A Randomized Controlled Trial of Guided Internet-deliveredCognitive Behavioral Therapy for Erectile Dysfunctionjsm_2391 2800..2809

Erik Andersson, MSc,*‡ Christian Walén, MSc,* Jonas Hallberg, MSc,* Björn Paxling, MSc,*Mats Dahlin, MSc,*§ Jonas Almlöv, MSc,* Reidar Källström, MD,¶ Klaas Wijma, PhD,**Per Carlbring, PhD,†† and Gerhard Andersson, PhD*†‡

*Department of Behavioural Sciences and Learning, Linköping University, Sweden; †Swedish Institute for DisabilityResearch, Linköping University, Sweden; ‡Department of Clinical Neuroscience, Division of Psychiatry, KarolinskaInstitutet, Stockholm, Sweden; §Psykologpartners, Linköping, Sweden; ¶Department of Urology, University HospitalLinköping, Linköping, Sweden; **Department of Clinical and Experimental Medicine, Linköping University, Linköping,Sweden; ††Department of Psychology, Umeå University, Umeå, Sweden

DOI: 10.1111/j.1743-6109.2011.02391.x

A B S T R A C T

Introduction. Men with erectile dysfunction are often worried about their condition, have interpersonal difficulties,and have a reduced quality of life. Internet-delivered cognitive behavior therapy (ICBT) has been shown effective fora number of health problems but evidence is limited concerning the treatment of erectile dysfunction.Aim. The study investigated the effects of ICBT for erectile dysfunction.Methods. Seventy-eight men were included in the study and randomized to either ICBT or to a control group,which was an online discussion group. Treatment consisted of a 7-week Web-based program with e-mail-basedtherapist support. Each therapist spent an average of 55 minutes per participant.Main Outcome Measure. The International Index of Erectile Functioning five-item version was administered viathe telephone at pretreatment, post-treatment, and 6 months after receiving ICBT.Results. At post-treatment, the treatment group had significantly greater improvements with regard to erectileperformance compared with the control group. Between-group differences at post-treatment were small (d = 0.1),but increased at the 6-month follow-up (d = 0.88).Conclusions. This study provides support for the use of ICBT as a possible treatment format for erectile dysfunction.Andersson E, Walén C, Hallberg J, Paxling B, Dahlin M, Almlöv J, Källström R, Wijma K, Carlbring P, andAndersson G. A randomized controlled trial of guided Internet-delivered cognitive behavioral therapy forerectile dysfunction. J Sex Med 2011;8:2800–2809.

Key Words. Cognitive Behavior Therapy; Erectile Dysfunction; Internet; Sexual Dysfunction

Introduction

E rectile dysfunction (ED) is defined as a con-tinuously present, or recurrent, inability to

achieve or maintain a sufficient erection inresponse to erotic stimulation [1]. ED is oftenexperienced by the individual as distressing and isfrequently associated with interpersonal difficul-ties [2]. ED is a common problem. Two studieshave shown a symptom prevalence ranging from19% to 22% [3,4]. However, when controlling fordissatisfaction of sex life, this number decreases to6.9% [3]. Another study stated a prevalence of EDrequiring clinical attention averaging 3–7% [5].

Even if methodology and definitions of ED havevaried between different epidemiological studies[6], it is known that the prevalence of ED increasesmarkedly with age [5]. The cause of ED is multi-factorial, i.e., it has both psychological and bio-logical components. Advances in biological modelshave been influential during the last decades [7,8].From a psychological point of view, ED could alsobe caused or worsened by intrusive negativethoughts, increased self-focus, and feeling tense[9,10].

Treatments for ED also mirror the multifacto-rial conceptualization with both medical and psy-chological treatments. Masters and Johnson’s [11]

2800

J Sex Med 2011;8:2800–2809 © 2011 International Society for Sexual Medicine

treatment, often referred to as “sex therapy,” is awell-known psychological treatment of ED[12,13]. The main treatment components involvesensate focus, communication, and relaxationtraining [11]. A recent systematic review [13] con-cluded that psychological treatments have somerelieving effects on erectile difficulties, althoughthe evidence base is limited.

