing a placebo control condition, which we have chosen, however, not to include. We considered this to be unethical, especially for the media-recruited participants. Randomization was performed by card drawing at the end of the initial assessment session. The unit of randomization was the couple. The method of block randomization was employed, with blocks of 10 cards, evenly divided in experimental and control group cards. A factorial design was employed with two between-groups comparisons, testing the main effects of treatment (cognitive behavioral bibliotherapy with telephone support vs. waiting-list control) and of sexual dysfunction type, as well as the interaction effect of treatment and dysfunction type. In clinical practice, many patients presenting different sexual dysfunction types nevertheless report impairment of several aspects of sexual functioning, including sexual desire, arousal, pain/discomfort, and orgasm. For this reason the same aspects of sexual, marital, and psychological functioning were evaluated in all participants. Rather than employing specific outcome parameters for different sexual dysfuntions, this also enables comparison of bibliotherapy effects across dysfunction types. Separate analyses were performed for male and female sexual dysfunctions. For a within-subjects comparison, assessment data of outcome variables at three points in time were used: (a) pretreatment baseline, (b) posttreatment after 10 weeks, and (c) follow-up after another 10 weeks.

Procedure

All data in the study were collected by a female research assistant, who is a trained psychologist. The research assistant was not involved in the design of the study, nor in the delivery of support to participants by telephone. Data collection, participant support, and data evaluation were kept separate to minimize experimenter bias. The initial assessment session lasted from 90 to 180 minutes. Partners were seen separately, while the other partner completed self-report questionnaires in an adjacent room. The research assistant conducted a structured interview to collect demographic, marital, and sexual histories, and treatment history data. Separately, for each partner, self-presented sexual dysfunctions were classified according to DSM-IV (APA, 1994). In case more than one sexual problem was presented, participants were asked to prioritize them. The highest priority dysfunction was used for further classification. No distinction was made between lifelong and acquired dysfunctions, nor between global and situational dysfunctions. Next, the Composite International Diagnostic Interview (CIDI: World Health Organization, 1990) was conducted. The variation of duration of the initial interview was largely due to the time needed to complete this part of the assessment. The CIDI is a structured diagnostic interview to evaluate the criteria for mental disorders, according to DSM-III-R (APA, 1987), Axis I. It enables data collection with simultaneous computer scoring. At the moment of data collection, no such instrument was available which used DSM-IV criteria. The interview is divided into 15 sections: demographics (A), disorders resulting from the use of tobacco (B), somatoform and dissociative disorders (C), phobic and other anxiety disorders (D), depressive and dysthymic disorders (E), manic and bipolar affective disorders (F), schizophrenia and other psychotic disorders (G), eating disorders (H), disorders resulting from the use of alcohol (I), obsessive-compulsive disorder (K), disorders resulting from the use of psychoactive substances (L), organic mental disorders (M), sexual dysfunction (N), interview observations (P), and interview ratings (X). Section I was conducted only if subjects scored 2 or higher on the CAGE, a four-question interview, indicative of the presence of alcohol abuse and alcohol dependence (Ewing, 1984). The sections concerning tobacco use (B), psychoactive drug abuse (L), and organically based mental disorder (M) were omitted from the interview. Although in the literature a rather consistent influence is found of tobacco smoking on vasculogenic erectile dysfunction (Condra, Morales, Owen, Surridge, & Fenemore, 1986; Hirshkowitz, Karacan, Howell, Arcasoy, & Williams, 1992), other dysfunctions have not been associated with tobacco abuse (Abel, 1985; Gilbert, Hagen, & D'Agostino, 1986). Abuse of psychoactive drugs and organic mental disorder were also found to be related to sexual dysfunction (Abel, 1985; Segraves, 1989). However, we did not expect these disorders to occur in our samples frequently enough to be of any influence. We placed confidence in the initial screening procedure by telephone in the volunteer subsample and the clinical interviewing and medical examination in the patient subsample to have effectively filtered out the individuals with. these disorders. The CIDI procedure allows the diagnosis of multiple mental disorders. Subjects can be diagnosed with a current diagnosis (disorder present within last 2 weeks), or a lifetime diagnosis (disorder having been present earlier in life). Interrater reliability (K = 1.00) and convergent validity of the CIDI, compared with the Schedules for Clinical Assessment in Neuropsychiatry (SCAN), were found to be satisfactory for current (r = 0.69) and for lifetime (r = 0.66) diagnoses of depression and anxiety disorders (Andrews, Peters, Guzman, & Bird, 1995), although another study found the reliability for phobias, social phobia, and agoraphobia to be less satisfactory (K-range from 0.46 to 0.63: Wittchen, Zhao, Abelson, Abelson, & Kessler, 1996).

Participants meeting any of the exclusion criteria were dropped from the study and, in case of entrance via an advertisement, given guidance to receive adequate professional help.

Couples in the experimental group received the bibliotherapy manual after payment of 25 Dutch guilders (US $12). They were given a verbal explanation of how to use the manual, and were encouraged to call a clinic staff member for support if any questions arose or if any difficulties were encountered. The research assistant introduced the staff member to the couple at the end of their visit, and a direct telephone number for use during office hours was given. At the end of the first assessment session new appointments were scheduled for posttreatment and follow-up assessment. Assessment at the end of the 10-week treatment period and at the end of the subsequent 10-week follow-up period took place in a similar manner. One partner was interviewed by the research assistant while the other partner completed questionnaires assessing outcome variables. In this interview, data regarding treatment compliance were collected.

The last 21 couples who entered the study in response to advertisements were not randomized, but assigned to another treatment regime. This was done to enable comparison of the effects of bibliotherapy with more frequent and scheduled telephone contacts on the one hand, and the same treatment with entirely self-initiated telephone support on the other hand (Gould & Clum, 1993). At the end of the initial assessment session, three appointments for telephone contact were scheduled during the 10-week treatment period, each separated by 3 weeks. The first appointment was after 1 week, at which one or both partners discussed their progress with the staff member. They were encouraged to prepare for the call by listing their questions and other topics for discussion. Self-initiated contact was encouraged in addition to the scheduled telephone contacts. In the scheduled-contact experimental group, the number of telephone contacts ranged from 0 to 6 (M = 2.5, SD = 2.0), and total duration of telephone contact ranged from 0 to 113 minutes (M = 32 min, SD = 35). In the own-initiative experimental group the number of telephone contacts ranged from 0 to 2 (M = 0.14, SD = 0.4), and total duration of telephone contact ranged from 0 to 20 minutes (M = 1.4 min, SD = 4.2). The contents of the bibliotherapy manual and the assessment procedures for the scheduled-contact group were identical to those for the other experimental group. Although number of telephone contacts (t = -5.26, df = 19.4, p [is less than or equal to] .001) and duration of contact (t = -3.93, df = 19.1, p [is less than or equal to] .001) differed significantly, preliminary analyses, however, revealed no univariate or multivariate differences with respect to posttreatment and follow-up outcome variables between the scheduled-contact group and the own-initiative group. The two groups were collapsed for further analyses, resulting in a larger experimental group.

Instruments

Four paper-and-pencil instruments were employed and are described below. A posttreatment session collected additional information.

Golombok Rust Inventory of Sexual Satisfaction (GRISS). This self-report questionnaire has separate forms for men and women, each consisting of 28 items. It measures the most common psychosexual complaints and was chosen to assess the degree of sexual dysfunction. Three main dimensions of sexual functioning in both partners were selected: male subscales measuring sexual infrequency (scoring range: 2-10), erectile dysfunction (scoring range: 4-20), and premature ejaculation (scoring range: 4-20), and female subscales measuring sexual infrequency (scoring range: 2-10), anorgasmia (scoring range: 4-20), and vaginismus (scoring range: 4-20). The vaginismus subscale also measures vaginal discomfort, as experienced by women with complaints of dyspareunia. Although complaints of infrequency of sexual contact do not completely cover the concept of hypoactive sexual desire disorder (HSDD), the GRISS subscale of infrequency was nevertheless chosen to represent problems with low sexual desire. Complaints of infrequency of both partners are, from our clinical experience, considered to adequately reflect the often occurring problem situation in couples with one partner with HSDD. The GRISS contains no scale of male sexual pain. However, our present study sample was found not to comprise male participants with this sexual disorder. Higher scores on all subscales indicate more dissatisfaction. Reliability (test-retest correlations for subscales range from 0.63 to 0.94) and construct validity of the GRISS have been found to be satisfactory in both English (Rust & Golombok, 1986) and Dutch samples (Ter Kuile, van Lankveld, Kalkhoven, & van Egmond, 1999). The subscales were found to be independent of social desirability (van Lankveld & Ter Kuile, 1999). Low correlations were found with the Lie-Scale of the Eysenck Personality Questionnaire Revised, Short Version (0.01 [is less than] r [is less than] 0.10).

Intiem Lichamelijk Kontakt Schalen (ILKS) (Intimate Bodily Contact Scales). This self-report instrument of Dutch origin measures the subjective meaning of sexuality on perceptual, behavioral, and evaluative levels. It employs a broad operationalization of sexual experience, conceiving of sexual dissatisfaction as the discrepancy between observed and expected behavior and experience during intimate physical and sexual contact. The following scales were selected for this study: perceived intimacy (19 items) and perceived arousal (7 items) during psysical interaction with the partner, and self-perception as sexual partner (7 items). Items are constructed as questions (e.g. "While making love with your partner, lately, do you feel spiritually attracted to him?"). Answering categories were 5-point Likert scales with item-specific categories indicating frequencies (never to always) or degree of experience (low to high). Some items provide an of no relevance answering category. The possible scoring range is 0 to 100. High scores on the intimacy and arousal subscales represent low perceived intimacy and perceived arousal. High self-perception scores indicate a positive self-perception as sexual partner. The reliability (Cronbach's [Alpha]: 0.65 [is less than] [Alpha] [is less than] 0.93) and validity of the ILKS have been found to be satisfactory in Dutch samples (Vennix, 1983).

Maudsley Marital Questionnaire (MMQ). From this originally English self-report questionnaire the subscales measuring dissatisfaction with the relationship in general (scoring range: 0-80) and with the sexual relationship (scoring range: 0-40) were selected. High scores indicate more dissatisfaction. Test-retest reliability (0.62 [is less than] r [is less than] 0.86) and construct validity (e.g., -0.38 [is less than] r [is less than] -0.64; p [is less than or equal to] .001 for the marital satisfaction subscale with the Marital Deprivation Scale) of the MMQ have been found to be satisfactory in Dutch samples (Arrindell, Boelens, & Lambert, 1983).

Self-rating Scale for the Evaluation of Treatment. A four-item self-report questionnaire was constructed for this study to assess client's ratings of change in problem status (problem status), and of distress associated with the sexual problem (problem distress). Items were constructed as 7-point Likert scales with item-specific answering categories. Answering categories were given for the extremes of each scale, and for the middle category. The intermediate categories (2, 3, 5, 6) did not have labels, only numbered compartments in which to place marks. Problem status was rated with one question: "To what extent has your sexual problem improved or worsened during the last 4 weeks?" (1 = gotten much worse, 4 = not gotten better or worse, 7 = gotten much better). Problem-associated distress was assessed by summing up the answer scores to three questions: "How much trouble has your sexual problem given you during the last 4 weeks?" (1 = a lot of trouble, 4 = some trouble, 7 = no trouble at all); "How much did your sexual problem interfere with your daily life in the last 4 weeks?" (1 = a lot, 4 = some, 7 = not at all); "How much cause for concern did your sexual problem give you during the last 4 weeks?" (1 = a lot of concern, 4 = some concern, 7 = no concern at all). Using a sum score was considered justified by the magnitude of bivariate correlations of these three items at the three assessment points for both males (correlation range: 0.48 to 0.77) and females (correlation range: 0.46 to 0.72). High scores represent, respectively, more improvement and lower problem associated distress.

