Randomized Controlled Trial of Internet-Based Cognitive Behavior Therapy for Distress Associated With Tinnitus

Randomized Controlled Trial of Internet-Based Cognitive Behavior Therapy for Distress Associated With Tinnitus GERHARD ANDERSSON, PHD, TRYGGVE STRÖMGREN, MSC, LARS STRÖM, MSC, AND LEIF LYTTKENS, MD, PHD Objective: The aim of this study was to investigate if cognitive behavior therapy (CBT) provided via the Internet results in significant decreases of distress in individuals with tinnitus. Methods: Participants were recruited through Web pages and newspaper articles and thereafter randomly allocated to a CBT self-help manual in six modules or to a waiting-list control group (WLC). All treatment and contact with participants were conducted via the Internet with Web pages and E-mail correspondence. Participants were 117 individuals with tinnitus of duration of more than 6 months. In the first randomized controlled phase of the study, 26 completed all stages of treatment (51% dropout), and 64 of the WLC group completed measures. At 1-year follow-up, all participants had been offered the program and 96 provided outcome measures (18% dropout rate from baseline). Tinnitus-related problems were assessed before and after treatment and at the 1-year follow-up. Daily diary ratings were included for 1 week before and 1 week following the treatment period. Results: Tinnitus-related distress, depression, and diary ratings of annoyance decreased significantly. Immediately following the randomized controlled phase (with a WLC), significantly more participants in the treatment group showed an improvement of at least 50% on the Tinnitus Reaction Questionnaire. At the uncontrolled follow-up, 27 (31%) of all participants had achieved a clinically significant improvement. Conclusions: CBT via the Internet can help individuals decrease annoyance associated with tinnitus. High dropout rates or delay in completing treatment can be a characteristic of treatment studies using the Internet but should be contrasted with the cost effectiveness and accessibility of the Internet. Key words: Internet, self-help, relaxation, cognitive-behavior therapy. ASI Anxiety Sensitivity Index; CBT cognitive behavior therapy; CI confidence interval; HADS Hospital Anxiety and Depression Scale; TRQ Tinnitus Reaction Questionnaire; VAS Visual Analogue Scale; WLC waiting-list control. INTRODUCTION Tinnitus is the perception of sound in the absence of any external or electrical stimulation and often consists of ringing or buzzing sounds (1). It can be experienced as extremely disturbing and disruptive to everyday activities and sleep (2) and is reported by 10 14% of the adult population (point prevalence) (3). Tinnitus is hence a common complaint, but for a majority of people with tinnitus, it is not a source of great distress. Still, for approximately one third of the adult population with chronic tinnitus, it results in significant psychological suffering, including depression, anxiety, concentration difficulties, and other related difficulties such as noise sensitivity and hearing impairment (2). Tinnitus often requires specialist consultation, and longitudinal data suggest that the rate of spontaneous improvement is From the Department of Audiology, University Hospital (G.A., L.L.), Uppsala, Sweden; and the Department of Psychology, Uppsala University (G.A., T.S., L.S.), Uppsala, Sweden. Address reprint requests to: Gerhard Andersson, Department of Psychology, Uppsala University, Box 12 25, SE-751 42 Uppsala, Sweden. Email: Gerhard.Andersson@psyk.uu.se Received for publication March 13, 2001; revision received October 24, 2001. DOI: 10.1097/01.PSY.0000031577.42041.F8 small (3). Comorbidity with anxiety and depression is not well investigated in representative samples, but studies on clinical samples suggest that tinnitus patients have diagnosable depression and anxiety to a greater extent than the normal population (2). For most forms of tinnitus, being associated with sensorineural hearing loss (4), no cure can be found, and treatment is aimed at ameliorating the suffering caused by tinnitus. A recent theory has advocated stimulation with sound (white-noise generators) slightly below the loudness of tinnitus as a relief, but empirical evidence is lacking (5). Other treatment