Internet-administrated treatments normallyconsist of a package of comprehensive self-helpmaterial in which the patient receives informa-tion and exercises on a weekly basis [14]. Admin-istration of psychological treatments via theInternet is believed to be cost effective and maybridge geographical distances [15]. Anotherbenefit of this treatment approach is that patientswho avoid seeking help, or who experience shamebecause of their problem, may have the opportu-nity to work with their problem in their homeenvironment [16]. There are a few studies sug-gesting that psychological treatment for ED canbe feasible and efficacious when delivered via theInternet. In a pilot study by Van Diest, VanLankveld, Leusink, Slob, and Gijs [17], anInternet-delivered treatment protocol was testedfor a variety of sexual dysfunctions. In this study(N = 39), 67% of the participants reportedimprovements in sexual functioning, and 47%reported that improvements were maintained at1-month follow-up. The treatment progressedfor 3 months and all therapist communicationwas conducted using an e-mail system. In a sub-sequent study [18], 58 men were randomized toInternet-delivered sex therapy for ED or to awaiting list. The results showed that the treat-ment condition was superior compared with awaiting list control group. A controlled trial byMcCabe and Price [19] reported similar levels ofimprovements between Internet-delivered cogni-tive behavior therapy (ICBT) and ICBT plusPDE5 inhibitors. The treatment (Rekindle) was a10-week program focusing on communicationand sensate focus exercises. One limitation withthis study was the small sample size (N = 12).Overall, the efficacy of ICBT for ED is stillunclear, and more research is needed.

Aim

The aim of this study was to investigate theeffects of ICBT for ED symptoms. The treat-ment was delivered in the form of a 7-week Web-based program with therapist support via e-mail.We expected the ICBT group to make superior

improvements compared with an active controlgroup. We chose an active control group toensure a control for attention and possible alle-viating effects of sharing one’s distress withothers.

The study protocol was approved by theregional ethical review board. Written informedconsent was obtained from all participants.

Method

Participants were recruited from media such asarticles and interviews on TV and in newspapers.Both homosexual and heterosexual persons wereeligible for participation.

To be included in the study, participants had to(i) have access to a computer with an Internetconnection; (ii) be �18 years old; (iii) have a scoreof <21 on International Index of Erectile Func-tioning 5 (IIEF-5) [20]; and (iv) have a stablepartner relationship for �3 months. Participantswere excluded if they had (v) a medical history thatcould jeopardize the subject’s health; (vi) severedepression or anxiety disorders; (vii) alcohol ordrug-related problems in need of immediate treat-ment; (viii) prior medical history that could be aprimary cause of ED (for example, surgery forprostate cancer, diabetes); and (ix) medication usethat could produce ED as a side effect.

As Figure 1 shows, the participants wererecruited in several steps. All interviews wereconducted via telephone. All assessors were psy-chology students in their last year of training.Inclusion was performed in collaboration with aurologist and a clinical psychologist whoreviewed all cases. Careful medical history andinformation on medication use were collectedduring the interviews. We did not conduct anydirect medical examinations but were careful notto include participants who had not been incontact with the health care system. After ran-domization, the treatment group received ICBTfor 7 weeks, and the control group had access toan Internet discussion forum during this time.The participants were allowed to use pro-erectilemedications during treatment. Immediately fol-lowing post-assessment, control group partici-pants were crossed over to ICBT. Primaryoutcome measure was administered via the tele-phone by the trained students with a restrictionthat the participant was not interviewed by hisonline therapist. Secondary outcomes wereassessed using online self-report questionnaires.At a 6-month follow-up, participants in the

Internet-delivered CBT for Erectile Dysfunction 2801

J Sex Med 2011;8:2800–2809

Figure 1 Flowchart.

2802 Andersson et al.

J Sex Med 2011;8:2800–2809

treatment group were assessed on the Internetand via telephone interviews. The control groupwas also assessed 6 months after ICBT comple-tion, but with telephone interviews only.

MeasuresPrimary OutcomePrimary outcome measure was the IIEF-5. Themain argument to use the IIEF-5 in this study wasthat it is easy to use and should be considered to bereliable when administered by a clinician. A cutoffvalue of 21 has been found to discriminatebetween diagnosis/no diagnosis [20]. In thepresent study, the IIEF-5 was presented over thetelephone, which has been done previously in epi-demiological research [21]. ED severity wasassessed using clinical guidelines from a previousstudy [22].

