Internet-Based CBT for Tinnitus

Abstract

Cognitive behaviour therapy (CBT) can reduce tinnitus distress but is not available for most patients. Therapist guided, internet-based CBT (ICBT) increase availability and has been shown to be effective. However, the initial positive results need to be replicated in larger samples, and treatment dropout has not been thoroughly studied. Moreover, it has not been evaluated if a low-intensity version of ICBT without therapist contact could be an alternative for patients who do not need or are able to manage the full ICBT-program. This study evaluated two parallel interventions delivered in regular care: ICBT for tinnitus distress (n = 293) and a low-intensity version of ICBT (n = 81) for patients with lower levels of tinnitus distress. We also explored predictors of dropout from ICBT and if dropout influences outcome. Tinnitus Reaction Questionnaire (Wilson, Henry, Bowen, & Haralambous, 1991) was used as the primary outcome. Secondary outcomes were measures of depression, anxiety, sleep, and sound sensitivity. Significant reductions following ICBT were found on all measures after treatment and also at a three-month follow-up. Patients receiving low-intensity ICBT showed a significant reduction in distress, even when they had low levels of distress initially. Treatment dropout was preceded by an increase in days spent at each treatment step but not by an increased distress. Early dropout was related to worse outcome. ICBT can be used in a regular clinical setting to reduce tinnitus distress. Early dropouts may need additional management. For help-seeking patients with lower distress, a low-intensity version of ICBT can be used.

Background

Tinnitus is defined as the experience of sound in the absence of any external stimulus (Lockwood, Salvi, & Burkard, 2002) and is found in about 10–15% of the population (Henry, Dennis, & Schechter, 2005). For 0.5–3% of the adult population, tinnitus constitutes a significant problem in their lives, with insomnia, concentration problems, anxiety/depression, and hearing problems being common complaints in association with tinnitus (Andersson 2002). Unfortunately, only a minority of tinnitus patients can be helped by medical interventions aimed directly at tinnitus (Baguley, Andersson, McKenna, & McFerran, in press). This indicates that at present, the best way to help patients suffering from tinnitus is to focus on decreasing the distress related to tinnitus. Cognitive behaviour therapy (CBT) is the most empirically supported intervention that aims to reduce tinnitus-related distress (Hesser, Weise, Zetterqvist Westin, & Andersson, 2011; Hoare, Kowalkowski, Kang, & Hall, 2011). However, there is a general shortage of CBT therapists (Gander, Hoare, Collins, Smith, & Hall, 2011). One way to address this problem is to develop treatments demanding less therapist time. Guided self-help treatments have been used for a wide range of conditions, very often with positive results (Watkins & Clum 2008). Therapist-guided Internet-based CBT (ICBT) has been successfully applied within the psychiatric area (Andrews, Cuijpers, Craske, McEvoy, & Titov, 2010; Johansson & Andersson 2012), and promising results have been reported for behavioural medicine problems (Cuijpers, van Straten, & Andersson, 2008). Four randomised controlled trials have shown positive results for therapist-guided self-help based on CBT for tinnitus: three used Internet to deliver the treatment (Andersson, Strömgren, Ström, & Lyttkens, 2002; Hesser et al., 2012; Kaldo et al., 2008) and one used a self-help book (Kaldo, Renn, Rahnert, Larsen, & Andersson, 2007). An open effectiveness trial at the Department of Audiology at Uppsala University Hospital in Sweden also showed positive results for ICBT in a regular clinical setting (Kaldo-Sandström, Larsen, & Andersson, 2004). Taken together, the effects are positive and equal to, or in some cases somewhat lower, face-to-face CBT for tinnitus, with effects remaining at one-year follow-up. The proportion of patients experiencing a clinically significant reduction in tinnitus distress, defined as at least a 50% reduction on the Tinnitus Reaction Questionnaire (TRQ), was 23–43%. Effect sizes (Cohen's d) ranged between 0.30 and 0.73. Tinnitus is not an indication for CBT unless it is perceived as a distressing condition, and a significant level of tinnitus-related distress is often needed to motivate the patient to complete CBT. In the daily clinical situation, it is sometimes difficult for both patient and therapist to decide whether the patients’ symptom profile, level of functioning, and current life situation indicate that a full CBT or ICBT treatment should be started or not. Clinical experience with this patient group suggests there is a risk that patients not enough distressed, motivated, or able to complete the treatment may be offered the treatment anyway, because they are help-seeking and it seems to be the only possible intervention at hand. This risk is even more pronounced in the case of ICBT, since it might be regarded as a “light version” of regular CBT. However, even though ICBT saves time for the therapist, it should not be regarded as a less intense treatment for the patient (Andersson et al., 2008). One way to better match the patients need and ability to the right level of treatment intensity would be to offer a self-help material to work with on their own pace. This would be presented as an opportunity to learn more about tinnitus and how to deal with it. After reading relevant parts, the patients decide if they want to try to work with some of the described methods. In line with this idea, some patients at the Department of Audiology were offered the possibility to access the modules in the ICBT program with the instruction to use them primarily for psychoeducational purposes and with no demand on doing or reporting homework. The patients were able to contact a therapist to ask questions, but no regular therapist contact or contact initiated by the therapist occurred. This way of delivering the treatment materials is less intense for the patient and is labelled low-intensity ICBT.