alternatives include antidepressants for selected patients with major depression (6), but apart from that, no medical treatment has been found to be effective (6). In sum, the outcome of various treatment attempts is largely negative when the aim is to silence the tinnitus (6). There is, however, evidence to suggest that psychological treatment and proper medical management can decrease the annoyance caused by tinnitus (4). In particular, cognitive behavior therapy (CBT), originally developed and evaluated with affective and anxiety disorders (7), has been found to convey coping strategies and relief for tinnitus sufferers (8). A recent meta-analysis summarized the results from eight small randomized studies and provided support for the use of CBT, in particular, for decreasing the annoyance caused by tinnitus (9). Access to the Internet is increasing (10), and the Internet is often used by patients to access health-related information (11). It is also becoming an accepted medium for interaction between physicians and patients (12). Most likely, the Internet will change the way health care is provided in the future, and hence there is an urgent need to evaluate the pros and cons of Internetadministrated treatment. Given the novelty of the me- 810 Psychosomatic Medicine 64:810 816 (2002) 0033-3174/02/6405-0810 Copyright 2002 by the American Psychosomatic Society

INTERNET TREATMENT OF TINNITUS dium, there is a number of issues exist that evolve when conducting treatment studies via the Internet, such as methods of recruitment, participant characteristics, and on-line behavior, that might differ from face-to-face communication (13). In a previous randomized controlled study, the effects of Internet-based CBT for headache was evaluated, showing decreased frequency of headache (14). In this trial, we investigated the effects of CBT provided via the Internet on distress related to tinnitus. The whole study was conducted via the Internet, from recruiting participants, screening, providing the treatment, and collecting the outcome data. The novel aspects of the study were that neither self-help nor any Internetbased treatment has been tested for tinnitus. The following hypotheses were tested: a) that the cognitive behavior therapy would be superior to a waiting-list control condition in reducing tinnitus-related distress, b) that depression and anxiety would decrease as a result of treatment, and c) that a proportion of participants would reduce their tinnitus distress by 50%, yielding considerable clinical improvement. METHODS Procedure Participants were recruited by means of newspaper articles in Swedish national and regional papers and on the Web page of the Swedish Hard of Hearing Association. Web pages for the study were created, including general information about tinnitus, an outline of the study, information regarding consent from the local ethics committee, and an application form (informed consent). Potentially eligible participants applied for participation by completing a screening questionnaire based on a structured interview for tinnitus history and complaints (15). Further, the screening questionnaire contained questions regarding overall health status and informed consent. Participants had to fulfill the following criteria: duration of tinnitus of at least 6 months, having seen a general practitioner (or ear, nose, and throat physician) on account of tinnitus, age between 18 and 70 years, and tinnitus a severe problem. Given the nature of recruiting participants, it was not possible to obtain audiograms or measures of the tinnitus characteristics by means of tinnitus matchings. The latter have, however, questionable clinical utility (1). Although tinnitus is, with a few rare exceptions, a nonmalignant symptom of hearing loss (4), the study only included participants who had been in contact with a physician because of their tinnitus. All participants had access to computer, modem, and an Internet Fig. 1. Trial profile. Psychosomatic Medicine 64:810 816 (2002) 811

