Transportability of Overall Survival in Multiple Myeloma From the US to Germany: A Benchmarking Study
Author(s)
Elsie Horne1, Marco Dibonaventura, PhD2, Per-Olof Thuresson, PhD3, Felipe Castro, PhD3, Christoph Buhl, MD4, Mohamed S. Ali, PharmD, SM5, Harlan Pittell, PhD5, Philani Mpofu, PhD5, Blythe Adamson, MPH, PhD5.
1Research Scientist, Flatiron Health UK, London, United Kingdom, 2Pfizer, New York, NY, USA, 3Hoffman-La Roche AG, Basel, Switzerland, 4Flatiron Health Germany, Berlin, Germany, 5Flatiron Health, New York, NY, USA.
1Research Scientist, Flatiron Health UK, London, United Kingdom, 2Pfizer, New York, NY, USA, 3Hoffman-La Roche AG, Basel, Switzerland, 4Flatiron Health Germany, Berlin, Germany, 5Flatiron Health, New York, NY, USA.
OBJECTIVES: The transportability of real-world evidence across countries is essential for health technology assessments. This study examines whether overall survival (OS) estimates from patients with multiple myeloma (MM) in the US can be transported to Germany.
METHODS: This retrospective cohort study included patients diagnosed with MM and treated in the first line (1L) between 2017-2024 from the GermanOncology registry and the US Flatiron Health Research Database. Patient characteristics and 1L treatment patterns were compared across the Germany and US cohorts. OS curves (Kaplan-Meier estimates) from 1L start were compared between cohorts, adjusting the patient characteristics and 1L treatment patterns of the US cohort to resemble that of the German cohort using inverse odds of sampling weights (IOSWs).
RESULTS: The Germany and US cohorts comprised 486 and 9881 patients, of which 47% and 45% were female, respectively. Median (IQR) ages in the Germany and US cohorts were 72 years (62-79) and 70 years (62-77), respectively. Classes of 1L induction therapy were different across cohorts. Compared with the US cohort, the German cohort was more likely to receive chemotherapy and a proteasome inhibitor (32% vs 8%), yet less likely to receive an immunomodulatory agent and a proteasome inhibitor (6% vs 45%). Revised ISS stage, ECOG, and cytogenetic tests were omitted from the IOSWs due to concerns about data quality. Five-year OS probabilities (95% CI) from 1L were 63% (56%-72%) in the Germany cohort, 55% (53%-56%) in the US cohort, and 46% (44%-48%) in the IOSW US cohort.
CONCLUSIONS: Uncertainty remains regarding the transportability of OS from the US to Germany in MM. Adjustments were limited by the exclusion of certain prognostic variables, due to concerns over the quality of the Germany data. While adjustments for 1L treatment patterns were made, there may be unmeasured factors influencing treatment decisions in the two countries.
METHODS: This retrospective cohort study included patients diagnosed with MM and treated in the first line (1L) between 2017-2024 from the GermanOncology registry and the US Flatiron Health Research Database. Patient characteristics and 1L treatment patterns were compared across the Germany and US cohorts. OS curves (Kaplan-Meier estimates) from 1L start were compared between cohorts, adjusting the patient characteristics and 1L treatment patterns of the US cohort to resemble that of the German cohort using inverse odds of sampling weights (IOSWs).
RESULTS: The Germany and US cohorts comprised 486 and 9881 patients, of which 47% and 45% were female, respectively. Median (IQR) ages in the Germany and US cohorts were 72 years (62-79) and 70 years (62-77), respectively. Classes of 1L induction therapy were different across cohorts. Compared with the US cohort, the German cohort was more likely to receive chemotherapy and a proteasome inhibitor (32% vs 8%), yet less likely to receive an immunomodulatory agent and a proteasome inhibitor (6% vs 45%). Revised ISS stage, ECOG, and cytogenetic tests were omitted from the IOSWs due to concerns about data quality. Five-year OS probabilities (95% CI) from 1L were 63% (56%-72%) in the Germany cohort, 55% (53%-56%) in the US cohort, and 46% (44%-48%) in the IOSW US cohort.
CONCLUSIONS: Uncertainty remains regarding the transportability of OS from the US to Germany in MM. Adjustments were limited by the exclusion of certain prognostic variables, due to concerns over the quality of the Germany data. While adjustments for 1L treatment patterns were made, there may be unmeasured factors influencing treatment decisions in the two countries.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
HTA340
Topic
Health Technology Assessment
Topic Subcategory
Decision & Deliberative Processes
Disease
Oncology