(114) Exploring the Feasibility and added Value of a Comorbidity Index Score in a Pharmacovigilance Database

Background: In pharmacovigilance (PV), the review of comorbidities is an important element in safety monitoring.
The Charlson Comorbidity Index (CCI) method is widely used in epidemiology on clinical and real world data to quantify multiple comorbidity conditions within a patient, but little has been published on the quantification of comorbidity conditions on individual case safety reports (ICSR) in PV databases. We aimed to transfer the CCI method for use on MedDRA coded data in PV databases.

Objectives: To explore feasibility and value add of a comorbidity index score on patient reports in a spontaneous adverse event database for efficient safety monitoring

Methods: We conducted a literature review to identify methods to quantify patient comorbidities utilizing a comorbidity index or risk score. CCI with its age-adjusted variation was selected as the method.
We aimed to map all CCI comorbidity conditions to established MSSO standard MedDRA queries, company standard MedDRA queries or High Level Terms (HLT).
The 16 comorbidity conditions are assigned scores (1, 2, 3 or 6) and summed to give a comorbidity score per patient. To retrieve these comorbidity conditions for the CCI, we included 3 elements of ICSRs: adverse events, indications for suspect and concomitant medications, medical history and concurrent conditions.
To test and operationalize its use on our corporate safety database, a Spotfire workbook was developed with specific views showing the comorbidity index at the aggregated case series level and at the patient level. We tested it for oncology products in typical safety evaluation tasks.

Results: We successfully implemented a comorbidity index score based on the CCI. Of the 16 comorbidity conditions, we were able to map 15 to standard MedDRA queries and HLTs and combinations thereof. This will allow efficient maintenance of these definitions.
Data completeness in ICSRs varies between products and report types and should be taken into account when interpreting results. In particular, the comorbidity score from ICSR cannot be used to predict mortality as in its original application.
The Spotfire workbook allows for efficient identification of patients with specific combinations of comorbidities and the score facilitates efficient sorting, filtering and clustering of patient reports during the safety review as well as observing trends over time in reported comorbidities.

Conclusions: Our results suggest that using a comorbidity index score in PV databases could be a valuable addition to existing comorbidity review practices. It may contribute to and enhance the efficient management of various safety review tasks, especially for medicines with indications in elderly and frail patients. Further testing is planned with a wider range of products and therapeutic areas.