Completeness of reporting of overnight hospitalisations to a psoriasis pharmacovigilance register and the National Health Service Wales Informatics Service

Mason, Kayleigh J.; Ali, Hassan; Reynolds, Nick J.; Burden, Anthony D.; Smith, Catherine H.; Evans, Ian; McElhone, Kathleen; Steer, Oliver; Tahir, Saliha; Griffiths, Christopher E. M.

Abstract

Background: The British Association of Dermatologists Biologic Interventions Register (BADBIR) is a longitudinal, observational, web-based pharmacovigilance register of patients with moderate-to-severe psoriasis. Patients are currently recruited from 153 dermatology centres in the UK and Republic of Ireland. The aim of BADBIR is to explore the long-term safety of biologic compared to conventional systemic therapies. Patients registered in 10 centres in Wales were flagged with the National Health Service Wales Informatics Service (NHS WIS) to maximise the capture of overnight hospitalisations.
Objective: To investigate the overlap in reporting of overnight hospitalisations to BADBIR and NHS WIS.
Methods: Data from 27/12/2007 to 28/12/2014 on 490 patients were received from NHS WIS. Overnight hospitalisations occurring after the start date of BADBIR registration therapy were classified as serious adverse events. Events were cross-referenced for admission and discharge dates (±7 days), and ICD-10 codes from NHS WIS with qualitative descriptions in BADBIR.
Results: Of the 490 patients flagged, 86 overnight hospitalisations for 62 patients were recorded with BADBIR and 244 overnight hospitalisations for 120 patients recorded with NHS WIS. 51 patients with overnight hospitalisation were reported in both datasets, but only 62 of 73 overnight hospitalisations were matched to events already on the BADBIR database. In total, 268 overnight hospitalisations were recorded for 131 patients.
Conclusion: The findings suggest that the reporting of overnight hospitalisations is more complete in NHS WIS than in BADBIR. Possible reasons contributing to the substantial proportion of missing overnight hospitalisations include: lag time in data entry (mean 283 days between admission date and event reporting to BADBIR; unknown for NHS WIS), admission to hospitals remote from the dermatology recruiting centre, patient recall bias, and incomplete information from recruiting centres. Failure to use linked sources may lead to underestimated risks of overnight hospitalisations.

Citation

Mason, K. J., Ali, H., Reynolds, N. J., Burden, A. D., Smith, C. H., Evans, I., …Griffiths, C. E. . M. (2017, August). Completeness of reporting of overnight hospitalisations to a psoriasis pharmacovigilance register and the National Health Service Wales Informatics Service. Paper presented at 33rd International Conference on Pharmacoepidemiology & Therapeutic Risk Management, Palais de Congres de Montreal, Montreal, Canada

• Presentation Conference Type: Conference Paper (unpublished)
• Conference Name: 33rd International Conference on Pharmacoepidemiology & Therapeutic Risk Management
• Conference Location: Palais de Congres de Montreal, Montreal, Canada
• Start Date: Aug 26, 2017
• End Date: Aug 30, 2017
• Deposit Date: Sep 6, 2023
• Publisher URL: https://onlinelibrary.wiley.com/doi/10.1002/pds.4275
• Additional Information: http://www.pharmacoepi.org/meetings/33ICPE/