A Bayesian multivariate approach to estimating the prevalence of a superordinate category of disorders

Abstract

Although our approach entails additional effort (e.g., contacting authors for individual participant data), the improved accuracy of the prevalence estimates obtained is significant and therefore recommended.

The Prevalence of Anxiety Disorders During Pregnancy and the Postpartum Period

“…47 Statistical simulations suggest that estimating the prevalence of a disorder category -such as ADs -by combining studies that differ with respect to the disorders measured requires special consideration to avoid catastrophically underestimating the true prevalence. 51 We address this concern by modelling the individual ADs using a modern Bayesian multivariate approach. Specifically, the current analysis estimates the probability of having one or more AD by combining individual disorder prevalences and simulating data from a large, typical sample to estimate the probability of having one or more of those disorders.…”

“…This modelling approach has been shown to outperform other means of estimating the prevalence of a disorder category and permits us to make probabilistic statements about other facets of the data (e.g., comorbidity) not possible using traditional approaches (for details and code, see Fawcett et al). 51 Furthermore, the current approach is not limited to studies that include multiple AD prevalences, allowing for additional studies to be included measuring only one AD. Anticipating heterogeneity, we also explore potential moderators.…”

“…This risks underestimating the prevalence of having an AD. For this reason, our analyses employed a Bayesian multivariate model, discussed in detail elsewhere, 51 to describe the data actually reported. This approach models the prevalence of each individual disorder, estimates the correlations amongst the disorders, uses the resulting information to produce an unbiased estimate of the probability of having one or more disorder in a typical sample, and also produces prediction intervals reflecting variation in the "true" underlying prevalence estimates across the distribution of included samples.…”

A Bayesian multivariate meta‐analysis of prevalence data

“…To the best of our knowledge, there is only one other existing multivariate meta‐analysis of multivariate prevalence data 24 . This serves as a case study illustrating the potential improvement in estimation by jointly modeling multivariate prevalence data, particularly when incorporating additional natural constraints into the model parameterization.…”

“…This widespread use of separate univariate models is especially prevalent in the context of observational data. For instance, we found only three examples of multivariate meta‐analysis of observational data in the literature: (1) Fawcett et al 24 presented a multivariate method that used data on prevalences of individual disorders in order to estimate their overall prevalence. To our knowledge, this is the only study to perform a multivariate meta‐analysis of prevalence and incidence data.…”