Many eHealth trials evaluating open-access, non-guided eHealth interventions aimed at the general population find that a substantial proportion of users drop-out before completion (Eysenbach 2005). Sometimes no more than 2% completes the interventions (Melville, Casey, & Kavanagh, 2010). However, therapist-guided Internet treatments with thorough assessment and visits before and after treatment often show drop-out rates similar to those found in traditional psychological treatments (Cuijpers et al., 2009). In the ICBT studies for tinnitus, drop-out rates have ranged from 20% to 51%, with the largest dropout in the first controlled trial by Andersson et al., (2002), which is also the trial with the lowest amount of therapist contact. This is on average more than that found in traditional group treatment, with a drop-out rate of 24% (Kaldo et al., 2008). However, when the number of accessed Internet modules was compared with the number of attended group sessions, the “received treatment dose” was 75% in both cases. It makes sense that dropout should be problematic since the patients end treatment before they can access and practice all the components included in treatment. Still, it is somewhat unclear how detrimental to outcome dropout actually is. Kaldo-Sandström et al., (2004) reported no correlation between the number of completed modules and outcome, while treatment compliance did correlate with outcome. One possible explanation is that dropout also occurs when patients experience an improvement and not only because they find the treatment to be ineffective or too demanding. When asked about their reason for dropping out, patients often states practical problems such as lack of time as the main reason (Andersson et al., 2002). However, the issue of dropout needs more empirical investigation, especially in a regular care setting, since the context there differs from previous efficacy trials. In addition, there is also a need to further explore factors that can predict dropout in therapist-guided ICBT. In 2002, the ICBT treatment and the technical platform at the Department of Audiology in Uppsala were updated. The content became more extensive to better cover and explain different problem areas and methods to deal with them, and the technical platform made the treatment more interactive and included more registrations for the patients in order to enable them and the therapist to better monitor the progress and give more specific feedback. One important feature of the updated ICBT platform is that patients rate their tinnitus distress and perceived tinnitus loudness at the end of each step in treatment. This, together with measures on numbers of days spent on each treatment step, gives an opportunity both to analyse patterns preceding dropout and to explore if dropouts seems to end treatment worse off than when they started. The updated ICBT platform was evaluated in a study by Kaldo et al. (2008 with positive results and with equal to or slightly better completion rates than in previous trials. However, no evaluation of the new ICBT platform regarding outcome or dropout has been performed with consecutive patients in a regular clinical setting. In addition, the low-intensity ICBT option that was included as a complement to full ICBT has not been evaluated previously.

The aims of this study were

to investigate if the new therapist-guided ICBT can reduce tinnitus distress and related problems (depression, anxiety, sleep, sound sensitivity) in consecutive tinnitus patients at a regular audiology clinic;

to evaluate if low-intensity ICBT can be helpful for patients with initial lower distress levels or not being able to participate in the full ICBT intervention; and

to explore dropout in ICBT, its impact on outcome, and if dropout can be predicted by patterns in tinnitus distress, perceived tinnitus loudness, and number of days spent on each step in treatment.