Compliance with treatment. In the posttreatment assessment session information was collected on the rate of compliance with the bibliotherapy method. The research assistant asked each partner individually to estimate the total time she or he had spent in reading the manual, the number of completed rational-emotive self-analyses, the number of different exercises performed as suggested by the manual, and the total number of times that exercises were performed. Questions were posed to each partner with respect to self-performed activity, as well as with respect to activity of the partner, thus providing a check on socially desirable answers. The number of telephone contacts and total duration of therapist contact by telephone were recorded for couples in the bibliotherapy group. Subjects in this group were asked to rate the perceived effort they and their partner invested in solving the sexual problem by means of the self-help method on an 11-point scale, ranging from 0 to 10, with 10 designating maximum effort.

TREATMENT

Treatment consisted of a minimal interventions approach. Participants received a bibliotherapy manual with a short verbal explanation by the research assistant of how to use the manual. The manual used is a book of 266 pages (van Lankveld, 1993). It presents a cognitive-behavioral strategy of self-help for sexual dysfunctions. In three introductory chapters the self-help strategy and the method of rational-emotive self-analysis (Maultsby, 1975) is explained. Depending on the constellation of sexual dysfunctions of both partners, other chapters were selected, focusing on distinct male and female dysfunctions in the context of masturbation and of sexual interaction. The last chapter is devoted to communication problems. Chapters commence with an introduction to the specific sexual dysfunction. Information is provided on biological and psychological factors and mechanisms involved in the development and maintenance of the dysfunction. The manual thus provides psychoeducational information, aimed at correcting lack of sexual knowledge and faulty or dysfunctional attitudes. Next, a step-by-step program of individual and partner exercises is delineated. The exercises generally follow the instructions of the sensate focus approach as described by Masters and Johnson (1970). Additional exercises are directed at, for instance, self-exploration of body and genitals, exploration of potentially arousing sexual stimuli, improvement of concentration on physical sensations, and exploration of sexual fantasy (c.f. Zilbergeld, 1993). Intermediate steps for sensate focus exercises are outlined in order to reduce the magnitude of steps and to enable gradual approximation of goals. For every dysfunction type, specific exercises are described, such as the "teasing exercise" for erectile disorder (Masters & Johnson, 1970), the "start-stop exercise" for premature ejaculation (Zilbergeld, 1993), or the finger exploration of the vagina in vaginismus. Throughout the chapters, rational-emotive self-analysis is employed as a trouble-shooting strategy (Maultsby, 1975). The partner who experiences problems during an exercise or in surrendering him/herself to an exercise is encouraged to use this method. An introductory chapter explains how to describe this experience and how to differentiate between cognitions, feelings, and behavior in the perceived situation. When a clear description of these aspects of the experience has been arrived at, new cognitions are generated, primarily aimed at the creation of better conditions for exercise. Imagery exercises, in which the newly formulated cognitions are applied, serve as stepping stones to resume sensate focus exercises. We intended to investigate the application of cognitive-behavioral bibliotherapy for sexual problems under conditions resembling the real-life application of the self-help manual as closely as possible. We have therefore chosen to keep the instruction for the use of the manual by the research assistant very limited, addressing three major points:

1. Participants should closely read the three introductory chapters of the manual in any case. These chapters explain the strategy used in the manual and the use of rational-emotive therapy as a trouble-shooting method when participants experience trouble in following the instructions for the sensate focus exercises.

2. They are encouraged to select for further reading the chapters in the rest of the manual which they find most appropriate in their personal problem situation, as they would do in the everyday self-help situation of people buying a self-help book and trying to apply it to their situation without any professional guidance.

3. Professional assistance by telephone by a trained sexologist (author #1) was offered to be made use of by participants when they experienced difficulties in their problem-solving efforts.

Statistical Analyses

All statistical analyses were performed using the Statistical Package for the Social Sciences 6.1 (SPSS: Norusis, 1994). Descriptive univariate statistics were calculated for demographic and outcome variables. The distributions of male and female sexual dysfunction types in the experimental group and in the control group were compared by means of Pearson's [chi square]. Further comparisons were multivariate and examined the influence of the independent variables (IV: treatment and dysfunction type) on the outcome variables. Before performing these analyses the data sets were screened for missing data, multivariate outliers, and other assumptions for multivariate analysis with frequencies, multiple linear regression, and scatter-plot. Multivariate outliers were deleted when their Mahalanobis distance proved greater than the critical value at [Alpha] = 0.001 after performance of multiple linear regression. No threats to the assumptions for multivariate analysis were found, unless reported at the appropriate place in the Results Section.

MANOVAs and MANCOVAs were employed with sequential adjustment for nonorthogonality. Order of entry of the independent variables was treatment, then dysfunction type. Levels of the treatment IV were bibliotherapy and waiting list. Levels of the dysfunction type IV for males were no sexual dysfunction, hypoactive sexual desire disorder, erectile disorder, and premature ejaculation. Levels of the dysfunction type IV for females were no sexual dysfunction, hypoactive sexual desire disorder, vaginismus, dyspareunia, and anorgasmia. Wilks' criterion was used to test the significance of main effects and interaction effects. For all reported effects, level of significance as well as effect size were calculated. Effect size was calculated as Glass' [Delta] (Glass, McGaw, & Smith, 1981), which standardizes the observed differences between experimental and control group means by the control group's standard deviation. The use of this effect size was chosen to enable comparison of the results of the present study with those of a meta-analysis of bibliotherapy for sexual dysfunctions (van Lankveld, 1998). To arrive at A, the [[Eta].sup.2] statistic was used, which represents the proportion of explained variance of the separate main and interaction ANOVA effects. First, Cohen's F was calculated by F = [([[Eta].sup.2]/1 - [[Eta].sup.2]).sup.1/2]. Subsequently, Glass' [Delta] is calculated by [Delta] = 2 x F (Cohen, 1988, p. 284). This results in an exact effect size for factors with two levels, and a conservative effects size estimator for factors with more than two levels.

Baseline levels at pretreatment of the outcome variables for the treatment groups and dysfunction types were analyzed with separate two-way MANOVAs for males and females. In case of pretreatment differences despite randomization, in the treatment factor or the interaction of treatment and dysfunction type, posttreatment and follow-up observations were adjusted for baseline scores.

Dependent upon the finding of pretreatment differences, separate two-way MANOVAs or MANCOVAs were performed on male and female outcome variables at posttreatment and follow-up, examining the influence of treatment and dysfunction type on sexual functioning variables.

Separate two-way MANCOVAs were performed on male and female outcome variables at posttreatment and follow-up, regarding participant evaluation of treatment, sexual and marital satisfaction, self-perception as sexual partner, and perceived intimacy and arousal during sexual interaction. Adjustment was made for sexual functioning variables by analysis of covariance, to separate the effects on this set of outcome variables from those on sexual functioning.

An intention-to-treat analysis was performed. Demographic and other personal characteristics, as well as pretreatment levels of outcome variables of couples lost at posttreatment and of follow-through couples, were compared. A similar comparison, supplemented with posttreatment outcome variables, was made of follow-through couples and couples lost between posttreatment and follow-up.

Finally, the associations with outcome of several aspects of compliance with treatment were analyzed by means of oneway MANOVAs.

RESULTS

Participant Flow

Figure 1 shows the flow of participants in this study. Two hundred and four couples from the outpatient clinic's waiting list were invited to participate. Some couples indicated their reasons for not participating. Most heard were lack of confidence in the potential value of the bibliotherapy method and the preference to wait for therapist-administered treatment. After refusal to participate and loss through initial assessment, 90 participating couples were randomized. Of 141 couples who applied for participation in response to media advertisements, 112 couples were randomized. Together, 202 couples were randomized. Twenty-one couples were added to the experimental group without randomization, and received bibliotherapy with scheduled-contact telephone support as described in the Methods section. Thus, the experimental group consisted of 125 couples, while the waiting list control group consisted of 98 couples. At posttreatment, 111 of these couples (89%) in the experimental group and 88 couples (90%) in the control group, were assessed. At follow-up, 100 couples (80%) in the experimental group, and 85 couples (87%) in the control group, were retained for assessment.

[ILLUSTRATION OMITTED]

Preliminary Analyses

Characteristics of 199 couples who followed through the study until posttreatment assessment were examined. Predominantly Caucasian couples, with both partners born in the Netherlands, participated in the study. Three percent of the males and 11% of the females were natives of other countries. Mean age of male participants was 39 years (SD = 11). Mean age of female participants was 36 years (SD = 11). Mean age of participants in our study was found to be 5 to 6 years higher than found in previous bibliotherapy studies of sexual dysfunction (van Lankveld, 1998) as well as in older outcome studies of therapist-administered sex therapy (De Amicis, Goldberg, LoPiccolo, Friedman, & Davies, 1984; Hawton & Catalan, 1986). A recent field study of therapist-administered sex therapy (Sarwer & Durlak, 1997), however, investigated a large sample of somewhat higher age. Average length of relationship was 14 years (SD = 11), which is equivalent to that found in the above-mentioned studies. Eighty-nine percent of the couples were living together and 65% of the couples were married. Forty-seven percent of the males and 44% of the females had no religious affiliation. Twenty-six percent of the males and 24% of the females were Protestant. Twenty-one percent of the males and 25% of the females were Roman Catholic. Six percent of the males and 7% of the females had other religious affiliations. Average duration of the sexual dysfunction was 8 years (SD = 8) for males, 8 years (SD = 7) for females. In previous bibliotherapy studies of sexual dysfunction (van Lankveld, 1998), mean problem duration was found to be 9.2 years (SD = 3.7 years). Demographic and sexual history characteristics of both treatment groups are shown in Table 1. No univariate differences on these characteristics were found between treatment groups. Oneway MANOVA treating the dichotomous dependent variables (DVs) as interval scaled variables and with simultaneous entry of all DVs showed no multivariate differences.

Table 1. Characteristics of Heterosexual Couples with
Sexual Dysfunctions

                                               Receiving
                                             bibliotherapy

                                             M         (SD)

Age male (years)                             38         10
Age female (years)                           35         11
Length of relationship (years)               13         10
Duration of male problem (years)              7          7
Duration of female problem (years)            7          7

                                                  %

Male education high (> 10 years)                  56
Female education high (> 10 years)                48
Have children together                            51
Sexual problem in previous relation male          13
Sexual problem in previous relation female        23
Former help other problem male                    23
Former help other problem female                  43
Current male anxiety disorder (DSM-IV)            11
Current female anxiety disorder (DSM-IV)          14

Male sexual problem type
  No sexual problem                               22
  Hypoactive sexual desire disorder               49
  Erectile disorder                               14
  Premature ejaculation                           16

Female sexual problem type
  No sexual problem                               18
  Hypoactive sexual desire disorder               53
  Anorgasmia                                       9
  Vaginismus                                      12
  Dyspareunia                                      8

                                              Waiting list

                                             M         (SD)

Age male (years)                             41         12
Age female (years)                           38         12
Length of relationship (years)               15         13
Duration of male problem (years)              8          8
Duration of female problem (years)            8          8

                                                  %

Male education high (> 10 years)                  64
Female education high (> 10 years)                60
Have children together                            41
Sexual problem in previous relation male          17
Sexual problem in previous relation female        32
Former help other problem male                    27
Former help other problem female                  42
Current male anxiety disorder (DSM-IV)             9
Current female anxiety disorder (DSM-IV)          17

Male sexual problem type
  No sexual problem                               26
  Hypoactive sexual desire disorder               48
  Erectile disorder                               17
  Premature ejaculation                            9

Female sexual problem type
  No sexual problem                               11
  Hypoactive sexual desire disorder               44
  Anorgasmia                                       9
  Vaginismus                                      17
  Dyspareunia                                     18

Note. N = 199.