G. ANDERSSON et al. connection and could print out the training instructions. Because all participants were Swedish citizens, all information and treatment was presented in Swedish. All registration forms and rating scales were converted into homepages with two different degrees of protection. General information about tinnitus, specific information about the study, and an application form were accessible to all users of the World Wide Web. Rating scales, as well as all other forms, were only accessible with a password, given by E-mail to the participants. In cases when the Internet failed to work or when participants had problems with their connection, the possibility of contacting the experimenters over the phone was given. This only occurred on a few occasions and was related to technical matters. After screening, participants were randomized to either treatment or the waiting-list control group (Figure 1). Randomization was not stratified and was based on random numbers. Allocation was carried out before any personal E-mail contact with the participants. Waiting-list group participants were informed that they had been randomized to a waiting-list condition and were offered the program later. The first phase will be referred to as the randomized controlled phase and the second 1-year follow-up will be referred to as the uncontrolled follow-up phase. Given the novelty of the medium and the lack of comparable studies, it was difficult to predict the number of participants needed to establish an effect. However, initially, we estimated that a sample size of 64 would have 80% power to detect a mean difference of 11 (SD 22) on the main outcome measure (the Tinnitus Reaction Questionnaire) (16) at a conventional 5% significance level. Assessments Standard self-report measures were used to assess tinnitus-related problems and psychological complaints as primary outcome measures. These assessments were carried out via the Internet at baseline, following treatment, and at 1-year follow-up. Tinnitus problems were assessed on the Tinnitus Reaction Questionnaire (TRQ), which consists of 26 items validated for use in tinnitus treatment research (16). Wilson et al. (16) reported an internal consistency of.96, a test-retest correlation of r.88, and a mean score of 32.5 (SD 22). The TRQ has been translated and validated in the Swedish language, with a Cronbach alpha of.97 and a mean score of 34.5 (SD 23.9) (17). Depression and anxiety were assessed on the 14-item Hospital Anxiety and Depression scale (HADS), which was divided into its two subscales, HADS-anxiety (HADS-A) and HADSdepression (HADS-D) (18). The Swedish version of this scale has been used in several previous studies (19). A population-based study found a mean score of 4.55 (SD 3.73) on the HADS-A and a mean score of 3.98 (SD 3.46) on the HADS-D, with corresponding Cronbach alphas of.82 and.90, respectively (20). This study also reported the percentages of probable cases of anxiety or depressive disorders using a cut-off of 11 points on each of the HADS subscales suggested by Zigmond and Snaith (18). For anxiety it was 8% and depression 6% (20). Fear of anxiety-related somatic sensations was measured with the Anxiety Sensitivity Index (ASI), a questionnaire containing 16 items (21). The scale has been translated into Swedish and, in a sample of tinnitus patients, a Cronbach alpha of.92 and a mean score of 17.5 (SD 12.4) were found (22). Ongoing research on the use of Web-administered questionnaires suggests that similar psychometric properties are found as when using paper-and-pencil test versions (23). In addition, participants completed daily registrations on the computer 1 week before treatment and 1 week immediately after treatment on visual analogue scales (with fixed response alternative 0 10), rating tinnitus loudness, annoyance caused by tinnitus, control over tinnitus, and quality of sleep. Cognitive Behavioral Self-Help Treatment A self-help manual was constructed following cognitive behavioral principles (7). In brief, cognitive behavior therapy is a relatively brief psychological treatment approach directed at identifying and modifying maladaptive behaviors and cognitions by means of behavior change and cognitive restructuring. The self-help manual was derived from previous studies on cognitive behavioral treatment of tinnitus (2, 8) and included 10 components presented in six modules on a weekly basis for 6 weeks. The first week included a treatment rationale and the first step of applied relaxation (tense-relax) (24). The second week continued