Secondary OutcomesThe IIEF is a 15-item questionnaire that mea-sures erectile functioning. Symptom dimensionsare (i) erectile functioning; (ii) intercourse satis-faction; (iii) orgasmic functioning; (iv) sexualdesire; and (v) overall satisfaction [23]. We alsoincluded the relationship assessment scale (RAS),a seven-item generic measure of relationship sat-isfaction. The RAS has shown good test–retestreliability and correlates fairly strong with otherRASs. The RAS has also shown good psychomet-ric properties across samples of ethnically diverseand age-diverse couples [24]. Symptoms ofgeneral anxiety were measured using the Beckanxiety inventory (BAI), which is a widespreadself-report measure that assesses the frequency of21 different symptoms of anxiety measured withanswer alternatives between 0 and 3 [25].Depressive symptoms were assessed by the Beckdepression inventory II (BDI-II) [26], whichincludes 21 items. Both the BAI and the BDI-IIhave good psychometric properties. Quality oflife was measured by the World Health Organi-zation Quality of Life-Brief Version(WHOQOL-BREF). This is a 26-item self-report questionnaire that yields an estimate of theparticipant’s overall quality of life. It measuresquality of life in four different domains: (i) psy-chological well-being; (ii) quality of social rela-tionships; (iii) physical health; and (iv)environmental satisfaction. The questionnaire isdesigned to be cross-cultural [27]. Finally, weused the clinical global impression scale (CGI-I;[28]) to measure global improvement. All second-

ary outcomes were administered via the Internetexcept for the CGI-I, which was administered viatelephone.

TreatmentThe treatment consisted of seven weekly modulesdelivered via the Internet using a secure Web-based contact-handling system (e.g., like Internetbanking). Participants were given consecutiveaccess to the treatment modules by completingand reporting a homework assignment to histherapist. The patient could contact the therapistat any time and expect a reply within 24 hours.Patient and therapist had no face-to-face or tele-phone contact during the treatment. In module 1,ED was presented from a biological, psychologi-cal, and social perspective and information wasgiven as to how various factors could be contribut-ing to the maintenance of the problem. In module2, the participant learned about the cognitivebehavioral therapy (CBT) perspective (operant andrespondent conditioning) and its relation to EDsymptoms. Module 3 consisted of relationship exer-cises where the participant was instructed to discussthe implications of the disorder with his partnerand invited the couple to attempt to conceptualizethe disorder from a CBT perspective together.Modules 4–6 consisted of exposure tasks thatengaged the participant in various structured exer-cises, designed to gradually expose them to sexualinteraction with their partner. The participant con-tinuously rated his level of anxiety and other mea-sures as he worked through each task. Modules 4–6also featured relaxation exercises and strategies formanaging intrusive thoughts. The last modulecomprised a relapse prevention program instruct-ing the participant about various strategies he couldutilize, alone or with his partner, to maintain anygains from the treatment. The therapists were allstudents in clinical psychology in their final year oftraining. Their main function was to provide thera-peutic support through clarifying information,reviewing progress, and giving the participantsfeedback on homework assignments. The thera-pists received weekly supervision meetings by alicensed psychologist.

This trial used an active control group, inwhich the participants had access to an onlinediscussion forum. In this forum, participantscould send messages anonymously to each otherfor the duration of the treatment (7 weeks). Theywere also encouraged to be active and discussiontopics were posted on the forum on a weeklybasis by the therapists to encourage activity.

Internet-delivered CBT for Erectile Dysfunction 2803

J Sex Med 2011;8:2800–2809

Online discussion forums have been shown to beeffective in reducing depressive symptoms andanxiety in persons with depression and breastcancer [29,30]. Thus, the rationale for employingthis design was to ensure a control for attentionand possible alleviating effects of sharing one’sdistress with others. One of the therapists in thestudy monitored the discussion forum on a dailybasis and participants were encouraged to discusstheir erectile difficulties and helpful ways ofcoping with it. Participation in the discussionforum was anonymous and participants used astudy code and not their own name.

Statistical AnalysisThe continuous outcome variables were analyzedusing a mixed-effects model. The mixed-effectsmodel combines two different levels to a singleframework. Level 1 (repeated measurements) isnested within level 2 (individuals) to form asingle framework with different covariance struc-tures at each level. Thus, mixed-effects modelstake into account independent random effects inthe model as well as nonindependence of longi-tudinal data. Mixed-effects models have theadvantage of accounting for correlations betweenrepeated measurements for each subject. In addi-tion, it is considered to be more flexible regard-ing time effects and handles missing data betterthan a standard analysis of variance [31]. Themodeling approach utilized all available datafrom all participants and measurement points,which made this an intent-to-treat analysis. Weused the mixed-effects model to examine the dif-ference in rates of change between treatmentgroup and control condition over two measure-ment points (i.e., pretreatment and post-treatment). We conducted paired t-tests amongpretreatment, post-treatment, and the 6-month

follow-up. Paired t-tests were also performed onthe control group, comparing scores before andafter the control period, as well as after receivingICBT. For ordinal scale variables, the Mann–Whitney U-test was used to investigate outcomescomparing independent samples. Within-groupand between-group effect sizes were estimatedusing Cohen’s d. Regression analysis was alsoconducted with treatment response as dependentvariable and number of completed modules asindependent variable. This was done to see ifthere was a relationship between erectileimprovement and number of completed modules.The sample size was based on a power calculationassuming that there was a chance higher than80% to detect a difference, given an effect size of0.5 and an alpha level of 0.05.

