Janse et al investigated the effect of two variants of internet-based cognitive–behavioural therapy (iCBT) for chronic fatigue syndrome (CFS): iCBT with protocol-driven feedback and iCBT with feedback on demand.Reference Janse, Worm-Smeitink, Bleijenberg, Donders and Knoop¹

First, it should be acknowledged that CBT trials for participants with CFS have a high preselection bias, i.e. self-selection, since, according to another study by two of the authors of Janse et al, patients seem to be sceptical about psychological interventions.Reference Twisk and Corsius² Janse et al’s study reported ‘clinically relevant depressive symptoms’ in both iCBT groups (protocol-driven feedback iCBT group 31%, feedback-on-demand iCBT group 29%), while depression and other psychological conditions that could explain ‘chronic fatigue’ exclude the diagnosis CFS.Reference Twisk and Corsius² It is feasible that many patients who improved had depression, not CFS.

Comparing the number of patients working full-time in this studyReference Janse, Worm-Smeitink, Bleijenberg, Donders and Knoop¹ with other studies, for example Sunnquist et al,Reference Sunnquist, Jason, Nehrke and Goudsmit³ the CFS (?) patients can be classified as ‘mild cases’.Reference Twisk and Corsius² Since CFS is a heterogeneous condition, the results of this studyReference Janse, Worm-Smeitink, Bleijenberg, Donders and Knoop¹ cannot be generalised to CFS.

Drop-out rates are not reported but the authors assumed a drop-out rate of 15% when deciding on sample sizeReference Janse, Worm-Smeitink, Bleijenberg, Donders and Knoop¹ and other studies by the same group have reported even higher drop-out rates.Reference Twisk and Corsius² They state ‘a substantial number of patients did not fully adhere to the interventions’.Reference Janse, Worm-Smeitink, Bleijenberg, Donders and Knoop¹ One could also question whether accessing treatment modules and email contact are ‘strict criteria’ to guarantee adherence to the graded activity protocol. Although the authors state that ‘The treatment is tailored to a patient's current activity pattern as assessed with actigraphy’, (increased) activity levels was not included in the adherence criteria.

According to the authors, both iCBT conditions are efficacious, since 29/80 (36%) in the protocol-driven feedback iCBT group and 34/80 (43%) in the feedback-on-demand iCBT achieved the ‘normal range’ for Checklist Individual Strength fatigue severity, compared with 12/80 (15%) in the waiting list group.Reference Janse, Worm-Smeitink, Bleijenberg, Donders and Knoop¹ However, the treatment effects of the protocol-driven feedback iCBT and feedback-on-demand iCBT in the study are by far insufficient to achieve ‘normal levels of fatigue’ (Checklist Individual Strength fatigue severity ≤27) as defined in another study by two of the authors of Janse et al.Reference Twisk and Corsius² Looking at the Sickness Impact Profile 8 and Medical Outcomes Survey Short Form-36 physical functioning scores after the intervention at the group level (scores at the individual level are not reported), both iCBT groups would still be qualified as ‘severely disabled’.

The effect of protocol-driven feedback iCBT and feedback-on-demand iCBT on objective measures are not reported, but other studies by the research group have shown that a CBT protocol has no effect on (low) physical activity levels, number of hours worked or cognitive test scores.Reference Twisk and Corsius²

The authors label their intervention CBT.Reference Janse, Worm-Smeitink, Bleijenberg, Donders and Knoop¹ However, looking at the protocol, the intervention investigated not only incorporated CBT, aimed at ‘behaviours and beliefs’ perpetuating ‘fatigue and impairment’, but also included a graded activity programme, known as graded exercise therapy (GET). Several large-scale patient surveys and studies, for example Cheshire et al,Reference Cheshire, Ridge, Clark and White⁴ indicate that CBT, especially when combined with GET, can cause iatrogenic harm and is not safe.Reference Twisk⁵

In conclusion, the study does not substantiate the claim that iCBT/GET for CFS is efficacious, while there are several indications CBT/GET is not a safe therapy.