The distributions of male and female sexual dysfunction types in the experimental group and control group are shown in Table 1. These differences were not statistically significant for males nor for females. With respect to the psychological, marital, and sexual characteristics of couples and individuals, the subject sample is found to be sexually dysfunctional, but otherwise normal. Neuroticism scores were in the normal range (male SCL-90: M = 114.5, SD = 23.5; female SCL-90: M = 126.5, SD = 32.0: Arrindell & Ettema, 1981), as were marital dissatisfaction scores (male MMQ-M: M = 14.4, SD = 10.5; female MMQ-M: M = 14.5, SD = 10.5; Arrindell, Boelens, & Lambert, 1983). Average general life dissatisfaction was in the dysfunctional range (male MMQGL: M = 8.6, SD = 4.9; female MMQ-GL: M = 8.8, SD = 4.7: Arrindell, Boelens, & Lambert, 1983). Sexual dissatisfaction, as measured by the MMQ-S was clearly in the dysfunctional range (male MMQ-S: M = 18.2, SD = 9.1; female MMQ-S: M = 20.6, SD = 8.7: Arrindell, Boelens, & Lambert, 1983).

Missing data on outcome variables at the three different assessment time-points were inspected and found to be randomly scattered over groups and variables. Cases with missing data on individual outcome variables were retained for comparisons. Missing data in these cases were replaced with the appropriate mean of, respectively, males or females from the same treatment group.

Effects of Cognitive-Behavioral Bibliotherapy on Sexual Problem Status

At posttreatment and follow-up measurement, participants rated the extent to which their sexual problem had improved or become worse during the last 4 weeks. Data are presented in Table 2. To test group differences, the answering categories were collapsed into three new categories: worse (1 = gotten much worse, 2, and 3), unchanged (answering category 4 = not gotten better or worse), and improved (answering categories 5, 6, or 7 = gotten much better). Compared to controls, more treated participants reported improvement at posttreatment: male, [chi square] (df = 2) = 22.6; p [is less than or equal to] .001; female, [chi square] (df = 2) = 15.5; p [is less than or equal to] .001. These differences were also found at follow-up for males, [chi square] (df = 2) = 14.6; p = .001, but not for females. Since participants were asked to rate improvement or worsening in the last 4 weeks, this may be interpreted to indicate further improvement with respect to the sexual problem for males, while the posttreatment situation for female participants essentially was unchanged. In both treatment and control group respectively 92% and 91.8% of the female participants answered that the problem had either not gotten better or had gotten worse, or that they had experienced smaller or larger improvement.

Table 2. Self-rated Improvement of Sexual Problem After
Cognitive-Behavioral Bibliotherapy

                                        Posttreatment(a)

                                             Waiting   Difference
                               Treatment %   list %        %

Males
1 gotten much worse                0.9         2.3        -1.4
2                                  3.6         1.1         2.5
3                                  9.9         9.1         0.8
4 not gotten better or worse      44.1        75.0       -30.9
5                                 28.8        10.2        18.6
6                                  9.9         2.3         7.6
7 gotten much better               2.7         0.0         2.7

Females
1 gotten much worse                0.9         1.1        -0.2
2                                  0.9         1.1        -0.2
3                                  4.5        10.2        -5.7
4 not gotten better or worse      49.5        69.3       -19.8
5                                 30.6        13.6        17.0
6                                 11.7         4.5         7.2
7 gotten much better               1.8         0.0         1.8

                                         Follow-up(b)

                               Treatment   Waiting   Difference
                                   %       list %        %

Males
1 gotten much worse               1.0        0.0         1.0
2                                 1.0        0.0         1.0
3                                12.1        3.5         8.6
4 not gotten better or worse     46.5       72.9       -26.4
5                                25.3       20.0         5.3
6                                13.1        2.4        10.7
7 gotten much better              1.0        1.2        -0.2

Females
1 gotten much worse               1.0        2.4        -1.4
2                                 1.0        0.0         1.0
3                                 6.0        5.9         0.1
4 not gotten better or worse     54.0       61.2        -7.2
5                                26.0       22.4         3.6
6                                11.0        8.2         2.8
7 gotten much better              1.0        0.0         1.0

Note. Posttreatment rating after 10 weeks, follow-up rating after 10
weeks following posttreatment. Question: "To what extent has your
sexual problem improved or worsened during the last 4 weeks?" Minus
sign indicates fewer subjects in treatment group giving this answer.

(a) Posttreatment rating after 10 weeks.

(b) Follow-up rating after 10 weeks following posttreatment.

In the subject sample recruited from the outpatient clinic for sexology, the effects at follow-up can be further demonstrated by the proportional difference between treatment and control group with respect to their drop-out rate from the waiting list for professional therapy. From the treated group (n = 46), 9 couples (19.6%) wished to discontinue, compared to 1 couple (2.6%) from the control group (n = 38), Fisher's Exact Test, two-tailed p = .02.

Effects of Cognitive-Behavioral Bibliotherapy on Sexual Dysfunction

Male participants. The effects of cognitive behavioral bibliotherapy on male sexual dysfunction were evaluated through three dependent variables (DVs): infrequency, erectile dysfunction, and premature ejaculation (GRISS subscales). To examine pretreatment differences between treatment groups and sexual dysfunction types (see Table 3 for mean scores), a 2 (treatment) x 4 (dysfunction type) factorial MANOVA was performed. Total N was 199. The combined DVs were significantly affected by dysfunction type, F(9,460) = 22.70, power = 1, p [is less than] .000 and by the interaction of treatment and dysfunction type, F(9,460) = 2.13, power = .79, p = .026, but not by treatment.

A 2 (treatment) x 4 (dysfunction type) MANCOVA was performed, with adjustment for the pretreatment level of the DVs. The combined DVs were significantly related to the combined covariates, approximate F(9,450) = 69.07, power = 1, p [is less than] .001, A = +2.02, and to treatment, F(3,185) = 5.06, power = .91, p = .002, [Delta] = +0.57, but not to dysfunction type or the interaction of treatment and dysfunction type.

The effect of treatment on DVs after adjustment for covariates was investigated in univariate and stepdown analysis. In the absence of an a priori rank order of the dependent variables on theoretical grounds, prioritization was carried out by ranking DVs according to their univariate observed F-statistic for the main effect of treatment, in descendent order. Table 4 shows the results of this analysis. By apportioning [Alpha] as shown in the last column, an experimentwise error rate of 5% for each effect was established. Infrequency of sexual interaction made a significant contribution to the composite DVs that differentiated between bibliotherapy and waiting-list control, stepdown F(1,187) = 10.44, power = .89, p = .001, [Delta] = +0.47. By univariate analysis, erectile dysfunction differentiated between treatment groups with marginal significance, F(1,187) = 5.03, power = .60, p = .026, [Delta] = +0.33, after adjustment for infrequency and the pretreatment covariares. Stepdown analysis, however, revealed that this difference was already accounted for by the covariates and sexual infrequency.

Table 4. Effects of Bibliotherapy on Male and Female Sexual Dysfunction
Variables.

                                            Posttreatment

                                            Univariate

Effect                DV                      F         df

Males
  Covariates       infrequency              39.62(a)   3/187
  (pretreatment)   impotence                68.70(a)   3/187
                   premature ejaculation    91.34(a)   3/187

  Treatment        infrequency              10.44(a)   1/187
                   impotence                 5.03      1/187
                   premature ejaculation     2.19      1/187

  Dysfunction      infrequency               0.10      3/187
  type             impotence                 3.07      3/187
                   premature ejaculation     1.67      3/187

  Treatment x      infrequency               0.36      3/187
  dysfunction      impotence                 0.53      3/187
  type             premature ejaculation     1.15      3/187

Females

  Covariates       infrequency              48.12(a)   3/186
  (pretreatment)   vaginismus               73.38(a)   3/186
                   anorgasmia              144.69(a)   3/186

  Treatment        infrequency              12.24(a)   1/186
                   vaginismus                2.98      1/186
                   anorgasmia                1.25      1/186

  Dysfunction      infrequency               0.88      4/186
  type             vaginismus                3.71(b)   4/186
                   anorgasmia                0.75      4/186

  Treatment x      infrequency               0.63      4/186
  dysfunction      vaginismus                6.47(a)   4/186
  type             anorgasmia                0.77      4/186

                                            Posttreatment

                                            Stepdown

Effect                DV                       F          df

Males
  Covariates       infrequency              39.62(***)   3/187
  (pretreatment)   impotence                63.82(***)   3/186
                   premature ejaculation    91.65(***)   3/185

  Treatment        infrequency              10.44(***)   1/187
                   impotence                 2.80        1/186
                   premature ejaculation     1.78        1/185

  Dysfunction      infrequency               0.10        3/187
  type             impotence                 3.18        3/186
                   premature ejaculation     1.79        3/185

  Treatment x      infrequency               0.36        3/187
  dysfunction      impotence                 0.48        3/186
  type             premature ejaculation     1.12        3/185

Females

  Covariates       infrequency              48.12(***)   3/186
  (pretreatment)   vaginismus               63.11(***)   3/185
                   anorgasmia              143.15(***)   3/184

  Treatment        infrequency              12.24(***)   1/186
                   vaginismus                1.64        1/185
                   anorgasmia                0.00        1/184

  Dysfunction      infrequency               0.88        4/186
  type             vaginismus                3.89(**)    4/185
                   anorgasmia                0.96        4/184

  Treatment x      infrequency               0.63        4/186
  dysfunction      vaginismus                6.02(***)   4/185
  type             anorgasmia                0.35        4/184

                                         Posttreatment   Follow-up

                                           Stepdown     Univariate

Effect                DV                   [Delta]      F         df

Males
  Covariates       infrequency                       37.57(a)   3/172
  (pretreatment)   impotence                         49.17(a)   3/172
                   premature ejaculation             66.34(a)   3/172

  Treatment        infrequency             0.47       2.27      1/172
                   impotence               0.33       2.49      1/172
                   premature ejaculation   0.22       2.79      1/172

  Dysfunction      infrequency             0.08       0.25      3/172
  type             impotence               0.44       3.61(c)   3/172
                   premature ejaculation   0.33       0.97      3/172

  Treatment x      infrequency             0.15       1.06      3/172
  dysfunction      impotence               0.18       0.72      3/172
  type             premature ejaculation   0.27       3.37(c)   3/172

Females

  Covariates       infrequency                       32.13(a)   3/172
  (pretreatment)   vaginismus                        59.37(a)   3/172
                   anorgasmia                        95.94(a)   3/172

  Treatment        infrequency             0.51       0.08      1/172
                   vaginismus              0.25       4.28      1/172
                   anorgasmia              0.16       0.46      1/172

  Dysfunction      infrequency             0.28       0.56      4/172
  type             vaginismus              0.56       6.39(a)   4/172
                   anorgasmia              0.25       1.01      4/172