Results

Patient Characteristics and AdherenceTable 1 displays the sample characteristics for par-ticipants in the treatment group as well as in thecontrol group. All participants were involved inheterosexual relationships. An independent t-testshowed no differences between the treatmentgroup and the control group regarding the meanIIEF-5 pretreatment score, t76 = -0.645; P = 0.52.The number of completed modules was low in thetreatment group with 54% of the participants onlyreaching modules 1–4 (see Table 2 for an overviewof adherence to the protocol). The averagenumber of completed modules was 3.9 (standarddeviation [SD] = 1.88). Throughout the entiretreatment, each therapist spent an average of 55minutes (SD = 28) per participant. Activities in thediscussion group varied with some men beingmore active, but most posting at least one message.

Table 1 Characteristics of the participants

Treatment group Control groupSignificant(N = 37) (N = 39)

Age in years: mean (SD) 57.62 (10.06) 55.50 (9.94) NSMin-max years 23–74 37–79IIEF-5 mean scores 12.44 (5.12) 13.13 (4.89) NSErectile dysfunction

Mild 30% 31% NSMild to moderate 32% 46% NSModerate 16% 8% NSSevere 22% 15% NS

Duration of erectile dysfunctionBetween 3 months and 1 year 24% 6% P < 0.05More than 1 year 76% 93% NS

SD, standard deviation; IIEF-5, International Index of Erectile Functioning 5; NS, nonsignificant.

2804 Andersson et al.

J Sex Med 2011;8:2800–2809

Treatment OutcomeThe mixed-effects model revealed a significantinteraction effect (group ¥ time) at post-treatmentfor the primary outcome measure (IIEF-5), F1, 67

= 3.92; P = 0.05, indicating superior improvementof the ICBT group compared with the controlgroup. A paired t-test also showed a significant

improvement from post-treatment to the 6-monthfollow-up for the treatment group, t29 = -3.00;P < 0.01. A Mann–Whitney U-test showed a sig-nificance between group effects on the CGI(P < 0.001) in favor of the treatment group(Figure 2). No other interaction effects were foundon any of the secondary outcome measures. Meansand SDs for continuous variables are presented inTable 3.

The within-group effect sizes on primaryoutcome at post-treatment were small for bothtreatment group (d = 0.42) and control group(d = 0.20). The between-group effect size wassmall (d = 0.10) at post-treatment, but when com-paring the 6-month follow-up data, the effect sizehad increased (d = 0.88) on this measure. A pairedt-test revealed a trend toward improvement in thecontrol group between pre and post, t34 = -1.82;

Figure 2 Clinical global impressionscale improvements.

Table 2 Treatment adherence

Module Frequency Percent

1 37 1002 30 813 28 764 24 655 17 466 7 197 3 8

Note: Participants were given consecutive access to the treatment modules.

Table 3 Means (SDs) at pretreatment, post-treatment, and 6-month follow-up

Measure

Treatment group Control group

Group ¥ time

Pre (N = 37) Post (N = 33) FU (N = 30) Pre (N = 39) Post (N = 35)

