Positive predictive value of an ICD-10-based operationalization of bipolar II disorder for register-based research.

Objective: Bipolar II disorder (BD-II) is a subtype of bipolar disorder characterized by recurrent episodes of depression and hypomania, without full manic episodes. Unfortunately, BD-II is not included as a diagnostic code in the ICD-10, which means that the Danish Psychiatric Central Research Register (DPCRR), where diagnoses are recorded according to the ICD-10, cannot be used to study BD-II without further ado. The aim of this study was to investigate whether BD-II can be operationalized retrospectively based on ICD-10 diagnoses with sufficient positive predictive value to allow for studies of BD-II using data from the DPCRR.

Materials and methods: We operationalized BD-II a priori based on a set of criteria (e.g. a minimum of two mood episodes labelled with ICD-10 diagnostic codes of hypomania or bipolar depression - at least one being bipolar depression - and no manic/mixed episodes). The positive predictive value of this operationalization was then examined by reviewing (two independent reviewers) the electronic health records (EHRs) of 147 patients from the Psychiatric Services of the Central Denmark Region matching the ICD-10-based operationalization of BD-II.

Results: For 107 of the 147 patients, the EHR review confirmed that BD-II was the most likely diagnosis, resulting in a positive predictive value of 73% for the ICD-10-based operationalization of BD-II.

Conclusions: This study suggests that, while not perfect in terms of positive predictive value, the proposed ICD-10-based operationalization will allow for studies of 'predominantly BD-II' using data from the DPCRR with sufficient validity.