  Treatment x      infrequency             0.23       0.97      4/172
  dysfunction      vaginismus              0.75       6.92(a)   4/172
  type             anorgasmia              0.26       0.85      4/172

                                               Follow-up

                                            Stepdown

Effect                DV                       F         df

Males
  Covariates       infrequency             37.57(***)   3/172
  (pretreatment)   impotence               44.99(***)   3/171
                   premature ejaculation   65.20(***)   3/170

  Treatment        infrequency              2.27        1/172
                   impotence                1.71        1/171
                   premature ejaculation    1.95        1/170

  Dysfunction      infrequency              0.25        3/172
  type             impotence                3.68(***)   3/171
                   premature ejaculation    0.70        3/170

  Treatment x      infrequency              1.06        3/172
  dysfunction      impotence                0.49        3/171
  type             premature ejaculation    2.87        3/170

Females

  Covariates       infrequency             32.13(***)   3/172
  (pretreatment)   vaginismus              49.96(***)   3/171
                   anorgasmia              89.23(***)   3/170

  Treatment        infrequency              0.08        1/172
                   vaginismus               4.28        1/171
                   anorgasmia               1.11        1/170

  Dysfunction      infrequency              0.56        4/172
  type             vaginismus               6.87(***)   4/171
                   anorgasmia               1.61        4/170

  Treatment x      infrequency              0.97        4/172
  dysfunction      vaginismus               6.99(***)   4/171
  type             anorgasmia               0.40        4/170

                                              Follow-up

Effect                DV                   [Delta]   [Alpha]

Males
  Covariates       infrequency                        .02
  (pretreatment)   impotence                          .02
                   premature ejaculation              .01

  Treatment        infrequency              0.23      .02
                   impotence                0.24      .02
                   premature ejaculation    0.25      .01

  Dysfunction      infrequency              0.13      .02
  type             impotence                0.50      .02
                   premature ejaculation    0.26      .01

  Treatment x      infrequency              0.27      .02
  dysfunction      impotence                0.22      .02
  type             premature ejaculation    0.49      .01

Females

  Covariates       infrequency                        .02
  (pretreatment)   vaginismus                         .02
                   anorgasmia                         .01

  Treatment        infrequency              0.04      .02
                   vaginismus               0.32      .02
                   anorgasmia               0.10      .01

  Dysfunction      infrequency              0.23      .02
  type             vaginismus               0.77      .02
                   anorgasmia               0.31      .01

  Treatment x      infrequency              0.30      .02
  dysfunction      vaginismus               0.80      .02
  type             anorgasmia               0.28      .01

Note. Tests of covariates, treatment, dysfunction type, and
interaction. Males: N = 199; Females: N = 199. [Delta] = Effect size by
Glass et al. (1981).

(a) Significance level cannot be evaluated but would be [is less than
or equal to] 001 in a univariate context;

(b) Significance level cannot be evaluated, but would be [is less than
or equal to] .01 in a univariate context;

(c) Significance level cannot be evaluated but would be [is less than
or equal to] .02 in a univariate context.

(*) p [is less than or equal to] .02.

(**) p [is less than or equal to] .01.

(***) p [is less than or equal to] .001.

At posttreatment, then, males in the bibliotherapy treatment group are characterized by fewer infrequency complaints compared to control group males when adjustments are made for pretreatment levels of sexual functioning.

A 2 (treatment)x 4 (dysfunction type) MANCOVA was performed on follow-up scores with adjustment for pretreatment level. The combined DVs were significantly related to the combined covariates, F(9,414) = 52.42, power = 1, p [is less than] .001, [Delta] = +1.85, but not to treatment or to dysfunction type, or to the interaction of treatment and dysfunction type. Thus, posttreatment gains of males with regard to sexual dysfunction proved to be lost at follow-up.

Female participants. The effects of cognitive behavioral bibliotherapy on female sexual dysfunction were evaluated through three dependent variables: infrequency, vaginismus, and anorgasmia (GRISS subscales). To assess pretreatment differences, a 2 (treatment) x 5 (dysfunction type) MANOVA was performed on pretreatment data. Total N was 199. The combined DVs were significantly affected by dysfunction type, F(9,424) = 8.92, power = 1, p [is less than] .000, but not by treatment or by the interaction of treatment and dysfunction type

Posttreatment effects of the bibliotherapy approach on female sexual dysfunction was examined with a 2 (treatment) x 5 (dysfunction type) MANCOVA, with adjustment for pretreatment levels. The combined DVs were significantly related to the covariates, F(9,448) = 88.31, power = 1, p [is less than] .001, [Delta] = +2.27; to treatment, F(3,184) = 4.62, power = .89, p = .004, [Delta] = +0.55; to dysfunction type, F(12,487) = 1.89, power = .85, p = .034, [Delta] = +0.40; and to the interaction of treatment and dysfunction type, F(12,487) = 2.26, power = .92, p = .009, [Delta] = +0.44. Subsequent univariate and stepdown analysis revealed that sexual infrequency made a significant contribution to the composite DVs that differentiated between bibliotherapy and waiting-list controls, stepdown F(1,186) = 12.24, power = .93, p = .001, [Delta] = +0.51. Table 4 shows the results of this analysis. After adjustment for the covariates and posttreatment infrequency, vaginismus made a significant contribution to the differentiation of dysfunction types at posttreatment, stepdown F(4,185) = 3.89, power = .88, p = .005, [Delta] = +0.56, as well as to the interaction effect, stepdown F(4,185) = 6.02, power = .99, p [is less than] .001, [Delta] = +0.75. No effect on anorgasmia complaints was observed.

The most interesting of these effects for this study is the interaction effect of treatment and dysfunction type. Post hoc analyses using Scheffe confidence intervals revealed that, compared to the waiting-list group, women in the treated group who were diagnosed with vaginismus reported substantially lower complaints of vaginismus at posttreatment. Females in the treated group with dyspareunia, however, reported more complaints of vaginismus at posttreatment.

A 2 (treatment) x 5 (dysfunction type) MANOVA was performed on follow-up scores with adjustment for pretreatment levels. The combined DVs were significantly related to the covariates, F(9,414) = 59.48, power = 1, p [is less than] .001, [Delta] = +1.96; to the interaction of treatment and dysfunction type, F(12,450) = 2.69, power = .96, p = .002, [Delta] = +0.50; and to dysfunction type, F(12,450) = 2.93, power = .98, p = .001, [Delta] = +0.52; but not to treatment. The main effect on infrequency at posttreatment thus proved to be lost at follow-up.

In subsequent univariate and stepdown analysis, only vaginismus was found to make a significant contribution to the composite DVs that revealed the significant interaction effect, stepdown F(4,171) = 6.99, power = .99, p [is less than] .001, [Delta] = +0.80. Table 4 shows the results of this analysis. Post hoc analyses using Scheffe confidence intervals revealed that, compared to the waiting-list group, women in the treated group who were diagnosed with vaginismus reported substantially lower complaints of vaginismus at follow-up. On the other hand, females in the treated group with dyspareunia reported more complaints of vaginal discomfort at follow-up, pointing to an adverse effect of bibliotherapy for this group. For women with other diagnoses no significant effects on follow-up scores of vaginismus, nor on other dependent variables, were found.

Effects of Cognitive-Behavioral Bibliotherapy on Other Aspects of Sexual and Marital Functioning, and on Client Evaluation of Treatment Effect

Male participants. Other effects of bibliotherapy were investigated through a set of seven outcome variables: problem status, problem distress, sexual satisfaction, intimacy, sexual arousal, self-perception, and marital satisfaction. Mean scores on these variables at different assessment points are shown in Table 3. In order to assess pretreatment differences, a 2 (treatment) x 4 (dysfunction type) MANOVA was performed. The combined DVs were significantly affected by dysfunction type, F(21,526) = 4.48, power = 1, p [is less than] .001, but not by treatment or by the interaction of treatment and dysfunction type. No adjustment for pretreatment levels of the DVs was found necessary.

Table 3. Bibliotherapy for Heterosexual Males and Females with
Sexual Dysfunctions: Outcome Data of Treated and Control Group
Participants.

                                          Males

                                                       Premature
                          Infrequency    Impotence    ejaculation

                           T      C       T      C      T      C

No        Baseline.        7.4   6.8     5.9    6.7    9.4    9.0
problem   Posttreatment    6.8   7.2     6.0    7.1    8.4    9.0
          Follow-up        6.6   7.0     5.5    6.7    7.0    8.6

HSDD      Baseline         7.9   7.9     6.0    6.7    8.6    9.0
          Posttreatment    7.4   7.9     5.8    7.0    7.9    8.4
          Follow-up        7.5   7.7     6.1    7.0    7.8    8.0

ED        Baseline         6.4   7.5    14.3   12.7    9.4   10.6
          Posttreatment    6.5   7.8    13.2   13.3    8.3   10.5
          Follow-up        6.6   7.5    12.8   12.6    8.5   10.6

PE        Baseline         7.1   6.8     6.5    7.9   13.5   10.5
          Posttreatment    6.7   7.0     6.3    7.0   12.6   10.0
          Follow-up        6.7   6.1     5.9    7.4   11.8    8.1

Total     Baseline         7.5   7.4     7.2    7.8    9.7    9.4
          Posttreatment    7.0   7.6     6.9    8.1    8.8    9.1
          Follow-up        7.1   7.3     6.8    7.9    8.3    8.6

                                        Females

                          Infrequency    Vaginismus    Anorgasmia

No        Baseline         7.0   6.6     5.9    6.0   10.3    9.5
problem   Posttreatment    7.1   7.3     5.6    5.8   11.0    9.9
          Follow-up        7.3   6.5     5.3    6.2   11.6   11.4

HSDD      Baseline         8.2   8.6     6.3    7.6   12.1   11.7
          Posttreatment    7.5   8.7     6.1    7.6   11.7   12.0
          Follow-up        7.6   8.4     6.3    7.3   12.2   11.3

FA        Baseline         6.9   6.3     5.9    7.9   15.5   17.8
          Posttreatment    6.9   7.1     6.9    7.5   15.4   16.9
          Follow-up        6.9   6.8     6.1    7.0   15.2   16.5

VAG       Baseline         7.5   7.2    17.2   17.5   12.5   13.0
          Posttreatment    6.6   7.7    14.0   17.5   12.3   14.2
          Follow-up        7.2   7.4    13.5   17.7   12.8   14.4

DYS       Baseline         6.9   7.7    12.1   15.0   14.2   12.9
          Posttreatment    7.0   7.9    15.1   14.1   14.0   12.8
          Follow-up        7.4   7.7    15.1   14.8   14.5   12.7

Total     Baseline         7.6   7.8     7.9   10.5   12.3   12.5
          Posttreatment    7.2   8.1     7.7   10.2   12.1   12.7
          Follow-up        7.4   7.7     7.8   10.4   12.6   12.6

                                       Males

                           Problem     Problem         Sexual
                           Status      distress     satisfaction

                           T     C     T      C      T      C

No        Baseline.       4.1   4.3   14.9   17.1   16.4   13.7
problem   Posttreatment   4.3   3.8   15.6   17.4   18.2   16.8
          Follow-up       4.5   4.1   15.7   17.7   17.1   15.1

HSDD      Baseline        3.9   4.1   13.9   14.3   20.0   20.8
          Posttreatment   4.5   4.0   14.7   14.7   17.7   21.8
          Follow-up       4.4   4.3   15.0   15.0   18.6   20.9

