Meta-analysis of cognitive behaviour therapy and selective serotonin reuptake inhibitors for the treatment of hypochondriasis: Implications for trial design

Abstract

BackgroundClassification of hypochondriasis as an obsessive-compulsive and related disorder in the International Classification of Diseases 11th Revision (ICD-11) has generated new heuristics for treatment of this common, chronic and disabling disorder. Standard treatment involves cognitive behaviour therapy (CBT) or selective serotonin reuptake inhibitors (SSRIs), but no meta-analysis has so far considered hypochondriasis as a structured diagnosis or assessed the role of medication. A clearer understanding of the relative effectiveness of these interventions and identification of clinically relevant factors moderating the treatment response is needed for clinical guideline development. MethodsThe current systematic review and meta-analysis of interventions for hypochondriasis was preregistered on PROSPERO (CRD42020185768) and follows PRISMA guidelines. We searched MEDLINE, PsycINFO, and Cochrane Library databases until July 2021 for randomized controlled trials (RCTs) of interventions for patients diagnosed with hypochondriasis (or historical diagnostic equivalents). We assessed aspects of study quality using: the CONSORT Checklist for evaluation of RCTs, the Cochrane Risk of Bias 2 tool, researcher allegiance and treatment fidelity. The primary outcome was improvement in hypochondriasis symptoms, comparing intervention and control groups at trial endpoint. Moderator variables were assessed using subgroup and meta-regression analyses. ResultsSearches identified 13 randomised controlled trials (RCTs) (N = 1405); 12 included CBT (N = 1212) and three included SSRI (N = 193) arms as the experimental intervention. Random effects meta-analysis yielded a moderate-to-large effect size for CBT versus all controls (g = −0.70 [95% CI -0.99 to −0.41], k = 18, I2 = 81.1%). Funnel plot asymmetry indicated possible publication bias and two potentially missing trials, reducing the effect size (g = −0.60 [95% CI -0.88 to −0.32]). Subgroup analysis showed that choice of control significantly moderated effect size, with those in CBT vs. wait-list (g = −1.32 [95% CI -1.75 to −0.90], k = 7, I2 = 0%) being double those of CBT vs. psychological or pharmacological placebo controls (g = −0.58 [95% CI -0.95 to −0.22], k = 7, I2 = 82%). Analysis of studies directly comparing CBT and SSRIs found a numerical, but not statistical advantage for SSRIs (g = 0.21 [95% CI -0.46 to 0.87], k = 2, I2 = 58.34%) and a modest effect size emerged for SSRIs vs. pill placebo (g = −0.29 [95% CI -0.57 to −0.01], k = 3, I2 = 0%). Most studies (11/13) were rated as high on potential researcher allegiance bias in favour of CBT. Meta-regressions revealed that effect sizes were larger in younger participants, and smaller in better quality and more recent RCTs and those with greater CBT fidelity. ConclusionCBT and SSRIs are effective in the acute treatment of hypochondriasis, with some indication that intervention at a younger age produces better outcomes for CBT. In the case of CBT, effect sizes appear to have been significantly inflated by the use of wait list controls, and researcher allegiance bias. We recommend that a definitive, adequately controlled trial, designed with respect to the methodological issues raised in this meta-analysis, is needed to determine the magnitude effects for CBT and SSRIs with confidence and the long-term effect of treatments, to inform mental health service provision for this overlooked patient group.