Stone commented that more studies should have been included in our meta-analysis. ^{Reference Koelen, Houtveen, Abbass, Luyten, Eurelings-Bontekoe and Van Broeckhuysen-Kloth1} In particular, he suggests that six studies (references 2, 3, 12-15 cited by Stone) that were published after the index date of our literature search (March 2010) and four other studies that were published before 2010 (references 5-8 cited by Stone) could have been included, and that one review that contains randomised trials in functional dysphonia might have met our inclusion criteria (reference 9 cited by Stone). Stone’s concern is that these omissions make our meta-analysis out of date.

Apparently, the rationale behind our inclusion criteria requires further clarification. Most importantly, while previous reviews were restricted to psychodynamic psychotherapy only, ^{Reference Abbass, Kisely and Kroenke2} predominantly involved patients with less severe disorder ^{Reference Kroenke3} or included medically unexplained physical symptoms according to divergent criteria, ^{Reference Kleinstäuber, Witthöft and Hiller4} our meta-analysis examined the effectiveness of psychotherapy for severe somatoform disorder. As mentioned in our publication, ‘severe’ was defined as a diagnosis of somatoform disorder according to established criteria and treatment offered in secondary or tertiary care settings.

We chose to utilise established criteria for somatoform disorder from the psychiatric nomenclature (ICD and DSM in particular). Our main rationale was that psychiatric diagnoses contain explicit criteria about impaired daily functioning in main areas of life (social, interpersonal and occupational), and about psychological factors implicated in the disorder. We opted for these criteria because in our view these would best capture the impairments and the complicated aetiology of these disorders. For this reason, we disregarded medical diagnoses that do not always consider psychological factors implicated in the disorder, and that often use less stringent criteria regarding the duration and severity of the disorder.

The search terms we used simply would not have yielded most of the studies mentioned by Stone, because for the reasons outlined above we did not search for dissociative seizures, pseudoseizures, psychogenic non-epileptic seizures and psychogenic movement disorders. One study was excluded because Escobar et al’s ^{Reference Escobar, Rubio-Stipec, Canino and Karno5} less stringent criteria for somatisation were used (reference 8 cited by Stone). Stone also mentions a review of randomised trials for functional dysphonia (reference 9 cited by Stone), but for the same reason, these studies do not meet our inclusion criteria. Two studies mentioned in Stone’s letter meet our inclusion criteria, one of which should probably have been included after revision in June 2013 (reference 12 cited by Stone), while the other was published in October 2013, and could not have been included (reference 13 cited by Stone).

We agree with Stone that there is a paradox in including somatoform disorder while excluding individual somatic syndromes, especially given the arbitrary cut-offs and the high overlap between seemingly distinct somatic syndromes. This is at least in part a reflection of the problematic nomenclature for these disorders, ^{Reference Creed, Guthrie, Fink, Henningsen, Rief and Sharpe6,Reference Mayou, Kirmayer, Simon, Kroenke and Sharpe7} which is divided between psychiatry and the remainder of medicine. ^{Reference Sharpe8} We concur with Stone that most patients with functional somatic disorders also have other symptoms, ^{Reference Fink and Schröder9} and may even meet criteria for somatoform disorder. Yet, we cannot be sure that all patients with individual somatic syndromes do meet criteria for somatoform disorder. For this reason, and for the reasons outlined above, we did not include individual somatic syndromes. At the same time, other reviews have already summarised the effectiveness of psychotherapy for medically unexplained physical symptoms using less stringent criteria, also including some - although not all - conversion disorders. ^{Reference Kleinstäuber, Witthöft and Hiller4}

To conclude, we consider it a strength of the current meta-analysis that it has a narrow and therefore specific focus on a precise diagnostic entity, because this clearly defines the boundaries for generalisation of the findings. We acknowledge that our findings cannot be extrapolated to all fields of medicine and somatic symptom disorders. The results from our meta-analysis specifically apply to patients with somatoform disorder according to established (psychiatric) diagnostic criteria that received psychotherapy in secondary and tertiary care.