Method

Design and procedure

This study used a non-controlled treatment evaluation design with consecutive patients in a regular clinical setting. The study can be described as an effectiveness trial (Shadish et al., 1997), as regular therapists provided the treatment in a regular clinical setting and no advertisements were used to recruit patients. Moreover, ICBT and low-intensity ICBT constituted two of five possible options, the other three being group CBT, individual CBT, and one-session group psychoeducation. Three measurement points were used for ICBT: pre- and post-treatment and follow-up three months after the end of treatment. For low-intensity ICBT, the post-measure was performed when the patients had had the opportunity to read and work with the intervention during three months.

Participants and choice of treatment

All consecutive patients who had started and either completed or dropped out of ICBT or low-intensity ICBT between June 2002 and February 2008 where included in this study (N3 = 374). These constituted a sub-group of 909 patients who were referred to CBT after a medical examination by an Ear Nose Throat specialist either at the department or from clinics in other parts of Sweden. All patients went through a structured psychological assessment (Andersson, Lyttkens, & Larsen, 1999) lasting between 1 and 1.5 hours. Patients were found eligible to try ICBT for tinnitus if they met the following criteria:

tinnitus was judged to be of major concern and the primary or one of the main presenting problems;

no other problem, medical, psychological, or social, in need of acute management or that would clearly interfere with CBT for tinnitus;

having a tinnitus duration of at least three months;

having enough time to work with the treatment, as estimated to be between 15 and 60 minutes a day for 6–10 weeks;

agreeing to work towards the main goal of treatment; to reduce tinnitus distress rather than the sound;

The decision concerning criterion 7 was influenced by a range of factors. One aspect was the patients experience with and preference for the Internet and how comfortable they felt using this media in a treatment. The patients’ need of face-to-face support and flexibility in treatment and some patients’ wish to meet other individuals with tinnitus were considered. Also, practicalities like how far patients had to travel to take part in a face-to-face treatment at the department had to be weighed against the other factors. Those patients receiving individual treatment were often in need of more support and/or had other problems that would interfere with Internet treatment or group treatment, e.g. social anxiety. If the psychologist found criterion 1 not to be fulfilled when interviewing the patients, but tinnitus still was of some concern, low-intensity ICBT (described below) was recommended as one possible option. This was also the case when the patients had little time to engage in the complete treatment (criterion 4) and when the patients themselves expressed doubts about completing the whole treatment package or only showing a specific interest in some parts of treatment. This procedure resulted in 293 patients starting ICBT and 81 starting low-intensity ICBT, representing 32.2% and 8.9% of the 909 patients remitted to the department during this period. A number of patients were included in group CBT (n = 238), individual CBT (n = 60), one-session group psychoeducation (n = 66), other treatment (n = 7), or CBT for sound sensitivity (rather than tinnitus distress; n = 70). The rest of the patients did not start any intervention because they were satisfied with the assessment and the information they received (n = 16), were expected not to benefit from treatment or referred to other treatment (n = 17), declined CBT or ICBT (n = 19), declined low-intensity ICBT (n = 2), did not show up for assessment (n = 19), or because of other reasons (n = 21).

Measurements

With the exception of the structured psychological assessment interview described above, all self-report measures were completed over the Internet. Internet administration of self-report instruments has been found to have the similar psychometric properties as their paper-and-pen equivalents (Buchanan 2003), which also included measures used in this study (Andersson, Kaldo-Sandström, Ström, & Strömgren, 2003).