M SD M SD M SD M SD M SD

IIEF-5 12.44* (5.12) 14.58† (5.02) 17.32‡ (5.55) 13.13§ (4.89) 14.11§ (4.91) SignificantIIEFErectile functioning 13.68* (5.96) 16.97† (7.81) 18.68† (9.33) 13.56§ (6.25) 16.56¶ (7.32) NSIntercourse satisfaction 6.68* (4.54) 7.41*,† (4.78) 8.81† (4.60) 6.28§ (3.80) 7.28¶ (4.47) NSOrgasmic functioning 7.70* (2.71) 7.82* (2.96) 7.89* (2.99) 7.69§ (2.72) 7.67§ (2.54) NSSexual desire 7.59* (1.62) 7.35* (1.82) 7.33* (1.96) 7.03§ (1.87) 7.06§ (1.80) NSOverall satisfaction 5.08* (2.25) 6.48† (2.11) 7.19† (1.81) 4.49§ (2.16) 5.42¶ (2.48) NSRAS 27.14* (5.57) 27.53* (5.61) 26.11* (6.64) 26.26§ (5.3) 25.56§ (5.64) NSBDI 5.86* (4.48) 4.91* (4.16) 5.07* (5.80) 7.26§ (6.86) 6.31§ (5.74) NSBAI 3.95* (4.25) 3.21* (3.15) 3.67* (4.66) 4.72§ (4.87) 4.03§ (4.67) NSWHOQOL-BREFPhysical health 77.95* (12.51) 77.84* (15.25) 75.96* (17.32) 77.03§ (16.88 76.43§ (16.37) NSPsychological 72.89* (11.34) 74.14* (13.52) 73.85* (16.42) 66.15§ (15.20) 68.92§ (14.20) NSSocial relationships 55.76* (14.83) 62.75† (14.92) 63.15† (19.73) 52.08§ (14.77) 52.14§ (13.61) NSEnvironment 80.35* (11.24) 78.58* (12.55) 77.41* (14.06) 79.51§ (10.45) 77.68§ (10.69) NS

Note: Means that share superscripts do not differ significantly (P < 0.05) using within-group t-tests.SD, standard deviation; FU, follow-up; M, mean; IIEF-5, International Index of Erectile Functioning 5; IIEF, International Index of Erectile Functioning; RAS,relationship assessment scale; BDI, Beck depression inventory; BAI, Beck anxiety inventory; WHOQOL-BREF, World Health Organization Quality of Life-BriefVersion; NS, nonsignificant.

Internet-delivered CBT for Erectile Dysfunction 2805

J Sex Med 2011;8:2800–2809

P = 0.08. After the control group received ICBT,within-group effects were significant on theIIEF-5, t25 = -3.25; P < 0.01, and CGI, Z = -2.57;P = 0.01. Thus, after treatment, the control groupmade similar improvements as the treatmentgroup (Figure 3).

Categorical Outcomes and the Role of AdherenceAt post-treatment, four participants in the treat-ment group had an IIEF-5 score �21, indicatingthat they no longer fulfilled the diagnostic criteriafor ED. In the control group, all participants stillhad a diagnosis of ED at this time of measurement.This difference was significant, c2 (1) = 4.45;P = 0.04. At the 6-month follow-up, 10 partici-pants (27%) in the treatment group fell outside thediagnostic range of ED according to the IIEF-5cutoff score. Finally, a regression analysis was per-formed using a number of completed modules(1–7) as independent variable and IIEF-5 changescores as dependent variable. A significant associa-tion was found between these two variables,F1, 31 = 6.86; P = 0.01, indicating that a highernumber of completed modules was associated withgreater improvements on the IIEF-5.

Discussion

The results of the present study suggest that ICBTprovides some improvements in erectile function-ing. Significant effects on the main outcomemeasures were found for the treatment groupcompared with the active control group when con-trolling for data loss. The between-group effectsizes were small at post-treatment but hadincreased at the 6-month follow-up. In addition,the control group showed similar improvementsafter receiving ICBT. However, no significant

effects could be found on any of the secondaryoutcomes and the treatment results should there-fore be interpreted with caution.

Pretreatment BAI and BDI scores were low.The nonsignificant results regarding depressionand anxiety could therefore be explained by flooreffects and/or by the fact that the study was under-powered to detect small BAI and BDI changescores. Low change scores on the RAS andWHOQOL-BREF could also possibly be theresult of relatively high relationship satisfactionand quality of life at pretreatment. One study [32]has shown that older men with ED have higherquality of life compared with healthy controls.Thus, it is reasonable to expect some ceiling effectsregarding these secondary outcomes. One possiblereason for the nonsignificant findings on the IIEFcould be that we administered it as a self-ratingscale and it is possible that the participants didnot fully understand the questions. One majordifference between the primary and secondaryoutcomes was that the former was administeredusing trained professionals whereas all secondaryoutcome measurements were self-administered onthe Internet. Thus, the mean difference betweenprimary and secondary outcomes could beexplained by the administration format.

We found significant differences in duration ofED at pretreatment. However, the screening ques-tions did not specifically ask for how long partici-pants had had ED problems on a continuous scale.Instead, the only distinction made was betweenmore or less than 1 year of difficulties. As the twogroups were equal regarding anxiety, depressivesymptoms, quality of life, and ED severity, we viewit as unlikely that the potential difference in termsof ED duration is indicative of general skewness interms of symptom burden.

