Contributions
Abstract: P1034
Type: Poster Sessions
Abstract Category: Clinical aspects of MS - Clinical assessment tools
Introduction: Annual relapse rate is commonly used as an indicator of effectiveness for MS disease-modifying therapy (DMT). An algorithm developed by Ollendorf et al. combining MS-related hospitalization and outpatient steroid use to identify relapse in claims is currently used by most research using claims data, and was validated on patients using glatiramer acetate, interferons or natalizumab. Since steroid is part of the treatment regimen for OCR initiation (per label), the conventional algorithm may overestimate post-treatment relapse in OCR patients.
Objective: To evaluate the potential of misclassification using an existing algorithm identifying relapse in insurance claims for multiple sclerosis (MS) patients receiving ocrelizumab (OCR).
Methods: MS patients initiating OCR from April 2017 to December 2017 were identified in the Pharmetrics Plus insurance claims database in the US. Patients were required to have at least 2 claims indicating MS diagnosis, at least 2 OCR IVs 12 to 18 days apart, and without other DMTs within 6 months of initiation. Steroid use in medical claims around each infusion date for OCR was assessed using HCPCS codes. Among patients with at least 3 months post-initiation enrollment, relapse events were identified in the 3 months post initiation using the traditional algorithm and an updated method excluding steroid use on the same day as the OCR IV.
Results: In total, 92.6% of the 955 MS patients who initiated OCR had claims indicating methylprednisolone or hydrocortisone injection on the date of 1st OCR infusion. Similar steroid use patterns were observed for the 2nd OCR loading dose. Among 266 patients who had at least 3 months of enrollment post OCR initiation, 91.0% had a “relapse” as defined by the traditional algorithm. When same day steroid use was excluded, the proportion of patients with a relapse dropped to 7.1%.
Conclusions: This study shows that recommended steroids for OCR initiation were used in the real world, consistent with the OCR label. The existing, widely used claims-based algorithm for identifying MS relapse may be inappropriate to assess outcomes in OCR patients. This misclassification should be addressed in future comparative effectiveness research to reduce bias.
Disclosure: Sponsored by Genentech, Inc. E. Yang is an employee of Genentech, Inc., and a shareholder of F. Hoffmann-La Roche Ltd; N.J. Engmann is an employee of Genentech, Inc., and a shareholder of F. Hoffmann-La Roche Ltd; W.S. Yeh is an employee of Genentech, Inc., and a shareholder of F. Hoffmann-La Roche Ltd; L. Julian is an employee of Genentech, Inc., and a shareholder of F. Hoffmann-La Roche Ltd.
Abstract: P1034
Type: Poster Sessions
Abstract Category: Clinical aspects of MS - Clinical assessment tools
Introduction: Annual relapse rate is commonly used as an indicator of effectiveness for MS disease-modifying therapy (DMT). An algorithm developed by Ollendorf et al. combining MS-related hospitalization and outpatient steroid use to identify relapse in claims is currently used by most research using claims data, and was validated on patients using glatiramer acetate, interferons or natalizumab. Since steroid is part of the treatment regimen for OCR initiation (per label), the conventional algorithm may overestimate post-treatment relapse in OCR patients.
Objective: To evaluate the potential of misclassification using an existing algorithm identifying relapse in insurance claims for multiple sclerosis (MS) patients receiving ocrelizumab (OCR).
Methods: MS patients initiating OCR from April 2017 to December 2017 were identified in the Pharmetrics Plus insurance claims database in the US. Patients were required to have at least 2 claims indicating MS diagnosis, at least 2 OCR IVs 12 to 18 days apart, and without other DMTs within 6 months of initiation. Steroid use in medical claims around each infusion date for OCR was assessed using HCPCS codes. Among patients with at least 3 months post-initiation enrollment, relapse events were identified in the 3 months post initiation using the traditional algorithm and an updated method excluding steroid use on the same day as the OCR IV.
Results: In total, 92.6% of the 955 MS patients who initiated OCR had claims indicating methylprednisolone or hydrocortisone injection on the date of 1st OCR infusion. Similar steroid use patterns were observed for the 2nd OCR loading dose. Among 266 patients who had at least 3 months of enrollment post OCR initiation, 91.0% had a “relapse” as defined by the traditional algorithm. When same day steroid use was excluded, the proportion of patients with a relapse dropped to 7.1%.
Conclusions: This study shows that recommended steroids for OCR initiation were used in the real world, consistent with the OCR label. The existing, widely used claims-based algorithm for identifying MS relapse may be inappropriate to assess outcomes in OCR patients. This misclassification should be addressed in future comparative effectiveness research to reduce bias.
Disclosure: Sponsored by Genentech, Inc. E. Yang is an employee of Genentech, Inc., and a shareholder of F. Hoffmann-La Roche Ltd; N.J. Engmann is an employee of Genentech, Inc., and a shareholder of F. Hoffmann-La Roche Ltd; W.S. Yeh is an employee of Genentech, Inc., and a shareholder of F. Hoffmann-La Roche Ltd; L. Julian is an employee of Genentech, Inc., and a shareholder of F. Hoffmann-La Roche Ltd.