Outcome measures

The TRQ was used to assess tinnitus annoyance and distress (Wilson et al., 1991). The TRQ consists of 26 zero- to four-point items summated into a total score from 0 to 104 and have a reported internal consistency (Chronbach's α) of α = .96 and a test–retest correlation of r = .88 (Wilson et al., 1991). A 50% reduction of a patient's TRQ score was used to define a clinically significant reduction in distress, an important complement to statistical significance. This definition has been used in previous tinnitus self-help trials and originates from previous CBT studies on tinnitus (Henry & Wilson 1996, 1998), and has been recommended for use in behavioural medicine trials (Blanchard & Schwartz 1988). The Hospital Anxiety and Depression Scale (HADS) (Zigmond & Snaith 1983) was used to measure symptoms of anxiety and depression. It has two subscales measuring depression (HADS-D) and anxiety (HADS-A) with seven items (response scale 0–3) each. HADS has good internal consistency, and test–retest correlations and is suitable for patients with somatic problems because of its focus on symptoms of anxiety and depression that are less affected by somatic symptoms. The psychometric properties for the Internet-administered HADS are very similar to the paper-and-pencil administered questionnaire (Andersson et al., 2003). The Insomnia Severity Index (ISI) (Bastien, Vallières, & Morin, 2001) was used to measure sleep problems. The ISI is a brief five-item (scored 0–4) questionnaire assessing sleep duration, sleep quality, and negative daytime symptoms. It has good Internet consistence and has been found suitable for detecting changes in perceived sleep difficulties. The question “Are you sensitive to sound?” was rated from 1 (“not more than others”) to 5 (“extremely”) to estimate the level of sound sensitivity.

Measures during treatment and definition of dropout

At the end of each of the six treatment steps, when reporting their homework, the patients also answered two questions to measure tinnitus distress and perceived tinnitus loudness, respectively, “How much have tinnitus bothered you the last week?” rated from 1 (“Not at all”) to 9 (“Extremely much”), and “How loud have tinnitus been the last week?” rated from 1 (“Completely silent”) to 9 (“Extremely loud”). It should be noted that these measures were not weekly measures or distributed with the same temporal distance from each other, since the time spent in each treatment step varied between patients. In order to measure how rapidly the patients progressed in treatment, the number of days from the start to the end of each treatment step was calculated. The level of treatment completion was represented by how many of the six treatment steps the patients had logged in to and started. This measure was preferred to the total number of modules since the number of modules during each treatment step could vary due to the patients’ individual needs and choices (see below). Patients who finished the final treatment step by reporting their homework with the modules in that step were considered having completed treatment, the rest were considered treatment dropouts.

Treatments

Therapist-guided (ICBT

The ICBT intervention is based on CBT methods used to handle tinnitus distress and related problems (Andersson & Kaldo 2006a). It consists of separate treatment modules, each focusing on a specific problem area and methods to manage it (Andersson & Kaldo 2006b). Each module includes psychoeducation, a rational for the method, instructions on how to practice and use the method, and finally online planning, registration, and reporting of homework. Some modules, for example, applied relaxation and focus exercises, are given in the same order to all patients and constitute the basic six-step treatment program. Others, for example, on sound sensitivity, sleep problems, and hearing difficulties, are chosen more freely by the patients after an initial discussion and prioritization with the therapist. During each of the six treatment steps, the patient can actively work with one to four treatment modules in parallel. In the end of each step, the patients report on all the modules they have worked with during that step to enable the therapist to give written feedback on the patient's progress. There is not a fixed date when the treatment ends. This gives the patients a rather large flexibility in how much time they spend on the treatment. Each treatment step is supposed to take 7–10 days. This gives an expected treatment time of 6–9 weeks. The treatment period could be extended if the patient and therapist agreed there was a good reason for this. During treatment, patients could contact the therapist via the Internet. A very important part in the therapist role is to monitor the progress of all patients and contact inactive patients via the Internet or by telephone. The treatment platform provides therapists with information and automatic notifications to make the monitoring more effective. Patients being inactive for more than 10 days were always contacted by the therapist. In total, nine different therapists assessed and treated the patients. All but one had an MSc psychology degree from 4the five-year clinical psychology program in Sweden, including 1.5 years of clinical training in CBT. Four had already received their license as a clinical psychologist and five first did their mandatory year of practice (under supervision) at the department and then received their license.

Low-intensity ICBT

The low-intensity ICBT includes all the text in the full ICBT package, but without the interactive homework planning, registration, and reporting. The patients are encouraged to read the most relevant modules and to try the methods described in them. They are informed that they are supposed to read and try out the material during three months and then they will fill out more questionnaires and receive feedback from the therapist. The patients have the opportunity to contact the therapist via the Internet to ask questions. However, the therapist does not actively monitor or contact the patients.

