Is CBT really losing efficacy?

Michael J. Scott feels the case is not proven.

In the July issue ‘Digest’ you report a meta-analysis of 70 CBT studies for depression conducted by Johnsen and Friborg (2015) and opined ‘CBT doesn’t seem to be helping reduce depression symptoms as much today as it used to when it was first developed in the 1970s’. But this conclusion may be premature, inspection of Table 1 of Johnsen and Friborg’s study shows that from 1977 up to and including the millennium 85 per cent of studies were randomised controlled trials (RCTs) but from 2001 to 2014 the comparable figure was 65 per cent. One of the hallmarks of an RCT is blind assessment, using a standardised diagnostic interview. Thus there can be no certainty that populations treated post the millennium are comparable to those before. The results of Johnsen and Friborg’s (2015) meta-analysis could more parsimoniously be interpreted as indicating that CBT does not work well when the population being addressed is poorly defined. Historically, CBT treatments have largely been diagnosis-specific, arguably failure to match disorder/difficulty with protocol will result in sub-optimal outcomes.

The authors of this meta-analysis did not analyse outcome in terms of loss of diagnostic status to determine whether this has changed since the 1970s – from a public health point of view changes on continuous measures, though important, are essentially surrogate measures. It would also be interesting to see a re-analysis of the included studies comparing outcomes in studies using blind standardised diagnostic assessments with those not using this ‘gold standard’.

Michael J. Scott



Johnsen, T.J. & Friborg, O. (2015). The effects of CBT as an anti-depressive treatment is falling. Psychological Bulletin [Advance online publication]. doi:10.1037/bul0000015

BPS Members can discuss this article

Already a member? Or Create an account

Not a member? Find out about becoming a member or subscriber