REPRESENTATIVENESS OF BROAD-CONSENT REAL-WORLD DATA: A CROSS-SECTIONAL ANALYSIS OF 3.5 MILLION INPATIENTS IN THE HELIOS SAFE MEDICAL DATA PLATFORM
Author(s)
Sabine Spänig, PhD1, Sven Hohenstein, PhD1, Sebastian Ortleb, M.Sc.2, Ulrike Wehkamp, MD2, Olaf Kannt, MD2, Ralf Kuhlen, MD3, Bollmann Andreas, MD1.
1Helios Health Institute GmbH, Leipzig, Germany, 2Helios Kliniken GmbH, Berlin, Germany, 3Fresenius SE & Co. KGaA, Bad Homburg, Germany.
1Helios Health Institute GmbH, Leipzig, Germany, 2Helios Kliniken GmbH, Berlin, Germany, 3Fresenius SE & Co. KGaA, Bad Homburg, Germany.
OBJECTIVES: Real-world data are used for outcomes research, but their value depends on how representative consenting patients are. Within the Helios Safe Medical Data (HeSaMeDa) platform covering the German Helios Hospital network, a subset provides broad consent for secondary data use. We compared this consent cohort with the entire inpatient cohort.
METHODS: This retrospective cross-sectional analysis used routinely collected data on all adults (aged ≥ 18 years) in inpatient treatment between January 2022 and September 2025. We compared case mix (ICD-10 chapters), demographics, Elixhauser comorbidity index, in-hospital outcomes, and guideline-directed medical therapy (GDMT) in a heart-failure subcohort. Odds ratios were estimated by logistic regression and metric variables by the Wilcoxon rank-sum test. Significance was set at 5%.
RESULTS: The entire cohort comprised 3,547,634 cases and the consent cohort 1,137,717 (32.1%). Case mix and demographics were broadly similar, with circulatory diagnoses in 17% vs. 18%, median age 66 vs. 68 years, and 48.7% vs. 51.0% female. The consent cohort was less acute, with fewer urgent admissions (26.4% vs. 55.5%; OR 0.29), less intensive care (8.8% vs. 13.8%; OR 0.60), lower mortality (0.8% vs. 2.8%; OR 0.29), and lower comorbidity (Elixhauser 0 vs. 2). Heart failure was the principal diagnosis in 2.1% vs. 3.6% (OR 0.57). GDMT was comparable for beta blockers (85.9% vs. 85.8%; OR 1.0) and mineralocorticoid receptor antagonists (43.7% vs. 41.1%; OR 1.1). In contrast, the consent cohort received more renin-angiotensin system inhibitors (93.7% vs. 90.7%; OR 1.5) and SGLT2 inhibitors (52.2% vs. 47.5%; OR 1.2).
CONCLUSIONS: The HeSaMeDa consent cohort provides real-world data on over one million inpatients who closely mirror the entire population. The principal difference is acuity. Because broad consent is obtained at admission, acutely ill and emergency patients are underrepresented. Consent-based real-world data are well suited to outcomes research, provided acute-care analyses account for this selection.
METHODS: This retrospective cross-sectional analysis used routinely collected data on all adults (aged ≥ 18 years) in inpatient treatment between January 2022 and September 2025. We compared case mix (ICD-10 chapters), demographics, Elixhauser comorbidity index, in-hospital outcomes, and guideline-directed medical therapy (GDMT) in a heart-failure subcohort. Odds ratios were estimated by logistic regression and metric variables by the Wilcoxon rank-sum test. Significance was set at 5%.
RESULTS: The entire cohort comprised 3,547,634 cases and the consent cohort 1,137,717 (32.1%). Case mix and demographics were broadly similar, with circulatory diagnoses in 17% vs. 18%, median age 66 vs. 68 years, and 48.7% vs. 51.0% female. The consent cohort was less acute, with fewer urgent admissions (26.4% vs. 55.5%; OR 0.29), less intensive care (8.8% vs. 13.8%; OR 0.60), lower mortality (0.8% vs. 2.8%; OR 0.29), and lower comorbidity (Elixhauser 0 vs. 2). Heart failure was the principal diagnosis in 2.1% vs. 3.6% (OR 0.57). GDMT was comparable for beta blockers (85.9% vs. 85.8%; OR 1.0) and mineralocorticoid receptor antagonists (43.7% vs. 41.1%; OR 1.1). In contrast, the consent cohort received more renin-angiotensin system inhibitors (93.7% vs. 90.7%; OR 1.5) and SGLT2 inhibitors (52.2% vs. 47.5%; OR 1.2).
CONCLUSIONS: The HeSaMeDa consent cohort provides real-world data on over one million inpatients who closely mirror the entire population. The principal difference is acuity. Because broad consent is obtained at admission, acutely ill and emergency patients are underrepresented. Consent-based real-world data are well suited to outcomes research, provided acute-care analyses account for this selection.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
RWD39
Topic
Methodological & Statistical Research, Real World Data & Information Systems, Study Approaches
Topic Subcategory
Data Protection, Integrity, & Quality Assurance, Health & Insurance Records Systems
Disease
No Additional Disease & Conditions/Specialized Treatment Areas