What is already known on this topic?

Insomnia is a widespread health problem, with cognitive behavioural therapy for insomnia (CBT-I) considered as first-line treatment.1 Unfortunately, access to CBT-I treatment is limited due to limited numbers of trained therapists and cost. Randomised controlled trials (RCTs) have provided strong evidence for web-based CBT-I as an effective treatment for insomnia.2 However, these studies have been of short duration and excluded people with comorbidities.3–5

Methods of the study

One thousand two hundred and twelve US-based participants were screened after indicating interest in the trial. The final study comprised 303 adults, aged 21–65 years (mean age: 43.28 years); 72% were female; 84% were white; and 78% had a college degree or higher. Participants were very computer literate with 96% checking their email daily, and had symptoms of chronic insomnia. Comorbid medical and psychiatric disorders were included unless severe (49.8% reported at least one medical or psychiatric comorbidity). The intervention was designed as an RCT, comparing a web-based CBT-I intervention, Sleep Healthy Using The Internet (SHUTi), to a control condition. SHUTi was a 6-week fully automated, interactive and tailored web-based programme that incorporated the primary tenets of face-to-face CBT-I. The control condition included non-tailored and fixed online material about insomnia, lifestyle, environmental and behavioural strategies, which had some overlap of content with web-based CBT-I. The primary outcomes measures for the study were severity of insomnia symptoms and distress around insomnia (Insomnia Severity Index—ISI) and measures of sleep duration (Sleep Onset Latency and Wake after Sleep Onset—WASO) from sleep diary data. Participants were randomised to either condition via a random number generator. This was a single-blind RCT, with participants blinded to condition. The study coordinator remained blinded to randomisation until after the individual was deemed eligible to participate and following the completion of the baseline assessment. It was not reported whether the assessors were blinded to participant condition when analysing the data. Participants’ data were collected at four time points: baseline, 9 weeks (end of intervention), 6 months and 1 year.

What this paper adds

• Web-based CBT-I is an effective treatment for insomnia that retains its effects up to 1 year. At 1-year follow-up, ISI reduced by 10 points for CBT-I versus 6 points with web-based patient education, and WASO reduced by 48.6 min versus 35.0 min. The number needed to treat (NNT) at 1 year to achieve a treatment response (drop of >7 points on ISI) was 3.7, and NNT was 3.3 to achieve remission (ISI score of <8).

• The presence of medical or psychiatric comorbidities did not reduce the effectiveness of web-based CBT-I.

• Web-based CBT-I was successful with no therapist assistance (fully automated) and dropout was surprisingly low (9.2% for intervention and control group combined postassessment) in this group of self-referred, highly educated, computer literate participants. Of note, the web-based CBT-I group was more adherent than the control group, with a median programme login count of 25 versus 1, over the 9-week intervention.

Limitations

• Results may not be generalisable to broader population or to people referred to web-based CBT-I by physicians. Participants were highly computer literate (98% used the internet daily) and highly educated (78% college degree or higher). They self-referred in to the study and sought out the trial via web-based posts or advertisements. Other studies have dropout rates of 47% for people referred by physicians.5

• Participants were excluded if they had another untreated sleep disorder, worked irregular schedules or had previous behavioural treatment for insomnia. This may exclude a significant proportion of people in the general population, with sleep disorders and shift work being common, and information on strategies to help sleep being widely available online.

What next in research?

Future RCTs comparing treatment response and remission rates between web-based and face-to-face CBT-I are required. Studies investigating what patient factors are predictive of a positive treatment outcome for web-based CBT-I are also important. For example, do personality, locus of control, motivation to change, rigidity of thinking, computer literacy or education level play a role? What are the future treatment implications for people who don’t respond to web-based CBT-I? Will they require a different treatment compared with treatment naïve patients?

Do these results change your practices and why?

Yes. The results of this study give us confidence as clinicians that highly motivated, computer literate, educated individuals could use web-based CBT-I as an effective option for treating insomnia, even if they have stable comorbidities. This study also shows that in the above group, clinicians could expect the benefits of web-based CBT-I to be sustained out to 1 year.