ED        Baseline        3.9   4.2   13.5   12.8   18.1   19.5
          Posttreatment   4.1   4.1   14.5   15.1   15.8   21.3
          Follow-up       4.1   4.3   15.2   16.0   17.0   19.9

PE        Baseline        3.8   4.0   14.1   15.1   14.4   17.0
          Posttreatment   4.3   4.1   14.1   14.0   14.3   17.9
          Follow-up       4.4   4.3   15.4   15.1   11.0   13.4

Total     Baseline        3.9   4.1   14.1   14.8   18.1   18.4
          Posttreatment   4.4   4.0   14.8   15.4   17.0   20.1
          Follow-up       4.4   4.2   15.2   15.9   17.0   18.5

                                        Females

                           Problem     Problem        Sexual
                           status      distress     satisfaction

No        Baseline        4.1   4.1   14.9   14.1   18.2   16.1
problem   Posttreatment   4.4   4.4   16.1   15.2   18.0   18.7
          Follow-up       4.3   4.1   16.3   15.0   18.5   16.6

HSDD      Baseline        3.7   3.9   13.6   13.7   21.9   24.0
          Posttreatment   4.5   3.9   15.7   14.6   17.8   24.2
          Follow-up       4.3   4.2   15.6   14.9   17.8   22.6

FA        Baseline        4.1   4.3   16.7   15.5   15.8   15.8
          Posttreatment   4.4   4.0   16.2   13.9   15.7   20.9
          Follow-up       4.3   4.5   15.2   14.3   16.9   16.6

VAG       Baseline        4.2   4.1   12.7   14.1   18.8   20.1
          Posttreatment   5.0   3.9   15.5   15.2   19.3   21.8
          Follow-up       4.9   4.3   15.9   15.5   17.4   21.0

DYS       Baseline        3.9   3.8   12.1   13.1   20.1   20.6
          Posttreatment   4.2   4.3   13.1   14.9   21.0   22.2
          Follow-up       4.3   4.4   13.3   15.2   23.1   21.5

Total     Baseline        3.9   4.0   13.9   13.9   20.1   21.1
          Posttreatment   4.5   4.1   15.6   14.8   18.1   22.5
          Follow-up       4.4   4.3   15.5   15.0   18.2   20.8

                                          Males

                                         Sexual
                           Intimacy      arousal      Self-perception

                           T      C      T      C        T      C

No        Baseline.       25.3   29.2   22.5   18.3     56.2   64.7
problem   Posttreatment   29.0   29.8   23.6   26.6     61.9   63.3
          Follow-up       33.7   28.6   28.1   21.2     61.1   63.4

HSDD      Baseline        42.7   42.4   28.8   26.0     50.4   49.7
          Posttreatment   39.9   42.9   27.0   31.1     59.1   56.8
          Follow-up       39.0   41.1   27.2   29.9     61.4   60.5

ED        Baseline        28.2   36.7   34.5   40.6     57.5   58.2
          Posttreatment   27.8   37.5   33.1   45.5     63.7   57.7
          Follow-up       29.6   39.2   31.4   44.4     61.5   57.9

PE        Baseline        36.4   32.9   18.9   21.5     42.6   53.5
          Posttreatment   34.8   35.8   19.5   26.9     51.6   63.1
          Follow-up       35.1   31.3   14.1   25.4     52.4   62.7

Total     Baseline        36.0   37.1   26.6   26.1     51.4   55.4
          Posttreatment   35.1   37.9   25.9   32.0     59.1   59.2
          Follow-up       36.0   36.6   26.1   29.6     60.1   61.0

                                         Females

                                         Sexual
                           Intimacy      arousal      Self-perception

No        Baseline        36.0   30.6   44.6   38.1     65.8   69.0
problem   Posttreatment   32.9   28.4   39.6   38.7     70.4   66.7
          Follow-up       35.6   29.3   45.1   40.8     68.8   69.1

HSDD      Baseline        45.5   44.8   57.7   57.4     76.5   67.1
          Posttreatment   40.9   44.6   51.2   55.1     77.9   69.6
          Follow-up       39.0   49.2   49.5   55.7     80.0   74.3

FA        Baseline        35.9   29.8   43.5   43.0     80.0   79.0
          Posttreatment   28.6   33.3   41.2   43.6     75.2   75.0
          Follow-up       30.9   29.5   38.4   39.8     79.8   78.1

VAG       Baseline        24.6   32.1   34.7   41.8     77.1   74.5
          Posttreatment   22.2   33.2   29.5   44.7     81.8   71.3
          Follow-up       24.6   28.3   32.4   39.1     81.2   74.9

DYS       Baseline        31.0   31.9   48.3   42.4     78.2   78.4
          Posttreatment   30.4   31.1   47.2   44.4     81.8   75.1
          Follow-up       30.3   33.0   44.1   44.8     73.5   75.3

Total     Baseline        39.3   37.3   50.6   48.5     75.1   71.7
          Posttreatment   35.3   37.3   45.3   48.5     77.1   71.0
          Follow-up       35.2   38.3   45.2   47.4     77.8   74.4

                             Males

                            Marital
                          satisfaction

                           T      C

No        Baseline.       11.6   10.4
problem   Posttreatment   11.4   10.9
          Follow-up       12.2   11.0

HSDD      Baseline        17.2   17.6
          Posttreatment   15.7   17.4
          Follow-up       15.2   16.2

ED        Baseline        10.8   11.9
          Posttreatment    9.9   10.3
          Follow-up       12.5   10.4

PE        Baseline        12.3   14.1
          Posttreatment   15.8   18.3
          Follow-up       15.2   16.7

Total     Baseline        14.3   14.4
          Posttreatment   14.0   14.6
          Follow-up       14.1   13.9

                            Females

                            Marital
                          satisfaction

No        Baseline        14.7   14.8
problem   Posttreatment   12.9   14.1
          Follow-up       13.0   11.9

HSDD      Baseline        17.7   18.1
          Posttreatment   15.7   17.9
          Follow-up       16.8   17.3

FA        Baseline        12.8   10.0
          Posttreatment   10.1   15.3
          Follow-up       12.0   14.9

VAG       Baseline         8.1    7.7
          Posttreatment   10.1   11.3
          Follow-up        9.8   11.6

DYS       Baseline        10.4   10.6
          Posttreatment   10.9   12.7
          Follow-up       10.1   15.3

Total     Baseline        15.0   13.8
          Posttreatment   13.6   15.1
          Follow-up       14.3   15.0

Note. HSDD = hypoactive sexual desire disorder;
ED = erectile disorder;
PE = premature ejaculation;
DYS = female dyspareunia;
FA = female anorgasmia;
VAG = vaginismus;
T = bibliotherapy;
C = waiting list.
Males: N = 199;
Females: N = 199.

A 2 (treatment) x 4 (dysfunction type) MANCOVA was performed on these seven posttreatment DVs. To separate the effect of treatment on these DVs from the effect already found for male sexual dysfunction variables, adjustment was made for posttreatment level of infrequency, erectile dysfunction and premature ejaculation. The combined DVs were significantly related to the covariates, F(21,517) = 7.67, power = 1, p [is less than] 001, [Delta] = +1.09; to treatment, F(7,180) = 2.35, power = .84, p = .026, [Delta] = +0.61; and to dysfunction type, F(21,517) = 2.06, power = .99, p = .004, [Delta] = +0.57; but not to the interaction of treatment and dysfunction type.

The effects of treatment on these outcome variables were investigated in subsequent univariate and stepdown analysis. Table 5 shows the results of this analysis. By apportioning [Alpha] as shown in the last column an experimentwise error rate of 5% for each effect was established. After adjustment for sexual dysfunction covariates and improvement of problem status, males in the bibliotherapy group were found to be characterized by lower distress rates associated with their problem compared to waiting-list males, stepdown F(1,185) = 10.65, power = .54, p = .001, [Delta] = +0.31.

Table 5. Posttreatment and Follow-Up Effect of Cognitive-Behavioral
Bibliotherapy for Male Dependent Variables.

                                             Posttreatment

                                     Univariate           Stepdown

Effect        Dependent
              Variable                   F         df       F

Covariates    Improvement             12.07(a)    3/186   12.07(**)
              Problem distress         4.48(b)    3/186    0.64
              Sexual satisfaction     40.41(a)    3/186   24.86
              Intimacy                 5.65(a)    3/186    2.87
              Sexual arousal          14.63(a)    3/186    6.52(**)
              Self-perception          7.11(a)    3/186    5.22(*)
              Marital satisfaction     1.40       3/186    1.14

Treatment     Improvement              4.64       1/186    4.64
              Problem distress         4.33       1/186   10.65(**)
              Sexual satisfaction      0.64       1/186    0.62
              Intimacy                 0.06       1/186    0.00
              Sexual arousal           0.50       1/186    0.04
              Self-perception          0.02       1/186    0.08
              Marital satisfaction     0.00       1/186    0.48

Dysfunction   Improvement              1.82       3/186    1.82
type          Problem distress         3.08       3/186    5.21(*)
              Sexual satisfaction      0.03       3/186    0.35
              Intimacy                 3.45       3/186    5.15(*)
              Sexual arousal           0.04       3/186    0.19
              Self-perception          0.86       3/186    0.46
              Marital satisfaction     3.72       3/186    1.37

Treatment x   Improvement              1.44       3/186    1.44
dysfunction   Problem distress         0.64       3/186    1.04
type          Sexual satisfaction      1.37       3/186    1.43
              Intimacy                 0.30       3/186    0.20
              Sexual arousal           0.56       3/186    0.98
              Self-perception          0.76       3/186    0.62
              Marital satisfaction     0.11       3/186    0.15

                                       Post-
                                     treatment   Follow-up

                                                 Univariate

Effect        Dependent
              Variable                  df            F         df

Covariates    Improvement             3/186        8.90(a)    3/170
              Problem distress        3/185        5.45(a)    3/170
              Sexual satisfaction     3/184       26.20(a)    3/170
              Intimacy                3/182        7.64(a)    3/170
              Sexual arousal          3/183       15.33(a)    3/170
              Self-perception         3/181        8.93(a)    3/170
              Marital satisfaction    3/180        0.86       3/170

Treatment     Improvement             1/186        1.20       1/170
              Problem distress        1/185        2.02       1/170
              Sexual satisfaction     1/184        0.40       1/170
              Intimacy                1/182        0.04       1/170
              Sexual arousal          1/183        0.10       1/170
              Self-perception         1/181        0.06       1/170
              Marital satisfaction    1/180        0.05       1/170

Dysfunction   Improvement             3/186        1.10       3/170
type          Problem distress        3/185        1.24       3/170
              Sexual satisfaction     3/184        1.26       3/170
              Intimacy                3/182        1.62       3/170
              Sexual arousal          3/183        0.37       3/170
              Self-perception         3/181        0.55       3/170
              Marital satisfaction    3/180        3.99       3/170

Treatment x   Improvement             3/186        0.88       3/170
dysfunction   Problem distress        3/185        1.57       3/170
type          Sexual satisfaction     3/184        1.38       3/170
              Intimacy                3/182        1.21       3/170
              Sexual arousal          3/183        2.49       3/170
              Self-perception         3/181        0.39       3/170
              Marital satisfaction    3/180        0.32       3/170

                                     Follow-up

                                     Stepdown

Effect        Dependent
              Variable                 F          df    [Alpha]

Covariates    Improvement            5.36(*)    3/169   .007
              Problem distress       5.45(**)   3/170   .007
              Sexual satisfaction    6.02(**)   3/186   .007
              Intimacy               2.44       3/164   .007
              Sexual arousal         5.88(**)   3/167   .077
              Self-perception        8.29(**)   3/166   .007
              Marital satisfaction   2.57       3/165   .007

Treatment     Improvement            2.82       1/169   .007
              Problem distress       2.02       1/170   .007
              Sexual satisfaction    0.76       1/168   .007
              Intimacy               0.41       1/164   .007
              Sexual arousal         0.06       1/167   .007
              Self-perception        0.00       1/166   .007
              Marital satisfaction   0.34       1/165   .007

Dysfunction   Improvement            1.97       3/169   .007
type          Problem distress       1.24       3/170   .007
              Sexual satisfaction    1.92       3/168   .007
              Intimacy               0.96       3/164   .007
              Sexual arousal         0.04       3/167   .007
              Self-perception        0.60       3/165   .007
              Marital satisfaction   5.43(**)   3/165   .007

Treatment x   Improvement            1.03       3/169   .007
dysfunction   Problem distress       1.57       3/170   .007
type          Sexual satisfaction    0.93       3/168   .007
              Intimacy               0.55       3/164   .007
              Sexual arousal         2.23       3/167   .007
              Self-perception        0.10       3/166   .007
              Marital satisfaction   0.05       3/165   .007

Note: Males: N = 199. Univariate and stepdown tests of covariates,
treatment, dysfunction type and interaction.

