Merck (France)

BackgroundSharing information derived from individual patient data is restricted by regulatory frameworks due to privacy concerns. Generative artificial intelligence can generate shareable virtual patient populations as proxies for sensitive reference datasets. Explicit demonstration of privacy is demanded.ObjectiveThis study evaluated whether a privacy-by-design technique called “avatars” can generate synthetic datasets replicating all reported information from randomized clinical trials (RCTs).MethodsWe generated 2160 synthetic datasets from two phase 3 RCTs for patients with multiple sclerosis (NCT00213135 and NCT00906399; n=865 and 1516 patients) with different configurations to select one synthetic dataset with optimal privacy and utility for each. Several privacy metrics were computed, including protection against distance-based membership inference attacks. We assessed fidelity by comparing variable distributions and assessed utility by checking that all end points reported in the publications had the same effect directions, were within the reported 95% CIs, and had the same statistical significance.ResultsProtection against membership inference attacks was the hardest privacy metric to optimize, but the technique yielded robust privacy and replication of the primary end points (in 72.5% and 80.8% of the 1080 generated datasets). Utility was uneven across the variables and end points, such that information about some end points could not be captured. With optimized generation configurations, we selected one dataset from each RCT replicating all efficacy end points of the placebo and approved treatment arms while maintaining satisfactory privacy (hidden rate: 85.0% and 93.2%).ConclusionsGenerating synthetic RCT datasets replicating primary and secondary efficacy end points is possible while achieving a satisfactory and explicit level of privacy. To show the potential of this method to unlock health data sharing, we released both placebo arms as open datasets.

BackgroundThis sub-analysis of the SPECTRA (Study of Prescribing patterns and Effectiveness of Ceftolozane/Tazobactam [C/T] Real-world Analysis) study aimed to describe the clinical characteristics, treatment patterns, and outcomes of hospitalized pneumonia patients treated with ceftolozane/tazobactam (C/T).Table 1.Patient characteristics and Microbiology2 Immunocompromised defined as: ‘Immunocompromised present’ [raw item] or ‘Hematologic Malignancy present’ or ‘Solid Tumor present’ or ‘Transplant=Yes’.MethodsSPECTRA was a multicenter observational study conducted in 7 countries. It included adult patients (age≥18yrs) who received ≥48-hours C/T treatment (n=617). Retrospective data were extracted from medical records covering a 6-month period prior to the index date, up to 30 days after the last dose of C/T or until death. Descriptive statistics were used to analyze clinical success, all-cause in-hospital mortality (ACHM), intensive care unit (ICU) admission, length of stay (LOS) in the subcategory of pneumonia (n=182).Table 2.Outcomes with Ceftolozane/tazobactam and Treatment duration (days) (Treatment interruptions not included)*Other conditions: Bile duct anostomic leakage, Brain-vascular hemorrhagic accident, CMV infection, Encephalic Death, Hemorrhage, Invasive fungal infection, Necrotizing pancreatitis, Post operative complication, Renal amyloidosis, Spontaneous rupture of the spleen, Septic shock, Hemodynamic dysfunction, Kidney failure, Stable angina grade I-II CCS, Probable Clostridium difficile colitis, Progression of AML, Refractory Hypoxemia, Refractory hematologic malignancy, Subarachnoid Hemorrhage, Solid tumor, TracheobronchitisResultsThe median duration of C/T treatment was 11.0 days, median time from index hospitalization admission to C/T initiation was 19.0 days, and from the first microbiological sample to C/T initiation 6.0 days. All-cause in-hospital mortality was 32.4% (n=59), with a median time from the index to death of 28.0 days. The causes of death included sepsis (38.9%) and pneumonia (18.6%) and 67% (n=122) patients were admitted to the ICU during the index hospitalization. Pseudomonas aeruginosa (PSA) was the most commonly isolated pathogen (88% of patients), with MDR PsA present in 68.4% (n=80) of pneumonia patients.ConclusionIn hospitalized patients with pneumonia treated with C/T in a real-world setting, clinical characteristics, treatment patterns, and outcomes provide insights into the use and effectiveness of C/T.DisclosuresEmre Yucel, PhD, Merck: I am a full time Merck Employee and own stocks in the retirement plan provided by Merck.|Merck: Stocks/Bonds (Public Company) David Paterson, bioMerieux: Grant/Research Support|bioMerieux: Honoraria|Merck: Advisor/Consultant|Merck: Grant/Research Support|Merck: Honoraria|Pfizer: Advisor/Consultant|Pfizer: Grant/Research Support|Pfizer: Honoraria|Shionogi: Grant/Research Support|Shionogi: Honoraria Florian Thalhammer, MD, MSD: Advisor/Consultant Stefan Kluge, Prof. Dr. med., Merck & Co: Advisor/Consultant|Merck & Co: Board Member Mike Allen, PhD, Merck: I am a full time Merck Employee and own stocks in the retirement plan provided by Merck.|Merck: Stocks/Bonds (Public Company) Brune Akrich, MD, Merck: I am a full time Merck Employee and own stocks in the retirement plan provided by Merck.|Merck: Stocks/Bonds (Public Company) Yanbing Zhou, PhD, Merck: I am a full time Merck Employee and own stocks in the retirement plan provided by Merck.|Merck: Stocks/Bonds (Public Company)