Statistical analyses

All outcome analyses were calculated according to the intent-to-treat principle (including treatment dropouts in the analyses), both for those filling out post- and follow-up measures (observed data), and with missing data replaced by the last observation carried forward (LOCF). Analysis using data assessed during treatment excluded four patients due to their very long breaks in treatment (over 200 days), which we regarded as outliers. Outcomes were calculated with two-tailed paired t-tests. Alpha level was 0.05. Effect sizes were calculated as Cohens’ d, where the mean of the post- or follow-up measure was subtracted from the pre-treatment measure and then divided by the pooled standard deviation.

Treatment outcome

In the ICBT condition, 63% (n = 184) filled out the post-treatment measures and 54% (157) filled out the three-month follow-up. Table 2 and Figure 1 show the outcome for the ICBT group on all measures (observed data).

Figure 1 Outcome of full Internet treatment (ICBT) on tinnitus distress and secondary outcome measures. (1) Tinnitus Reaction Questionnaire. (2) Hospital Anxiety and Depression Scale—Depression. (3) Hospital Anxiety and Depression Scale—Anxiety. (4) Insomnia Severity Index. (5) Single question.

Significant decreases were found on all measures and all measurement points, both when calculated with the observed data and with LOCF (Table 2). The effect sizes were in the medium range, except for a small effect size found for sound sensitivity. The criterion for clinical significant reduction (50% reduction or more on the TRQ) was reached for 69 of the patients at post-treatment. These constitute 37.5% of the patients with observed data at post-treatment and 23.5% of all patients starting treatment (Intent-to-treat; ITT). The corresponding figures at three-month follow-up where 63 patients reached clinical significance were 40.1% (observed data) and 21.5% (ITT). On average, the patients receiving ICBT stayed in treatment for 94.4 days (SD = 75.7) when one outlier with 1360 days and 11 patients who never started the first step in treatment were excluded. This was markedly longer than the expected 46–60 days. Treatment dropouts had an average treatment time of 83.4 days (SD = 86.9), which did not differ significantly from patients completing all six treatment steps (M = 99.4 days; SD = 69.7). In the low-intensity ICBT group, the post-measure (three months after access to the intervention) were filled out by 63% (n = 51) of the patients. The initial mean of 22.1 (SD = 17.5) for the TRQ was as expected lower than for those who received full ICBT (t(372) = 5.68, p < .001). However, in spite of their lower entry scores, the 51 patients who filled out the follow-up measures showed a small (Cohen's d = 0.26) but significant reduction from 20.4 (SD = 16.3) to 17.5 (SD = 17.1) on the TRQ (t(50) = 2.08, p = .043). When analysed as LOCF, the TRQ was reduced from 22.1 (SD = 17.5) to 20.3 (SD = 18.2), which remained significant (t(80) = 2.05, p = .043) but with a considerable smaller effect size (Cohens’ d = 0.10).

Treatment drop-out patterns and effects on outcome

In ICBT, 62% (n = 181) finished the whole treatment. The average number of accessed treatment steps was 4.6 (i.e. 77% out of the maximum of six steps). No differences were found between dropouts and completers with regard to gender or pre-treatment levels in the primary and secondary outcomes, years with tinnitus, number of previous treatments, hearing deficits, or type of referral (external vs. internal). However, dropouts were younger (M = 44.8 years, SD = 13.3) than completers (M = 49.1, SD = 13.6), and this difference was significant (t[288] = 2.66; p < .01). Concerning differences in outcome, the patients who ended treatment prematurely lacked follow-up data to a large extent, with an attrition of 88% at post-treatment and 89% at follow-up. Therefore, the ratings of tinnitus distress and loudness from the patients’ last step in treatment were used to compare dropouts with treatment completers. After the first step in treatment, the average for tinnitus subjective loudness was 6.5 (SD = 1.5) and the average for tinnitus distress was 6.2 (SD = 1.6). On the last available rating of tinnitus subjective loudness, dropouts had a mean rating of 6.2 (SD = 1.9), which compared to the completers mean of 5.6 (SD = 1.8) was significantly higher (t(247) = 2.13, p < .05). The same pattern was found for the last available rating of tinnitus distress (dropouts: M = 6.0, SD = 1.9; completers: M = 5.3, SD = 1.9; t(247) = 2.75, p < .01). These results were stable also when controlling for pre-treatment levels of tinnitus distress as measured by the TRQ. A sub-group analysis of these data showed that it was only the group of patients dropping out of treatment after the first step in treatment that had significantly higher distress than the completers (t(243) = 3.23, p < .001), and also higher loudness ratings (t(243) = 2.88, p < .01). As can be seen in Figure 2, the trend is otherwise that regardless of when patients drop out from treatment, they have a lower level of distress and often a lower level of loudness compared to their initial level. Figure 2 The relationship between treatment dropout and tinnitus distress, tinnitus loudness, and days per step in treatment. Note. Data collected with a one-item rating at the end of each treatment step. Each line represents patients dropping out during a certain step in treatment or completing all six steps. Four patients were excluded to avoid outlier effects.