(a) Significance level cannot be evaluated but would be [is less than
or equal to] .001 in a univariate context.

(b) Significance level cannot be evaluated but would be < .007
in a univariate context.

(*) p [is less than or equal to] .007.

(**) p [is less than or equal to] .001.

Next, a 2 (treatment) x 4 (dysfunction type) MANCOVA was performed on these seven DVs at follow-up, with adjustment for follow-up level of infrequency, erectile dysfunction, and premature ejaculation. The combined DVs were significantly related to the covariates, F(21,471) = 6.84, power = 1, p [is less than] .001, [Delta] = +1.08, and to dysfunction type, F(21,471) = 1.73, power = .97, p = .024, [Delta] = +0.54, but not to treatment or to the interaction. This indicates loss of the posttreatment gains for males in the treated group.

Female participants. A 2 (treatment) x 5 (dysfunction type) MANOVA was performed on the seven pretreatment DVs of female participants. The combined DVs were significantly affected by dysfunction type, F(28,650) = 2.80, power = 1, p [is less than] .001, but not by treatment or by the interaction of treatment and dysfunction type. No adjustment for pretreatment covariates was found necessary.

A 2 (treatment) x 5 (dysfunction type) MANCOVA was performed on these seven female DVs at posttreatment. To separate the effect of treatment on these DVs from the effect already found for female sexual dysfunction variables, adjustment was made for posttreatment level of infrequency, vaginismus, and anorgasmia. The combined DVs were significantly related to the covariates, F(21,512) = 7.21, power = 1, p [is less than] .001, [Delta] = + 1.06, but not to treatment, to dysfunction type, or to the interaction of treatment and dysfunction type. Univariate and stepdown statistics can be found in Table 6.

Table 6. Posttreatment and Follow-up Effect of Cognitive-Behavioral
-Bibliotherapy for Female Dependent Variables.

                                             Posttreatment

                                     Univariate           Stepdown

Effect        Dependent variable         F         df       F

Covariates    Improvement             11.36(a)    3/184   11.36(**)
              Problem distress         6.17       3/184    1.80
              Sexual satisfaction     49.65(a)    3/184   21.61(**)
              Intimacy                14.58(a)    3/184   10.35(**)
              Sexual arousal          13.85       3/184    2.60
              Self-perception          1.75       3/184    1.50
              Marital satisfaction     1.12       3/184    0.87

Treatment     Improvement              3.92       1/184    3.92
              Problem distress         0.06       1/184    0.12
              Sexual satisfaction      0.79       1/184    0.30
              Intimacy                 1.30       1/184    2.59
              Sexual arousal           0.36       1/184    0.08
              Self-perception          2.77       1/184    2.69
              Marital satisfaction     0.11       1/184    0.37

Dysfunction   Improvement              0.40       4/184    0.40
type          Problem distress         0.57       4/184    0.55
              Sexual satisfaction      0.44       4/184    0.66
              Intimacy                 4.92(a)    4/184    5.52(**)
              Sexual arousal           4.91(a)    4/184    1.07
              Self-perception          1.49       4/184    1.47
              Marital satisfaction     1.15       4/184    0.11

Treatment x   Improvement              1.78       4/184    1.78
dysfunction   Problem distress         1.09       4/184    0.82
type          Sexual satisfaction      0.89       4/184    1.73
              Intimacy                 0.39       4/184    0.33
              Sexual arousal           0.36       4/184    0.23
              Self-perception          0.24       4/184    0.23
              Marital satisfaction     0.18       4/184    0.29

                                       Post-
                                     treatment   Follow-up

                                            Univariate

Effect        Dependent variable       df       F         df

Covariates    Improvement             3/184    3.91      3/196
              Problem distress        3/178    4.16(b)   3/169
              Sexual satisfaction     3/181   30.77(a)   3/169
              Intimacy                3/182   12.76(a)   3/169
              Sexual arousal          3/180   16.96(a)   3/169
              Self-perception         3/183    1.87      3/169
              Marital satisfaction    3/179    0.55      3/169

Treatment     Improvement             1/184    0.07      1/169
              Problem distress        1/178    0.18      1/169
              Sexual satisfaction     1/181    1.00      1/169
              Intimacy                1/182    0.17      1/169
              Sexual arousal          1/180    0.00      1/169
              Self-perception         1/183    1.01      1/169
              Marital satisfaction    1/179    0.06      1/169

Dysfunction   Improvement             4/184    1.81      4/169
type          Problem distress        4/178    1.06      4/169
              Sexual satisfaction     4/181    0.97      4/169
              Intimacy                1/182    4.83(a)   4/169
              Sexual arousal          4/180    4.85(a)   4/169
              Self-perception         4/183    2.71      4/169
              Marital satisfaction    4/179    2.10      4/169

Treatment x   Improvement             4/184    0.81      4/169
dysfunction   Problem distress        4/178    0.86      4/169
type          Sexual satisfaction     4/181    0.82      4/169
              Intimacy                4/182    0.92      4/169
              Sexual arousal          4/180    0.22      4/169
              Self-perception         4/183    0.21      4/169
              Marital satisfaction    4/179    0.17      4/169

                                           Follow-up

                                      Stepdown

Effect        Dependent variable      F          df     [Alpha]

Covariates    Improvement            0.35       3/165   .007
              Problem distress       1.33       3/167   .007
              Sexual satisfaction    7.97(**)   3/168   .007
              Intimacy               1.83       3/166   .007
              Sexual arousal         1.88       3/163   .007
              Self-perception        1.87       3/169   .007
              Marital satisfaction   5.17(*)    3/164   .007

Treatment     Improvement            0.00       1/165   .007
              Problem distress       1.04       1/167   .007
              Sexual satisfaction    0.68       1/168   .007
              Intimacy               0.03       1/166   .007
              Sexual arousal         0.11       1/163   .007
              Self-perception        1.01       1/169   .007
              Marital satisfaction   0.58       1/164   .007

Dysfunction   Improvement            1.15       4/165   .007
type          Problem distress       1.25       4/167   .007
              Sexual satisfaction    0.74       4/168   .007
              Intimacy               4.74(**)   4/166   .007
              Sexual arousal         0.94       4/163   .007
              Self-perception        2.71       4/169   .007
              Marital satisfaction   0.67       4/164   .007

Treatment x   Improvement            0.71       4/165   .007
dysfunction   Problem distress       1.00       4/167   .007
type          Sexual satisfaction    0.72       4/168   .007
              Intimacy               0.46       4/166   .007
              Sexual arousal         0.13       4/163   .007
              Self-perception        0.21       4/169   .007
              Marital satisfaction   0.81       4/164   .007

Note. Females: N = 199. Univariate and stepdown tests of covariates,
treatment, dysfunction type and interaction.

(a) Significance level cannot be evaluated but would be [is less than
or equal to] .001 in a univariate context.

(b) Significance level cannot be evaluated but would be < .007
in a univariate context.

(*) p [is less than or equal to] .007.

(**) p [is less than or equal to] .001.

A 2 (treatment) x 5 (dysfunction type) MANCOVA was performed on seven female DVs at follow-up, with adjustment for follow-up level of infrequency, vaginismus, and anorgasmia. The combined DVs were significantly related to the covariates, F(21,469) = 5.42, power = 1, p [is less than] .001, [Delta] = +0.96, and to dysfunction type, F(28,589) = 1.72, power = .98, p = .012, [Delta] = +0.54, but not to treatment or to the interaction of treatment and dysfunction type.

For female participants, then, no effect of bibliotherapy was found at either posttreatment or follow-up assessment on variables reflecting other aspects of individual, sexual, and marital functioning, which were not already accounted for by the effects on sexual dysfunction variables.

Intention-to-Treat Analysis

Despite efforts to follow up as many participants as possible, a small number of couples dropped out of the study after pretreatment assessment and between posttreatment and follow-up assessment, resulting in partial loss of end-point data.

Pretreatment data of follow-through couples and couples Who dropped out between pretreatment and posttreatment assessment (early dropouts) were compared. Of 223 couples who were assigned to treatment and waiting list group, 24 dropped out after pretreatment assessment, of which 14 couples (11%) were from the treatment group, and 10 couples (10%) were from the control group. This difference was not significant. No differences were found between the distributions of male or female sexual dysfunction types of dropout and follow-through couples. No univariate or multivariate differences were found with regard to demographic characteristics or pretreatment levels of outcome variables except for fewer pretreatment vaginismus complaints in female participants of early dropout couples, stepdown F(1,209) = 4.60, power = 0.57, p = 0.033.

A comparison was made of data of follow-through couples and couples who dropped out between posttreatment and follow-up assessment (late dropouts). Of 199 couples who completed posttreatment assessment, 14 dropped out before follow-up assessment, 11 couples (9%) from the treatment group and 3 couples (3%) from the control group. This difference was not significant. No differences were found with regard to the distribution of female sexual dysfunction types. Male sexual dysfunction types were associated with differences in late drop out, [chi square] (3) = 8.08, p = .044. Males with premature ejaculation were found to drop out more often before follow-up than other male dysfunction types. No differences in demographic and other characteristics were found. Dropout and follow-through couples were also compared with respect to posttreatment levels of outcome variables. Separate MANOVAs for male and female participants were conducted, with simultaneous entry of all outcome variables regarding sexual dysfunction as well as regarding other aspects of individual, sexual, and marital functioning. A significant multivariate difference for male posttreatment values was found, F(10,176) = 2.18, power = 0.90, p = 0.021. Subsequent univariate and stepdown testing revealed that dropout males had higher posttreatment scores on problem-associated distress, stepdown F(1,185) = 7.55, power = 0.78, p = 0.007, and on premature ejaculation, stepdown F(1,183) = 3.89, power = 0.48, p = 0.05. No multivariate differences for female posttreatment outcome data were found.