Figure 3 Mean International Index of Erectile Functioning 5 score between the two treatment conditions.Note: Post 1 is the post-treatment data for the Internet-delivered cognitive behavior therapy (ICBT) group (treatment) and thedata for the post-attention intervention for the control group. Post 2 is the post-treatment data for the initial control group afterit had the ICBT intervention. FUP 1 and FUP 2 are the data from the 6-month follow-up for both groups after receiving the ICBT.

2806 Andersson et al.

J Sex Med 2011;8:2800–2809

The number of completed modules was low,which possibly influenced the overall treatmentresponse. At post-treatment, only three partici-pants (8%) had reached the final module and54% was still between modules 1–4. We investi-gated the role of treatment adherence and aregression analysis showed a higher treatmentresponse for each module completed, thus indi-cating a dose–response relationship betweennumber of completed modules and treatmentsuccess. As research has pointed out previously,both medical and psychological treatments forED have been associated with high rates of dis-continuation [33]. McCabe, Price, and Piter [34]have reported dropout rates as high as 70% intheir ICBT program Rekindle, which was partlyexplained by limited motivation of the treatmentparticipants. Many of the participants in thepresent study reported lack of time as the mainreason for dropping out. Other reasons for dis-continuing were related to the participant’s part-ners; they were unwilling to participate in gradedexposures or communication exercises and attimes saw the treatment components as unsuit-able for them. These reasons were also given bydropouts in the Rekindle study [34]. The fact thatthis study only progressed for 7 weeks could alsohave affected the high dropout rates. Perhaps alonger treatment period could increase thenumber of completed modules at post-treatment.One possible explanation for the larger effect atfollow-up could be that achieving improvement iscontingent upon repeated use of the exercises inthe program. This suggests that ICBT for EDmight benefit from a longer treatment periodthan the 7 weeks employed in the present study.

There are few randomized controlled trialsthat have been conducted in the field of Internet-based psychological interventions for ED. Thefew trials that have been carried out have oftenbeen limited by small sample sizes. This studyused a randomized controlled trial design andhad reasonable power to detect between-groupdifferences. Another advantage in this study wasthat the primary outcome was administered by aclinician.

This study has several limitations. One meth-odological limitation is the lack of experimentalcontrol regarding the 6-month follow-up data,which in this study were compared with thecontrol group post-treatment data. We thereforelost experimental control for spontaneousimprovements at the 6-month follow-up. Theo-retically, then, the between-group effects may be

caused by other factors than ICBT. However, sincethe IIEF-5 includes measurement of erectile diffi-culties during the last 6 months, it could be some-what biased to only use post-treatment data, whichpartly cover a time period before treatment actu-ally began. Another limitation of the study is theinability to control for and include all medicalaspects of the sample participants. The inclusioncriteria relied on self-report measures and theurologist in charge did not perform any physicalexamination of the participants. As a consequence,it is possible that we failed to detect possible caseswhere the symptoms could be better explained bymedical conditions unknown for the participants(because we relied on self-report). In addition, wedid not have young age or significant psychologicaldistress as an inclusion criterion, which would haveyielded a smaller group but potentially moreresponsive to psychological treatment. Theseaspects could partly explain the low effect sizes andtherefore, the results should be interpreted withcaution. One final limitation is the choice ofIIEF-5 as primary outcome measure. Questionshave been raised on limited applicability regardingthis questionnaire and that it does not provide afull clinical picture of the ED diagnosis [35]. Thechoice of a short questionnaire could have affectedthe results.

A recommendation for further studies is to dif-ferentiate among different age groups to have abroader ED sample. The mean participant in thissample was 57 years of age, and it would be inter-esting to investigate treatment response in ayounger sample. Another recommendation wouldbe to utilize an experimental design with a higherlevel of control for medication intake. In this treat-ment, the participants had access to psychologicaltreatment but were allowed to use ED medicationsimultaneously. It is unclear if medications havehad some influence on the outcome. Results froma previous study did not show any significant dif-ferences between ICBT alone or in combinationwith PDE5 inhibitors [19], but more research isneeded. Finally, greater effect sizes were observedat 6-month follow-up after receiving ICBT. It istherefore recommended that future studies uselonger experimental time periods.

