We analyzed the NIS database to select the patients between 18 and 75 years. Based on the validated International Classification of Diseases, 9th Revision, and Clinical Modification (ICD-9-CM) diagnosis codes, the patients with a primary diagnosis (DX1) of bipolar disorder and a comorbid diagnosis (DX2 to DX25) of BPD were identified as the cases. This patient cohort was compared with the control cohort which included patients with a primary diagnosis (DX1) of bipolar disorder only. In HCUP databases, more than 14,000 ICD-9-CM diagnosis codes are mentioned [16]. Bipolar disorder was identified using ICD-9-CM diagnosis codes 296.40, 296.41, 296.42, 296.43, 296.44, 296.50, 296.51, 296.52, 296.53, 296.54, 296.60, 296.61, 296.62, 296.63, 296.64, and 296.7; and BPD was identified using ICD-9-CM diagnosis code 301.83. We also performed a retrospective cohort analysis to derive a control cohort that had similar demographic characteristics to the cases. The controls were matched for age, gender and race using the case-control matching application in SPSS version 23 (IBM Corp., Armonk, NY, USA). Finally, a separate cohort was extracted from the NIS of patients (18–75 years) with primary diagnoses of BPD to compare their sociodemographic characteristics and hospital outcomes with bipolar patients with and without BPD.