Compliance

At posttreatment assessment after 10 weeks, compliance to the bibliotherapy method was found to vary widely. Estimated time spent in reading the manual varied from 0 to 40 hours (M = 5.4; SD = 6.3) for male participants and from 0 to 30 hours (M = 6.1; SD = 5.6) for female participants. The number of completed rational-emotive self-analyses ranged from 0 to 35 (M = 1.2; SD = 4.1) for male participants and from 0 to 20 (M = 1.1; SD = 3.1) for female participants. The number of different exercises ranged from 0 to 6 (M = 1.0; SD = 1.3) for male participants and from 0 to 6 (M = 1.2; SD = 1.5) for female participants. The total number of times exercized ranged from 0 to 40 (M = 5.1; SD = 8.5) for male participants and from 0 to 40 (M = 4.3; SD = 7.4) for female participants. Spearman [Rho] between partners' compliance scores was significant, but small (0.39, p [is less than] .001; Cohen, 1988), for time spent reading and for the number of completed rational-emotive self-analyses (0.28, p [is less than] .001). For exercises, some of which require the simultaneous effort of both partners, correlations were of medium size. Spearman's [Rho] was 0.50 (p [is less than] .001) for number of different exercises and 0.49 (p [is less than] .001) for total number of times exercises were performed.

To check on socially desirable answers to questions regarding compliance, Spearman correlations were calculated between each partner's ratings of their own compliance behavior and the other partner's ratings of this behavior. For the male participants, [Rho] was found to vary between 0.44 for number of completed rational-emotive self-analyses and 0.68 for estimated time spent in reading. For female participants, p varied between 0.40 for number of completed rational-emotive self-analyses and 0.65 for total number of exercises performed. Reading and pondering over a rational-emotive self-analysis can be done in solitude, going unnoticed for the other partner. Many of the exercises, however, require both partners' presence, and this will, obviously, increase the concordance of both partners' reports. From the size of these correlations, the answers to compliance questions can be considered a fairly reliable data source regarding the daily practice of bibliotherapy. It must be recognised, however, that this type of check on compliance cannot rule out the possibility of recall bias.

The association of compliance with treatment outcome was examined through multiple regression. For every participant, sum scores were calculated of the gain scores between baseline and posttreatment or follow-up on the three DVs reflecting sexual functioning, as listed in the Results section. To balance the influence of the infrequency variable on this sum score, its gain score was doubled before being added to the sum score. Treatment outcome was thus entered as DV in the regression equation. IVs were seven compliance variables: estimated time spent in reading, number of completed rational-emotive self-analyses, number of different exercises performed, total number of times exercises were performed, number of telephone contacts, and self-rated and partner-rated effort to apply the self-help method. They were entered stepwise into the regression equation. A significant regression model was built for the compliance variables with respect to male outcome at posttreatment, R = 0.44, F(1) = 23.53, p [is less than or equal to] .001. The only variable that proved to contribute to the model was partner-rated male effort (t = 4.85, p [is less than or equal to] .001). Male participants making more efforts to solve the sexual problem, as rated by their partner, showed stronger posttreatment effects. A similar association (R = 0.52, F(1) = 30.98, p [is less than or equal to] .001) between partner rating of male effort and treatment effect was found at follow-up, with t = 5.57, p [is less than or equal to] .001. A significant association of compliance with female outcome at posttreatment was found, R = 0.40, F(3) = 6.21, p = .001. Partner-rated effort (t = 3.07, p = .003), the number of completed rational-emotive self-analyses (t = 2,85, p = .005), and the number of times therapist support by telephone was initiated (t = 2.02, p = .046) were found to contribute significantly to the regression model. No association of compliance variables was found with treatment effects for female participants at follow-up.

Compliance with the bibliotherapy requirements of the male participants, evaluated as the male's therapeutic efforts as rated by their female partners, was thus found to be significantly related to the male participants' treatment outcome at both posttreatment and follow-up assessment. Compliance with the bibliotherapy requirements of the female participants was also found to be significantly related to female treatment outcome at posttreatment, but not at follow-up assessment. More specifically, the higher the regard of the male partner for the female's effort to solve the sexual problem, the larger the number of rational-emotive self-analyses completed, and the more frequent the supportive contacts by telephone during the 10-week treatment period, the better was the effect of treatment on female sexual functioning.

DISCUSSION

As hypothesized, cognitive-behavioral bibliotherapy was found to lead to positive results at the end of the 10-week treatment period for both male and female participants. On a self-report measure indicating their satisfaction with treatment, significantly more treated male participants reported improvement of their sexual functioning both at posttreatment and follow-up, compared to male controls. On the same measure, significantly more treated female participants reported significant improvement of their sexual functioning at posttreatment, compared to female controls. At follow-up, no further improvement was found. When compared to the waiting-list control group, males as well as females were found to have fewer complaints of infrequency of sexual interaction. At follow-up these posttreatment gains, although partially maintained, were no longer significant. The treatment also resulted in lower male ratings of distress associated with the sexual dysfunction at posttreatment. This effect, too, was lost at follow-up. This erosion of effect with regard to specific sexual complaints during follow-up was also found in a meta-analysis of studies on bibliotherapy of sexual dysfunctions (van Lankveld, 1998) and may have several causes. Sexual interaction, which is characteristically avoided by many couples with sexual dysfunction, may be disinhibited by following the guidelines and suggestions for exercises in the manual, but perhaps only temporarily. If further improvement does not come forth, for instance because sexual physiology remains impaired, pre-existent interactional patterns may return. In a review of the treatment of sexual dysfunctions by standard sex therapy, Hawton (1995) reported that treatments sessions "... mostly occur weekly (...). The programme lasts on average between eight and 20 sessions over a period of 3-9 months" (p. 308). The length of the treatment period in our study may have been too limited to consolidate changes which took place during the initial treatment stage. Moreover, consolidation may be dependent upon the continued performance of exercises and the use of the cognitive trouble-shooting strategies. In line with the findings in studies of therapist-administered sex therapy (Hawton & Catalan, 1986, 1990; Hawton, Catalan, & Fagg, 1991), compliance to the bibliotherapy method was positively associated with treatment effect at posttreatment. When compliance information at posttreatment and follow-up is compared, large differences appear. Both male and female participants had spent less time reading the manual and had made fewer rational-emotive self-analyses in the period between posttreatment and follow-up, compared to the first 10 weeks of treatment. The number of times they had performed exercises did not differ between these periods, but male and female participants gave lower ratings of their own effort as well as their partners' effort to apply the self-help method. Partner-rated male effort was found to be the only variable that proved to predict treatment outcome for male participants at posttreatment and follow-up. This finding may well be explained by the greater value placed by women on their male partner's long-term investment in overall relationship quality (Buss, 1998; Regan & Berscheid, 1996). When their partner displays greater effort at solving his sexual problem, female partners may be more inclined to engage in sexual interaction and to provide the necessary assistance to exercises. The significant association of treatment effects and partner-perceived problem-solving effort augments the importance of continued compliance to bibliotherapy requirements. Taken together, this points to diminishing efforts in the follow-up period when support by a professional is no longer offered. Continued therapist support over a longer period of time may be necessary to maintain posttreatment gains. Lack of compliance may, furthermore, reflect dysfunctionality of the relationship, with partner conflicts that may have gone unrecognized by the therapist. Partners, too, may be unaware of the impact of their sometimes subtle strife on sexual interaction.

In our treatment approach participants received an offer to make use of the available therapist support by telephone. Regular supportive contacts were not scheduled in advance to enhance the ecological validity of the therapeutic intervention. It was intended to study the effect of the self-help approach in conditions as closely as possible resembling the conditions of totally unassisted self-help. It can be considered a methodological pitfall to investigate a self-help method under therapist-administered conditions (Glasgow & Rosen, 1978). Moreover, making appointments for regular telephone calls may be experienced by participants as an invasion of their privacy, while they had decided to try out a self-help approach to protect their privacy. This may be an important reason why the offer of professional assistance was not accepted very often.

Although the effects of different elements of the cognitive-behavioral treatment as presented in the bibliotherapy manual were not examined directly, some speculations can be made concerning the differential effectivenes of these elements in the context of the present study. The low average number of rational-emotive analyses which participants reported to have completed makes it unlikely that the rational-emotive troubleshooting method was very helpful in improving sexual problems through bibliotherapy. Therapist support may be a prerequisite for the proper use of such cognitive interventions. Exercises, many of which were modified elements of standard sex therapy, have probably played a larger role since participants reported substantial frequencies of exercise performance.

A differential effect of treatment was found for women with different types of sexual dysfunctions. Female participants diagnosed with vaginismus had fewer complaints of vaginismus at posttreatment as measured with a self-report instrument. Female participants who were diagnosed with dyspareunia, however, reported more complaints of vaginal discomfort at posttreatment. This difference was also found at follow-up. It is conceivable that this finding could be explained by unrecognized organic causes of dyspareunia. Dyspareunia is the sexual dysfunction for which organic etiology is most frequently found (Quevillon, 1993). Of 9 female participants with dyspareunia in the treated group, however, only one entered the study after recruitment via an advertisement without prior gynecological examination. The other participants in the treated group entered the study while being on the waiting list for sex therapy at the outpatient sexology clinic. All were thoroughly examined by a gynecologist to exclude organic causes before being placed on the waiting list. It is conceivable, however, that 10 weeks of treatment by means of sensate focus exercises may have been too short a period for women with dyspareunia to allow healing of tender areas of the vaginal vestibulum (Quevillon, 1993) which are often found in women with sexual pain disorder, both with and without organic causes for sexual pain or vestibular tenderness. Unguided self-help through bibliotherapy simply may not be the best avenue to pursue for couples in which the female partner suffers from dyspareunia.

The rates of drop out from the study were limited to, respectively, 10% before posttreatment and another 6% before follow-up assessment. The most conspicuous characteristic of participants who dropped out, compared to other male dysfunction types, was the diagnosis of premature ejaculation. This agrees with high drop-out rates in previous bibliotherapy studies of men with premature ejaculation (Trudel & Proulx, 1987; Zeiss, 1978). No differences were found with regard to drop out between female sexual dysfunction types. Compared to follow-through couples, the male partners of couples who dropped out before posttreatment assessment were characterized by not only higher baseline erectile dysfunction complaints and higher marital dissatisfaction, but also by higher perceived intimacy during sexual interaction by the male partner and by lower baseline vaginismus scores of the female partners. Of dropout couples before follow-up, males had higher posttreatment scores on problem-associated distress and on premature ejaculation. These differences with regard to drop out from bibliotherapy are in need of further explanation. The professional in primary care may, nevertheless, want to carefully consider the recommendation of a self-help approach to a couple in which the male partner suffers from premature ejaculation because of the risk of low compliance.

Although the drop-out rates in this study were low, part of this drop out, as well as the observed low compliance to the treatment method, may have to do with the specific situation in the Netherlands with respect to treatment facilities for sexual problems. Professional treatment of sexual dysfunctions is available for most categories in the adult population through coverage of costs by health insurance companies. In particular, this is the case for help provided in hospitals and community facilities (such as Regionaal Instituut voor Ambulante Geestelijke Gezondheidszorg) for mental health problems. Easy access to professional help may have a deleterious effect on the motivation of individuals and couples with such problems to apply for self-help, and to comply with the requirements of the self-help method for prolonged periods of time.