the applied relaxation (relax only) and also included positive imagery, sound enrichment by means of external sounds (25), hearing tactics (26), and advice regarding noise sensitivity (which is a common problem among tinnitus patients) (15). The latter two components were optional. Week 3 involved controlled breathing (as part of applied relaxation) and cognitive therapy (7), which was adjusted to deal with negative thoughts and beliefs relating to tinnitus. The module given at week 4 included differential relaxation (24) and behavioral sleep management (27). In the fifth module, rapid relaxation was presented (24), as was advice regarding concentration difficulties, exercises of concentration (mindfulness), and advice on physical activity. The final module at week 6 included continued practice of applied relaxation, relapse prevention, and a summary of the contents of the treatment program. All modules involved homework assignments and weekly reports on a report Web page to be submitted weekly. Participants were encouraged to ask questions regarding the treatment, and all queries were answered as promptly as possible by the investigators depending on their area of expertise. When submitting a week s report, the participant was sent an encouraging E-mail with the instruction to go to the next module. Statistical Analysis Results were analyzed with the statistical package STATISTICA for Macintosh (version 4.1). Confidence intervals were calculated for within-group differences and t tests for differences between change scores. 2 was used to test differences between the groups and repeated measures analysis of variance for testing within-group differences. Outcome variables were checked for skewness and kurtosis and were acceptable. RESULTS Details of recruitment are presented in Figure 1. A total of 117 participants fulfilled entry criteria and completed assessments before treatment. Of these, 66 completed full assessments after the treatment period and 51 part of assessment. The nonresponse rate was 27 in the treatment group (51%) and none in the waiting-list group. At the 1-year follow-up, 96 participants (82%) completed questionnaire measures. At this point, two participants had to be excluded because of change of E-mail address and could thus not be located in the data matrix. There were 50 participants in the waiting-list group and 46 in the treatment group in the follow-up. Patient Characteristics Baseline characteristics of the 117 participants are presented in Table 1. There was no statistically significant between-group difference at baseline. When using the 812 Psychosomatic Medicine 64:810 816 (2002)

INTERNET TREATMENT OF TINNITUS TABLE 1. Descriptive Characteristics of the Treatment and Waiting-List Control Groups from that scale, improvements were sustained for the treatment group. Treatment Group Waiting List Mean age years (SD) 48.5 (12.3) 47.2 (15.0) Mean duration years (SD) 6.2 (5.6) 6.4 (6.8) Proportion female (%) 46 48 Localization (%) Right ear only 8.5 5 Right ear dominant 10.5 21 Left ear only 17 19 Left ear dominant 23 14.5 Both ears 28 19.5 In the head 13 21 Tinnitus character (%) Tones 39.5 38 Buzzing 19 14 Both tones and 41.5 48 buzzing Sleep problems (%) 60 65 Interference with 90 84 concentration (%) Hearing problems (%) 69 68 Noise sensitive (%) 60 65 cut-off of 11 points (18) for detection of probable cases, a 23% (N 27) prevalence of anxiety and a 15% (N 18) point prevalence of depression were found. Efficacy of Treatment Mean scores at baseline, posttreatment, and at follow-up are presented in Table 2. Included are change scores ( 95% CI), ie, pretreatment minus posttreatment and pretreatment minus follow-up, and t tests for between-group differences in change scores. Also included are repeated measures ANOVA at follow-up for the whole group and for both groups separately. For the treatment group in the randomized controlled phase, statistically significant differences in change scores were found on all measures except for quality of sleep ratings, for which CIs overlapped zero. In addition, the treatment group differed from the waiting-list control group on all measures except sleep ratings, on which no improvements were found. The strongest effects on the questionnaires were found on the TRQ and on the