Figure 2 also shows the clear pattern that completers finished each treatment step much quicker than the dropouts. This difference is more obvious than the difference on the ratings of distress and loudness.

Discussion

The aim of this study was to investigate the effects of a therapist-guided ICBT intervention on tinnitus distress and related problems for patients treated at a regular care audiology clinic. Effects of a low-intensity version of ICBT and dropout from treatment were also explored and adherence and dropout rates were analysed. Guided ICBT significantly reduced tinnitus distress, symptoms of depression and anxiety, as well as sleep problems. This replicates the findings of Kaldo-Sandström (2004) with similar within-groups effect sizes at post-treatment (d = 0.58 vs. 0.66) and follow-up (d = 0.55 vs. 0.68), and confirms that ICBT for tinnitus distress can be delivered to consecutive tinnitus patients who receive treatment in a regular clinical setting. The attrition rate of 37% at post-treatment was somewhat higher than that in our previous study. However, analyses using LOCF to replace missing data still showed significant improvements. A new finding in this study was that the treatment material on the Internet platform also could be used as a low-intensity intervention for less distressed patients not suitable for full ICBT and for patients who for different reasons could not participate in full ICBT. Although the effects were rather small, probably due to the low initial distress level, as this has previously (Abbott et al., 2009; Malouff, Noble, Schutte & Bhullar, 2010) been found to reduce the effect size, this is an important finding. A lack of options for these patient groups might result in them nevertheless participating in full treatment due to their help-seeking behaviour. Offering full treatment to patients for whom a less intensive treatment is sufficient is not in line with the idea of stepped care and can lead to resources being used in an inefficient way as well as disappointing or overloading the patients (Cima et al., 2012). Compared to a more brief, purely psychoeducational written material, low-intensity ICBT offers the patients the opportunity to work more thoroughly with methods of specific interest if they chose to. The Internet also provides a good opportunity to ask the therapist for help on specific issues. Taken together, these opportunities probably give help-seeking patients a sense of control that might in itself be beneficial, but this has to be tested in future trials, preferably comparing low-intensity ICBT to a purely psychoeducational material. Compared to the one-session psychoeducational group intervention, low-intensity ICBT has the advantages of being accessible at a distance and letting the patients work through the material in a flexible way and read and repeat the parts most relevant for them. However, one observation regarding low-intensity ICBT is that patients only very rarely used the opportunity to contact the therapist. This is probably partly due to a generally lower motivation to use the material, which generates fewer questions, but it may also indicate that the strategy to encourage patients to contact the therapist when needed leads to fewer contacts and potentially worse outcome (Kenwright, Marks, Graham, Franses, & Mataix-Cols, 2005). The term “guided self-help” should therefore be reserved for treatments where an active, not a passive, therapist is involved (Andersson 2009). In the full ICBT-intervention, the treatment completion (62%), and amount of accessed treatment steps (77%) were very similar to when the same treatment was tested in a previous controlled trial (Kaldo et al., 2008), and slightly better compared to the earlier version of treatment when used in the same clinical setting (Kaldo-Sandström et al., 2004). The tendency for dropouts to not complete symptom measures after dropping out of treatment makes it difficult to estimate the effects of non-completion of a treatment. However, in this study, the one-item ratings of distress and loudness made at the end of each treatment step presented a good opportunity to explore this. We found that dropout from treatment directly after the first step was related to lower effects, but that dropout later on in the treatment did not seem to be associated with outcome. One interpretation of this could be that early dropout reflects a mismatch between patient and treatment or less ability to utilise or find time for the treatment, while later dropout reflects lower motivation for treatment because some benefits have been achieved by then. The clinical implication of this would be to routinely make further assessment of early dropouts and consider other treatment options more rapidly. No pre-treatment characteristic except for age could predict dropout. During treatment, the number of days spent on each treatment step seems to be a more useful predictor than the ratings of distress and subjective tinnitus loudness. Figure 2 shows that on a group level, dropout tends to occur after exceeding 20 days to complete the latest step in treatment, while treatment completers steadily lies below 20 days for each step.