The multivariate effect sizes at posttreatment of the present study with respect to the sexual functioning of males (Glass' [Delta] = 0.57) and of females ([Delta] = 0.55) are slightly larger than the effect size of the combined previous studies, when weighted for sample size, [Delta] = 0.50 (van Lankveld, 1998). The observed multivariate effect sizes at follow-up for males ([Delta] = 0.38) and females ([Delta] = 0.36), although slightly larger than the combined mean effect size from previous studies ([Delta] = 0.30), were not significantly different from zero. In essence, this warrants the conclusion that the cognitive-behavioral bibliotherapy used here produced effects that are largely equal to those of former bibliotherapy approaches, which were entirely based on the directed practice therapy of sexual dysfunctions. Although it has been hypothesized that bibliotherapy might only be effective for people with minor sexual difficulties (Hawton, 1995), the results of the present study indicate that a medium-size posttreatment effect is also demonstrated in a mixed subject sample composed of both referred clinical patients and self-referred volunteers.

The largest effect of treatment at posttreatment was found on male and female complaints of low frequency of sexual interaction. Although a substantial portion of the subject couples had a primary diagnosis of hypoactive sexual desire disorder, an interaction effect of treatment and dysfunction type was found only for female participants with a diagnosis of either vaginismus or dyspareunia, with regard to posttreatment vaginismus scores. The absence of an interaction effect on infrequency points to a global effect of cognitivebehavioral bibliotherapy, regardless of dysfunction type, on the frequency of sexual interaction and the perception and evaluation of this frequency by the individual partners. It has not been shown whether the actual frequency of sexual interaction increased or an unchanged frequency was differently evaluated, since absolute frequency of sexual interaction was not assessed.

We can only speculate about the high proportion in our study of males and females with low sexual desire. Self-presented sexual problems of the participants were recorded to allocate them to the respective dysfunction types, without the research assistant putting any pressure on them to reveal possible underlying dysfunctions. We decided to do so for the purpose of approximating as closely as possible the real-life situation in which people buy a self-help book on the basis of their own perception of what constitutes their sexual problem. The average duration of the sexual problems was rather high (between 7 and 8 years). In many cases of such long problem histories, sexual contact is avoided by identified patients. The lack of desire to initiate sexual interaction or to respond to their partner's inititiave may dominate the way they experience their situation. This may have resulted in the large rate of self-defined desire problems in our sample. The sizeable proportion of participants with sexual desire problems in our sample may have had considerable decreasing impact on the overall treatment effect of bibliotherapy. In the sexological literature, sexual desire problems, of all sexual dysfunctions, are considered the hardest to treat and the most resistant to change (Beck, 1995).

Further effort should be directed at identifying individual and couple characteristics that predict the effect of cognitive-behavioral bibliotherapy. With the aid of such predictors, physicians in general practice, gynecologists, and urologists, who are frequently consulted for sexual dysfunctions, could more appropriately recommend the use of bibliotherapy.

REFERENCES

Abel, E. L. (1985). Psychoactive drugs and sex. New York: Plenum Press.

American Psychiatric Association (1987)..Diagnostic and statistical manual of mental disorders (3rd ed. - Revised). Washington, DC: Author.

American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.

American Psychological Association. (1995). Training in and dissemination of empirically-validated psychological treatments: Report and recommendations. The Clinical Psychologist, 48, 3-24.

Andrews, G., Peters, L., Guzman, A. M., & Bird, K. A. (1995). Comparison of two structured diagnostic interviews: CIDI and SCAN. Australian and New Zealand Journal of Psychiatry, 29, 124-132.

Arrindell, W. A., & Ettema, H. (1981). Dimensionele structuur, betrouwbaarheid en validiteit van de Nederlandse bewerking van de Symptom Check List (SCL-90): Gegevens gebaseerd op een fobische en een `normale' populatie [Dimensional structure, reliability and validity of the Dutch adaptation of the Symptom Check List (SCL-90): Data from a phobic and a `normal' population]. Nederlands Tijdschrift voor de Psychologie, 36, 77-108.

Arrindell, W., Boelens, W., & Lambert, H. (1983). On the psychometric properties of the Maudsley Marital Questionnaire (MMQ): Evaluation of self-ratings in distressed and `normal' volunteer couples based on the Dutch version. Personality and Individual Differences, 4, 293-306.

Beck, J. G. (1995), Hypoactive sexual desire disorder: An overview. Journal of Consulting and Clinical Psychology, 63, 919-927.

Buss, D. M. (1998). Sexual strategies theory: Historical origins and current status. The Journal of Sex Research, 35, 19-31.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences. New York: Academic Press.

Condra, M., Morales, A., Owen, J. A., Surridge, D. H., & Fenemore, J. (1986). Prevalence and significance of tobacco smoking in impotence. Urology, 27, 495-498.

De Amicis, L., Goldberg, D. C., LoPiccolo, J., Friedman, J., & Davies, L. (1984). Three-year follow-up of couples evaluated for sexual dysfunction. Journal of Sex and Marital Therapy, 10, 215-228.

Dow, M. G. T. (1983). A controlled comparative evaluation of conjoint counselling and self-help behavioural treatment for sexual dysfunction. Unpublished doctoral thesis, University of Glasgow, Scotland.

Everaerd, W., & Dekker, J. (1985). Treatment of male sexual dysfunction: Sex therapy compared with systematic desensitization and rational emotive therapy. Behavior Research and Therapy, 23, 13-25.

Ewing, J. A. (1984). Detecting alcoholism, the CAGE questionnaire. Journal of the American Medical Association, 252, 1905-1907.

Gilbert, D. G., Hagen, R. L., & D'Agostino, J. A. (1986). The effects of cigarette smoking on human sexual potency. Addictive Behaviors, 11, 431-434.

Glasgow, R. E., & Rosen, G. M. (1978). Behavioral bibliotherapy: A review of self-help behavior therapy manuals. Psychological Bulletin, 85, 1-23.

Glass, G. V., McGaw, B., & Smith, M. L. (1981). Meta-analysis in social research. Beverly Hills, CA: Sage.

Gould, R. A., & Clum, G. A. (1993). A meta-analysis of self-help treatment approaches. Clinical Psychology Review, 13, 169-186.

Hawton, K. (1995). Treatment of sexual dysfunctions by sex therapy and other approaches. British Journal of Psychiatry, 167, 307-314.

Hawton, K., & Catalan, J. (1986). Prognostic factors in sex therapy. Behavior Research and Therapy, 24, 377-385.

Hawton, K., & Catalan, J. (1990). Sex therapy for vaginismus: Characteristics of couples and treatment outcome. Sexual and Marital Therapy, 5, 39-48.

Hawton, K., Catalan, J., & Fagg, J. (1991). Low sexual desire: Sex therapy results and prognostic factors. Behaviour Research and Therapy, 29, 217-224.

Heiman, J. R., & Meston, C. M. (1997). Empirically validated treatment for sexual dysfunction. In R. C. Rosen, C. M. Davis, & H. J. Ruppel (Eds.), Annual review of sex research; An integrative and interdisciplinary review. Vol. 8 (pp. 148-194). Lake Mills, IA: The Society for the Scientific Study of Sexuality.

Hirshkowitz, M., Karacan, I., Howell, J. W., Arcasoy, M. O., & Williams, R. L. (1992). Nocturnal penile tumescence in cigarette smokers with erectile dysfunction. Urology, 39, 101-107.

Masters, W. H., & Johnson, V. E. (1970). Human sexual inadequacy. Boston: Little, Brown.

Maultsby, M. C. (1975). Help yourself to happiness through rational self-counseling. Boston: Esplanada.

Munjack, D. J., Schlaks, A., Sanchez, V. C., Usigli, R., Zulueta, A., & Leonard, M. (1984). Rational-emotive therapy in the treatment of erectile failure: An initial study. Journal of Sex and Marital Therapy, 10, 170-175.

Norusis, M. J. (1994). SPSS, advanced statistics 6.1. Chicago: SPSS Inc. Quevillon, R. P. (1993). Dyspareunia. In W. O'Donohue & J. H. Geer (Eds.), Handbook of sexual dysfunctions: Assessment and treatment (pp. 367-380). Boston: Allyn & Bacon.

Regan, P. C., & Berscheid, E. (1996). Beliefs about the state, goals, and objects of sexual desire. Journal of Sex and Marital Therapy, 22, 110-120.

Rosen, R. C., & Leiblum, S. R. (1995). Treatment of sexual disorders in the 1990s: An integrated approach. Journal of Consulting and Clinical Psychology, 63, 877-890.

Rust, J., & Golombok, S. (1986). The GRISS: A psychometric instrument for the assessment of sexual dysfunction. Archives of Sexual Behavior, 15, 157-165.

Sarwer, D. B., & Durlak, J. A. (1997). A field trial of the effectiveness of behavioral treatment for sexual dysfunctions. Journal of Sex and Marital Therapy, 23, 87-97.

Segraves, R. T. (1989). Effects of psychotropic drugs on human erection and ejaculation. Archives of General Psychiatry, 46, 275-284.

ter Kuile, M. M., van Lankveld, J. J. D. M., Kalkhoven, P., & van Egmond, M. (1999). The Golombok Rust Inventory Of Sexual Satisfaction (GRISS): Psychometric properties within a Dutch population. Journal of Sex and Marital Therapy, 25, 59-71.

Trudel, G., & Proulx, S. (1987). Treatment of premature ejaculation by bibliotherapy: An experimental study. Sexual and Marital Therapy, 2, 163-167.

van Lankveld, J. J. D. M. (1993). Zelfje seksuele relatie verbeteren [How to improve your sexual relationship]. Cothen, the Netherlands: Servire.

van Lankveld, J. J. D. M. (1998). Bibliotherapy in the treatment of sexual dysfunctions: A meta-analysis, Journal of Consulting and Clinical Psychology, 66, 702-708.

van Lankveld, J. J. D. M., & ter Kuile, M. M. (1999). The Golombok Rust Inventory of Sexual Satisfaction (GRISS): Predictive validity and construct validity in a Dutch Population. Personality and Individual Differences, 26, 1005-1023.

van Lankveld, J. J. D. M., Grotjohann, Y., van Lokven, B. M. E., & Everaerd, W. (1999). Sexual, psychological, psychiatric and marital characteristics of couples applying for bibliotherapy via different recruitment strategies: A multivariate comparison. Journal of Sex and Marital Therapy, 25, 197-209.

Vennix, P. (1983). De ILKS-gedachte en andere operationalisaties [The ILKS-concept and other operationalizations]. Nisso Onderzoeksrapport nr. 37. Zeist, the Netherlands: NISSO.

Wittchen, H. U., Zhao, S., Abelson, J. M., Abelson, J. L., & Kessler, R. C. (1996). Reliability and procedural validity of UM-CIDI DSM-III-R phobic disorders. Psychology in Medicine, 26, 1169-1177.

Wolf, F. M. (1986). Meta-analysis: Quantitative methods for research synthesis. Beverly Hills, CA: Sage.

World Health Organisation. (1990). Composite International Diagnostic Interview. Geneva, Switzerland: Author.

Zeiss, R. A. (1978). Self-directed treatment for premature ejaculation. Journal of Consulting and Clinical Psychology, 46, 1234-1241.

Zilbergeld, B. (1993). The new male sexuality. New York: Bantam Books.

Manuscript accepted November 27, 2000

Jacques J. D. M. van Lankveld Academic Hospital Maastricht, The Netherlands

Walter Everaerd University of Amsterdam, The Netherlands

Yvonne Grotjohann University of Utrecht, The Netherlands

This research was supported by Research Grant no. 4142 from the Netherlands National Foundation of Mental Health.

Address correspondence to J.J.D.M. van Lankveld, Department of Medical, Clinical and Experimental Psychology, Academic Hospital Maastricht, PO Box 5800, 6202 AZ Maastricht, The Netherlands; e-mail: [e-mail].