HADS-depression scale. On the diary ratings, tinnitus annoyance decreased the most. At the 1-year uncontrolled follow-up phase, a statistically significant repeated measures effect was found on the TRQ for both groups, but apart from that, improvements were only found in the treatment group. All control participants had been offered the program by then, but their progress was not monitored. The effect on the ASI did not remain statistically significant, but apart Clinically Important Improvement Clinically significant improvement was defined as a 50% reduction of the total score on the TRQ. Using this criterion, 7 of 24 (29%) individuals in the treatment group and 2 of 48 (4%) individuals in the waiting-list group reached clinical significance at posttreatment. The difference was significant by means of a chi-square test ( 2 9.14, df 1; p.003). At 1-year follow-up, the proportion of the total sample showing a clinically significant improvement was 27/86 (31%), with 28% in the treatment group and 34% in the waiting-list group improving. The difference was not significant ( 2 0.38, df 1; p.54). A more conservative assessment of clinical significance is to include the dropouts (intention-to-treat). Counting these as treatment failures, the percentage of clinically improved cases decreases to 7 of 53 (13%) at the controlled randomized phase and 27 of 117 (23%) of the full sample at the 1-year uncontrolled follow-up phase. No significant differences between the groups were found at either posttreatment ( 2 1.14, df 1; p.29) or at the 1-year follow-up ( 2 2.6, df 1; p.16). Dropouts and Delayed Responders During the first phase of the study, E-mails were sent to the potential dropouts in order to probe for the causes for dropping out or being delayed. A majority of responders replied that the main cause for dropping out or being delayed was lack of time (N 22), and four participants had just recently begun working with the modules. Additional comments were that the program had been too fast, that there was a lack of peace and quiet at home to do the exercises, and that they had just not been sending in any reports and were still working with the program. A few comments concerned the negative aspects of the program such as being impersonal and that the program was too extensive. Unpaired t tests revealed no significant differences between the completers and the nonresponders of the first randomized controlled phase of the study on any of the pretreatment measures (all p values.05). Although the response rate was higher in the WLC group in the first phase, at the 1-year uncontrolled follow-up, the groups did not differ in response rate ( 2 1.45, p.23). DISCUSSION Participants receiving CBT via the Internet improved to a greater extent than a control group on a waiting list. Psychosomatic Medicine 64:810 816 (2002) 813

G. ANDERSSON et al. TABLE 2. Mean (SD) Scores at Pretreatment, Posttreatment, and 1-Year Follow-Up; Mean Change Scores (95% Confidence Intervals) Pretreatment Minus Posttreatment and Pretreatment Minus 1-Year Follow-Up; Results Presented for Treatment (CBT) and Waiting-List Control (WLC) Groups; Positive Change Scores Indicate Improvement Except for control and sleep Tinnitus Reaction Questionnaire a Anxiety Sensitivity Index HADS Anxiety HADS Depression CBT WLC CBT WLC CBT WLC CBT WLC Pretreatment 42.6 (21.6) Posttreatment 29.5 (22.2) Follow-up 27.0 (19.1) 38.2 (24.03) 35.4 (23.0) 26.8 (20.2) 20.6 (10.2) 16.2 (7.8) 16.8 (9.4) 19.1 (12.7) 17.8 (12.1) 15.3 (11.2) 7.9 (4.6) (N 63) 5.9 (3.6) 6.1 (3.5) 7.8 (4.9) 6.8 (4.8) 6.4 (5.0) 6.9 (4.0) 5.2 (4.1) 5.3 (3.8) 6.2 (3.8) 6.0 (3.79) Change pre to post 13.4 (7.4 19.5) 0.9 ( 2.4 4.3) 5 (2.6 7.4) 0.5 ( 1.9 2.8) 2.6 (1.2 3.9) 0.4 ( 0.4 1.2) 2.9 (1.3 4.4) 0.3 ( 1.1 0.7) Difference between t 3.99, df 70, p.002 t 2.49, df 70, p.015 t 3.05, df 70, p.004 t 3.14, df 70, p.002 change scores Change pre-follow-up 7.6 (2.2 13.0) 5.5 (0.8 10.2) 2.9 (0.02 5.7) 0.2 ( 2.4 2.8) 1.3 (0.2 2.3) 0.5 ( 0.6 1.6) 1.1 (0.4 2.4) 0.9 ( 0.7 1.3) Repeated measures a) F 17.9, df 1,94; p.0001 a) F 1.9, df 1,94; p.17 a) F 7.4, df 1,94; p.008 a) F 5.4, df 1,94; p.02 ANOVA 5.3 (4.0) b) F 13.8, df 1,45; p.0006 b) F 3.8, df 1,45; p.06 b) F 11.8, df 1,45; p.002 b) F 8.3, df 1,45; p.006 c) F 5.5, df 1,49; p.023 c) F 0.2, df 