A related issue is the fact that on average, the patients spent much more time in treatment than expected. This is probably due to a number of factors. The flexible way the treatment was administered, without an absolute deadline for treatment, could make the both the patient and the therapist more inclined to allow more time to pass before finishing each step in treatment. Many other ICBT studies have a fixed treatment time that spans between 10 and 12 weeks, and this probably creates a deadline effect motivating both the patient and the therapist to move forward. Future studies should evaluate the effects of having a clear deadline (Nordin, Carlbring, Cuijpers, & Andersson, 2010). Another related factor is the use of parallel modules in each treatment step, something that is less common in other ICBT interventions (but see Paxling et al., 2011, for one exception). This increases the workload for the patients, and taken together it might have been too much to manage in only 46–60 days. The actual time in treatment found in this study (94 days) more closely resembles the treatment period of 10–12 weeks commonly found in other ICBT trials. Finally, in some cases, the therapist agreed in letting the patient take a planned pause in the treatment that could last for several weeks and even months. Unfortunately, these cases (estimated to be 5–10% of all patients) were not registered and therefore no sub-analyses without these cases could be made. The characteristics of the patients who received full ICBT were rather similar to the previous effectiveness study using the older treatment platform (Kaldo-Sandström et al., 2004). The number of male patients was higher (64% compared to 52%), external referrals were markedly higher (60% compared to 30%), and the use of at least one hearing aid was higher (22% compared to 10%). The mean distress level according to TRQ was somewhat higher (36.5 compared to 33.0), but not if the patients receiving low-intensity ICBT was included (giving an overall mean of 33.4 on the TRQ). These comparisons should be seen in the light of the fact that our current sample was much larger and that Internet use and access increased during the period between the two studies. This study has limitations. Since there was no control group, the passage of time alone might have caused the decrease in distress. Regarding the full ICBT option, this is not very likely since previous randomised controlled trials shows a superiority of the treatment when compared to wait-list control group (Andersson, et al., 2002; Hesser et al., 2012). It is, however, more likely that the rather small decrease in distress after the low-intensity ICBT would also have occurred without the intervention (Hesser, Weise, Rief, & Andersson, 2011). Even if this would be the case, it is still an important finding that it is not harmful to use this resource-saving strategy to give these patients the low-intensive and non-guided version instead of the full treatment package. We conclude that guided ICBT can be successfully used in a regular clinical setting to reduce distress among patients suffering from tinnitus. In addition, a low-intensity version of the intervention could be used for patients with lower distress levels or with difficulties or low motivation to participate in full treatment. Early dropout from treatment is related to lower effects and should be regarded as an indication for further assessment and referral. Interestingly, levels of distress are not related to treatment dropout. Dropout is probably better predicted by a lack of engagement in terms of a decrease in treatment pace and activity rather than by an increase in distress. Continue your exploration of Gamification and Game-Based e-Learning in the field of information technology with our related content.

Acknowledgements

We thank the staff at the Department of Audiology, Uppsala University Hospital, for support over the years. We also thank Leif Lyttkens and Olafur Jacobson for making the implementation of ICBT for tinnitus possible. Finally, the Swedish council for working and life research and Hörselfonden are thanked for support in the development of ICBT for tinnitus.