Conclusions

The results suggest that ICBT could be a possibletreatment alternative for ED. Despite its limita-tions, the study adds valuable knowledge about thetreatment of ED, both regarding the method of

Internet-delivered CBT for Erectile Dysfunction 2807

J Sex Med 2011;8:2800–2809

delivery and contents of treatment. Still, furtherstudies of ICBT for ED are warranted.

Acknowledgments

This study was sponsored in part by a grant to ProfessorAndersson from Linköping University (professor’scontract).

Corresponding Author: Erik Andersson, MSc,Department of Clinical Neuroscience, Karolinska Insti-tutet, M46 Huddinge sjukhus, Stockholm 14186,Sweden. Tel: +46(0)736716335; Fax: +46(0)8-779 54 16;E-mail: erik.m.andersson@ki.se

Conflict of Interest: None.

Statement of Authorship

Category 1(a) Conception and Design

Erik Andersson; Christian Walén; Jonas Hallberg;Gerhard Andersson; Per Carlbring; ReidarKällström

(b) Acquisition of DataErik Andersson; Christian Walén; Jonas Hallberg;Jonas Almlöv; Björn Paxling; Mats Dahlin; KlaasWijma

(c) Analysis and Interpretation of DataErik Andersson; Christian Walén; Jonas Hallberg;Gerhard Andersson; Per Carlbring; Björn Paxling;Mats Dahlin; Reidar Källström; Jonas Almlöv;Klaas Wijma

Category 2(a) Drafting the Article

Erik Andersson; Christian Walén; Jonas Hallberg;Jonas Almlöv; Mats Dahlin; Klaas Wijma; BjörnPaxling

(b) Revising It for Intellectual ContentErik Andersson; Christian Walén; Jonas Hallberg;Gerhard Andersson; Per Carlbring; ReidarKällström

Category 3(a) Final Approval of the Completed Article

Erik Andersson; Christian Walén; Jonas Hallberg;Gerhard Andersson; Per Carlbring; Björn Paxling;Mats Dahlin; Reidar Källström; Jonas Almlöv; KlaasWijma

References

1 American Psychiatric Association. Diagnostic and statisticalmanual of mental disorders: DSM-IV-TR. 4th edition. Wash-ington, DC: American Psychiatric Association; 2000.

2 Porst H, Vardi Y, Akkus E, Melman A, Park NC, Seftel AD,Teloken C, Wyllie M. Standards for clinical trials in malesexual dysfunctions. J Sex Med 2010;7(1 Pt 2):414–44.

3 Braun M, Wassmer G, Klotz T, Reifenrath B, Mathers M,Engelmann U. Epidemiology of erectile dysfunction: Resultsof the “Cologne Male Survey”. Int J Impot Res 2000;12:305–11.

4 Laumann EO, Paik A, Rosen RC. The epidemiology oferectile dysfunction: Results from the National Health andSocial Life Survey. Int J Impot Res 1999;11(Suppl 1):S60–4.

5 Kubin M, Wagner G, Fugl-Meyer AR. Epidemiology of erec-tile dysfunction. Int J Impot Res 2003;15:63–71.

6 Segraves RT. Considerations for diagnostic criteria forerectile dysfunction in DSM V. J Sex Med 2010;7(2 Pt 1):654–60.

7 Pallas J, Levine SB, Althof SE, Risen CB. A study using Viagrain a mental health practice. J Sex Marital Ther 2000;26:41–50.

8 Althof S. Sex therapy in the age of pharmacotherapy. AnnuRev Sex Res 2006;17:116–32.

9 Barlow DH, Durand VM. Abnormal psychology: An integra-tive approach. Pacific Grove, CA: Brooks/Cole Pub. Co.;1995.

10 Nobre PJ. Psychological determinants of erectile dysfunction:Testing a cognitive-emotional model. J Sex Med 2010;7(4 Pt1):1429–37.

11 Masters WH, Johnson VE. Human sexual inadequacy.London: Churchill; 1970.

12 van Lankveld J. Self-help therapies for sexual dysfunction. JSex Res 2009;46:143–55.

13 Melnik T, Soares BG, Nasello AG. The effectiveness of psy-chological interventions for the treatment of erectile dysfunc-tion: Systematic review and meta-analysis, includingcomparisons to sildenafil treatment, intracavernosal injection,and vacuum devices. J Sex Med 2008;5:2562–74.