1,49; p.87 c) F 0.8, df 1,49; p.38 c) F 0.3, df 1,49; p.56 VAS Loudness VAS Annoyance VAS Control VAS Sleep Quality CBT WLC CBT WLC CBT WLC CBT WLC Pretreatment 6.3 (1.7) Posttreatment 6.2 (2.3) 6.4 (2.0) 6.4 (2.1) 6.4 (1.9) 5.3 (2.3) (N 25) 5.5 (2.2) 5.8 (2.2) Change pre to post 0.7 (0.2 1.3) 0.1 ( 0.3 0.4) 1.2 (0.5 1.8) 0.2 ( 0.4 0.1) 1.1 ( 2.2 0.1) 0.2 ( 0.6 0.2) 0.9 ( 1.7 0.2) 0.3 ( 0.7 0.1) Difference between t 2.10, df 81, p.04 t 4.57, df 82, p.001 t 2.09, df 82, p.05 t 1.68, df 81, p.10 change scores 3.9 (2.5) 5.0 (3.0) (N 25) 3.8 (3.0) 4.2 (2.9) 6.5 (2.7) 7.3 (2.3) 6.4 (2.2) 6.7 (2.1) (N 58) a) Whole group, b) treatment group, c) waiting-list group following treatment 814 Psychosomatic Medicine 64:810 816 (2002)

INTERNET TREATMENT OF TINNITUS Improvements were found on tinnitus-related distress, depression, anxiety, anxiety sensitivity, and diary ratings of annoyance caused by tinnitus. Significantly more participants in the treatment group showed an improvement of 50% on the TRQ. A high nonresponse rate was initially observed in the treatment group in the controlled randomized phase. In the uncontrolled follow-up phase, several participants were brought back into the fold and results suggest that the improvement was maintained for a proportion of participants. There are no previous similar tinnitus studies with which to compare the result, and the only direct comparison that can be made concerns an Internet-based study on headache (14), in which 50% of the treated participants reached a 50% reduction in headache activity (clinical significance). The efficacy of self-administered treatment for benign headache is well established, but for tinnitus, there is no controlled study relying on self-help only. Therefore, it was expected to find effects in the headache study, but it was less obvious that decreases of tinnitus-related distress would be found in this study. CBT has been found to convey relief from tinnitus (9), but previous studies have included face-to-face treatment, either individually or in groups. This study represents a novel way to conduct empirical trials, and there are caveats that need to be corrected in future trials. The most evident is the initially high rate of participants not responding to the posttreatment questions. One potential influencing factor might be that participants recruited via the Internet differ from those in ordinary studies. A tendency for Internet users to be disinhibited has been observed (28), and it might be that it is easier to send in E-mail than it is to respond to an advertisement via telephone or letter. This can influence the decision to drop out or delay starting the program when confronted with the demands associated with the treatment. As seen in Figure 1, most of the dropouts/late responders occurred at the beginning of the treatment, and we need to consider if we have set the bar too high when we regard these participants as dropouts. When we asked the participants to state the causes for their dropout, the most common one was lack of time. In addition, the 1-year uncontrolled follow-up showed that many of the dropouts most likely started their treatment later and even after the posttreatment assessment. In an ongoing replication study, we have assessed treatment credibility and have found that tinnitus patients rate Internet treatment as credible, but we do not know if this was the case in the present study. Another influencing factor might be interaction with the therapists. Participants had the possibility of contacting us during the whole course of the study, and many did so via E-mail, but the treatment manual could be constructed to foster continuous interactions at each step of treatment. Finally, inclusion criteria can be more conservative and participants could be asked to complete a pretreatment phase before inclusion in the study. How representative our study participants are in comparison with clinically recruited samples can be debated. Although a majority of Swedish citizens have access to the Internet, we do not know if this is the case with tinnitus patients. However, when recruiting the participants to the study, articles appeared in Swedish newspaper journals and in the journal of the Swedish Hard of Hearing Association. Although all participants were selfrecruited