Direction of arrival (DoA) estimation is a common sensing problem in radar, sonar, audio, and wireless communication systems. It has gained renewed importance with the advent of the integrated sensing and communication paradigm. To fully exploit the potential of such sensing systems, it is crucial to take into account potential hardware impairments that can negatively impact the obtained performance. This study introduces a joint DoA estimation and hardware impairment learning scheme following a model-based approach. Specifically, a differentiable version of the multiple signal classification (MUSIC) algorithm is derived, allowing efficient learning of the considered impairments. The proposed approach supports both supervised and unsupervised learning strategies, showcasing its practical potential. Simulation results indicate that the proposed method successfully learns significant inaccuracies in both antenna locations and complex gains. Additionally, the proposed method outperforms the classical MUSIC algorithm in the DoA estimation task.

This comparative non-interventional study using data from the French National Health Database (Système National des Données de Santé) investigated real-world (cumulative) live birth outcomes following ovarian stimulation, leading to oocyte pickup with either originator recombinant human follicle-stimulating hormone (r-hFSH) products (alfa or beta), r-hFSH alfa biosimilars, or urinaries including mainly HP-hMG (menotropins), and marginally u-hFSH-HP (urofollitropin). Using data from 245,534 stimulations (153,600 women), biosimilars resulted in a 19% lower live birth (adjusted odds ratio (OR) 0.81, 95% confidence interval (CI) 0.76-0.86) and a 14% lower cumulative live birth (adjusted hazard ratio (HR) 0.86, 95% CI 0.82-0.89); and urinaries resulted in a 7% lower live birth (adjusted OR 0.93, 95% CI 0.90-0.96) and an 11% lower cumulative live birth (adjusted HR 0.89, 95% CI 0.87-0.91) versus originator r-hFSH alfa. Results were consistent across strata (age and ART strategy), sensitivity analysis using propensity score matching, and with r-hFSH alfa and beta as the reference group.

Abstract Background: In France, no specific information on the Highly Active Relapsing-Remitting Multiple Sclerosis (HA-RRMS) population is available.Objective: To describe the epidemiology and therapeutic management of HA-RRMS patients in France.Methods: In this cohort study, HA-RRMS patients were identified in the health data system with a new algorithm using outpatient healthcare consumption and hospital discharge data. Results: Over 2010–2015, 9,596 incident HA-RRMS patients were identified (sex ratio: 2.8; mean age 39.9 years old) and followed-up for 4.0 years, on average. In 2015, the incidence and mortality rates in patients aged 20 and above were 3.6 and 389 per 100,000, respectively (lower mortality than in the MS population but twice as high as in the French population). During the study, 39.7% of patients took a disease-modifying drug (DMD) and 83.5% a high efficacy DMD (HE-DMD, fingolimod or natalizumab, with a mean treatment duration of 3.5 years). When patients treated by an HE-DMD required to switch to another treatment, it was more often to the other HE-DMD (more often from natalizumab to fingolimod) than to a DMD. Conclusions: According to this first real-world study, HA-RRMS patients could represent 8% of MS patients. The results validated the algorithm to identify HA-RRMS patients.