14 Andersson G. Using the Internet to provide cognitive behav-iour therapy. Behav Res Ther 2009;47:175–80.

15 Barak A, Klein B, Proudfoot JG. Defining Internet-supported therapeutic interventions. Ann Behav Med 2009;38:4–17.

16 Marks IM, Cavanagh K, Gega L. Computer-aided psycho-therapy: Revolution or bubble? Br J Psychiatry 2007;191:471–3.

17 Van Diest SL, Van Lankveld JJ, Leusink PM, Slob AK, Gijs L.Sex therapy through the Internet for men with sexual dysfunc-tions: A pilot study. J Sex Marital Ther 2007;33:115–33.

18 van Lankveld JJ, Leusink P, van Diest S, Gijs L, Slob AK.Internet-based brief sex therapy for heterosexual men withsexual dysfunctions: A randomized controlled pilot trial. J SexMed 2009;6:2224–36.

19 McCabe M, Price E. Internet-based psychological and oralmedical treatment compared to psychological treatment alonefor ED. J Sex Med 2008;5:2338–46.

20 Rosen RC, Cappelleri JC, Smith MD, Lipsky J, Peña BM.Development and evaluation of an abridged, 5-item version ofthe International Index of Erectile Function (IIEF-5) as a diag-nostic tool for erectile dysfunction. Int J Impot Res 1999;11:319–26.

21 Wu CJ, Hsieh JT, Lin JS, Hwang TI, Jiann BP, Huang ST,Wang CJ, Lee SS, Chiang HS, Chen KK, Lin HD. Compari-son of prevalence between self-reported erectile dysfunctionand erectile dysfunction as defined by five-item InternationalIndex of Erectile Function in Taiwanese men older than 40years. Urology 2007;69:743–7.

22 Rhoden EL, Telöken C, Sogari PR, Vargas Souto CA. The useof the simplified International Index of Erectile Function(IIEF-5) as a diagnostic tool to study the prevalence of erectiledysfunction. Int J Impot Res 2002;14:245–50.

2808 Andersson et al.

J Sex Med 2011;8:2800–2809

23 Rosen RC, Riley A, Wagner G, Osterloh IH, Kirkpatrick J,Mishra A. The international index of erectile function (IIEF):A multidimensional scale for assessment of erectile dysfunc-tion. Urology 1997;49:822–30.

24 Hendrick SS, Dicke A, Hendrick C. The relationship assess-ment scale. J Soc Pers Relat 1998;15:137–42.

25 Beck AT, Epstein N, Brown G, Steer RA. An inventory formeasuring clinical anxiety: Psychometric properties. J ConsultClin Psychol 1988;56:893–7.

26 Beck AT, Steer RA, Ball R, Ranieri W. Comparison of BeckDepression Inventories -IA and -II in psychiatric outpatients. JPers Assess 1996;67:588–97.

27 Skevington SM, Lotfy M, O’Connell KA. The World HealthOrganization’s WHOQOL-BREF quality of life assessment:Psychometric properties and results of the international fieldtrial. A report from the WHOQOL group. Qual Life Res2004;13:299–310.

28 Guy W, ed. Clinical global impressions. Rockville: US Depart-ment of Health and Human Services; 1976.

29 Winzelberg AJ, Classen C, Alpers GW, Roberts H, KoopmanC, Adams RE, Ernst H, Dev P, Taylor CB. Evaluation of an

Internet support group for women with primary breast cancer.Cancer 2003;97:1164–73.

30 Houston TK, Cooper LA, Ford DE. Internet support groupsfor depression: A 1-year prospective cohort study. Am J Psy-chiatry 2002;159:2062–8.

31 Gueorguieva R, Krystal JH. Move over ANOVA: Progress inanalyzing repeated-measures data and its reflection in paperspublished in the Archives of General Psychiatry. Arch GenPsychiatry 2004;61:310–7.

32 Guest JF, Das Gupta R. Health-related quality of life in aUK-based population of men with erectile dysfunction. Phar-macoeconomics 2002;20:109–17.

33 McCabe MP, Price E. Attrition from an Internet-based psy-chological intervention for erectile dysfunction: Who is likelyto drop out? J Sex Marital Ther 2009;35:391–401.

34 McCabe MP, Price E, Piterman L, Lording D. Evaluationof an Internet-based psychological intervention for the treat-ment of erectile dysfunction. Int J Impot Res 2008;20:324–30.

35 Vroege JA. The sexual health inventory for men (IIEF-5). IntJ Impot Res 1999;11:177.

Internet-delivered CBT for Erectile Dysfunction 2809

J Sex Med 2011;8:2800–2809