and not randomly selected, it is likely that we reached a broader range of tinnitus patients, at least geographically, than we do when recruiting participants locally. Finally, recent clinical application of the same treatment as presented here has shown that regular clinic patients can be helped when they receive the treatment as an option when seeking help for their tinnitus (ie, those who have not been self-recruited to participate in an Internet study). In clinical practice, we do not treat patients without prior consultation because this is incompatible with the health-care system in Sweden. Although our study protocol was approved by the local ethics committee, there is concern that the lack of any independent mechanism of verification might influence the results. So far, we have conducted several controlled trials on Internet-based self-help, and to the best of our knowledge, we have not had any problem with unsolicited participation. Future studies should be designed so that unsolicited participation is prevented. It is worth restating that participation is time consuming and that an individual would have little to gain from misrepresenting aspects such as age and history. In addition, the inclusion criterion we used regarding prior medical consultation for tinnitus was concealed in the sense that the question was asked without any information that it would serve as an exclusion criteria. As seen in Figure 1, 11 potential participants were excluded for that reason. Psychological treatment for tinnitus is not aimed at eliminating the tinnitus; rather, reduced annoyance is the goal (9). Less than a third of the treated participants achieved substantial improvement (50%), but our impression from E-mails is that a majority of completers found the treatment to be beneficial. At this stage, it is not possible to discern differential effects of the treatment ingredients. Studies should be designed for this purpose to dismantle differential effects of different components. In order to be able to determine if the Internet can be used more regularly, there is a need to compare Internet treatment with ordinary treatment within the same study. Providing treatment via the Internet has advantages over self-help books in that advice can be given on a continuous basis without delay. In comparison with ordinary treatment, it is cost effective and it also makes the Psychosomatic Medicine 64:810 816 (2002) 815

G. ANDERSSON et al. treatment available to persons living far from the specialist center. However, we do not regard Internet treatment as an alternative but rather as a complement to tinnitus management at the home clinic. It is likely that the Internet expansion will result in more Internet applications of previously tested CBT treatments and other clinical applications (29 31). However, doing controlled trials via the Internet may be more difficult than we expected when we began. The treatment program must be read carefully step by step, and the homework must be done for the treatment to have any effect. It is likely that participants in a treatment trial via the Internet have to be monitored closely (eg, by means of telephone calls), and potentially a proper screening of motivation would result in a better response. We conclude that the Internet can be used as an adjunct treatment tool for tinnitus sufferers and that there is a need to further develop and evaluate Internet-based treatments. This research was supported through a project grant from the Swedish Council for Social Research and a grant from the Swedish Hard of Hearing Association. REFERENCES 1. Jastreboff PJ. Phantom auditory perception (tinnitus): mechanisms of generation and perception. Neurosci Res 1990;8: 221 54. 2. McKenna L, Andersson G. Hearing disorders. In: Hersen M, Bellack A, editors. Comprehensive clinical psychology. Oxford: Pergamon Press; 1998. p. 69 83. 3. Davis A, Rafie EA. Epidemiology of tinnitus. In: Tyler RS, editor. Tinnitus handbook. San Diego: Singular Thomson Learning; 2000. p. 1 23. 4. Vesterager V. Tinnitus investigation and management. BMJ 1997; 314:728 31. 5. The Wessex Institute for Health Research and Development. Development and Evaluation Committee Report (no. 83): Tinnitus retraining therapy. March 1998. [accessed in September 1998]. Available from: URL: http://www.epi.bris.ac.uk/rd 6. Dobie RA. A review of randomized clinical trials of tinnitus. Laryngoscope 1999;109:1202 11. 7. Hawton K, Salkovskis PM, Kirk J, Clark DM. Cognitive behavior therapy for psychiatric problems. Oxford: Oxford University Press; 1989. 8. Andersson G, Melin L, Hägnebo C, Scott B, Lindberg P. A review of psychological treatment approaches for patients suffering from tinnitus. Ann Behav Med 1995;17:357 66. 9. Andersson G, Lyttkens L. A meta-analytic review of psychological treatments for tinnitus. Br J Audiol 1999;33:201 10. 10. Eysenbach G, Sa ER, Diepgen TL. Shopping around the Internet today and tomorrow: towards the millennium of cybermedicine. BMJ 1999;319:1294. 11. Eysenbach G, Diepgen TL. Towards quality management of medical information on the Internet: evaluation, labelling, and filtering of information. BMJ 1998;317:1496 1502. 12. Kuppersmith RB. Is e-mail an effective medium for physicianpatient interaction? Arch Otolaryngol Head Neck Surg 1999;125: 468 70. 13. Grohol JM. Future clinical directions: professional development, pathology, and psychotherapy on-line. In: Gackenbach J, editor. Psychology and the Internet. Intrapersonal, interpersonal and transpersonal implications. San Diego: Academic Press; 1998. p. 111 140. 14. Ström L, Pettersson R, Andersson G. A controlled trial of self-help treatment of recurrent headache conducted via the Internet. J Consult Clin Psychol 2000;68:722 7. 15. Andersson G, Lyttkens L, Larsen HC. Distinguishing levels of tinnitus distress. Clin Otolaryngol 1999;24:404 10. 16. Wilson PH, Henry J, Bowen M, Haralambous G. Tinnitus reaction questionnaire: psychometric properties of a measure of distress associated with tinnitus. J Speech Hear Res 1991;34:197 201. 17. Andersson G, Vretblad P, Larsen H-C, Lyttkens L. Longitudinal follow-up of tinnitus complaints. Arch Otolaryngol Head Neck Surg 2001;127:175 9. 18. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67:361 70. 19. Herrmann C. International experience with the hospital anxiety and depression scale a review of validation data and clinical results. J Psychosom Res 1997;42:17 41. 20. Lisspers J, Nygren A, Söderman E. Hospital anxiety and depression scale (HADS): some psychometric data for a Swedish sample. Acta Psychiatr Scand 1997;96:281 6. 21. Reiss S, Peterson RA, Gurksy DM, McNally RJ. Anxiety sensitivity, anxiety frequency and the prediction of fearfulness. Behav Res Ther 1986;34:1 8. 22. Andersson G, Vretblad P. Tinnitus and anxiety sensitivity. Scand J Behav Ther 2000;29:57 64. 23. Buchanan T. Potential of the Internet for personality research. In: Birnbaum MH, editor. Psychological experiments on the Internet. San Diego: Academic Press; 2000. p. 121 40. 24. Öst L-G. Applied relaxation: description of a coping technique and review of controlled studies. Behav Res Ther 1987;25:379 409. 25. Jastreboff PJ, Jastreboff MM. Tinnitus retraining therapy (TRT) as a method for treatment of tinnitus and hyperacusis patients. J Am Acad Audiol 2000;11:162 77. 26. Andersson G, Green M, Melin L. Behavioral hearing tactics: a controlled trial of a short treatment program. Behav Res Ther 1997; 35:523 30. 27. Lichstein KL, Riedel BW. Behavioral assessment and treatment of insomnia: a review with an emphasis on clinical application. Behav Ther 1994;25:659 88. 28. Joinson A. Causes and implications of disinhibited behavior on the Internet. In Gackenbach J, editor. Psychology and the Internet. Intrapersonal, interpersonal and transpersonal implications. San Diego: Academic Press; 1998. p. 43 60. 29. Winzelberg AJ, Eppstien D, Eldredge KL, Wilfley D, Dasmahapatra R, Dev P, Taylor CB. Effectiveness of an Internet-based program for reducing risk factors for eating disorders. J Consult Clin Psychol 2000;68:346 50. 30. Lange A, Schrieken B, van den Ven J-P, Bredweg B, Emmelkamp PMG, van der Kolk J, Lydsdottir L, Massaro M, Reuvers A. Interapy : the effects of a short protocolled treatment of posttraumatic stress and pathological grief through the Internet. Behav Cognit Psychother 2000;28:175 92. 31. Carlbring P, Westling BE, Ljungstrand P, Ekselius L, Andersson G. Treatment of panic disorder via the Internet a randomized trial of a self-help program. Behav Ther 2001;32:751 64. 816 Psychosomatic Medicine 64:810 816 (2002)