A5048360798- Healthcare Policy and Management
- Health Systems, Economic Evaluations, Quality of Life
- Obesity, Physical Activity, Diet
- Genetic Associations and Epidemiology
- Health and Medical Studies
- Global Health Care Issues
- Nutritional Studies and Diet
- Genetic and phenotypic traits in livestock
- Chronic Disease Management Strategies
- Data-Driven Disease Surveillance
- Bayesian Methods and Mixture Models
- Health disparities and outcomes
- Economics of Agriculture and Food Markets
- Advanced Causal Inference Techniques
- Statistical Methods in Clinical Trials
- Chronic Obstructive Pulmonary Disease (COPD) Research
- Statistical Distribution Estimation and Applications
- Amyotrophic Lateral Sclerosis Research
- Statistical Methods and Inference
- Medication Adherence and Compliance
- Scientific Computing and Data Management
- Statistical Methods and Bayesian Inference
- Privacy-Preserving Technologies in Data
- Radiomics and Machine Learning in Medical Imaging
- demographic modeling and climate adaptation
Novartis (Germany)
2024-2025
Ludwig-Maximilians-Universität München
2017-2023
Helmholtz Zentrum München
2016-2022
University of Stuttgart
2022
Deutsches Diabetes-Zentrum e.V.
2019-2020
German Center for Diabetes Research
2017-2020
Heinrich Heine University Düsseldorf
2019-2020
Munich Business School
2020
Institute of Health Economics
2016-2019
Center for Environmental Health
2019
Machine-learning classifiers mostly offer good predictive performance and are increasingly used to support shared decision-making in clinical practice. Focusing on practicability, this study evaluates prediction of patient-reported outcomes (PROs) by eight supervised including a linear model, following hip knee replacement surgery.NHS PRO data (130,945 observations) from April 2015 2017 were train test predict binary postoperative improvement based minimal important differences. Area under...
This article discusses the augmented inverse propensity weighted (AIPW) estimator as an for average treatment effects. The AIPW combines both properties of regression-based and probability (IPW) is therefore a "doubly robust" method in that it requires only either or outcome model to be correctly specified but not both. Even though this has been known years, rarely used practice. After explaining proving double robustness property, I conduct simulation study compare efficiency with IPW...
Abstract Consumer wearables and sensors are a rich source of data about patients’ daily disease symptom burden, particularly in the case movement disorders like Parkinson’s (PD). However, interpreting these complex into so-called digital biomarkers requires complicated analytical approaches, validating sufficient unbiased evaluation methods. Here we describe use crowdsourcing to specifically evaluate benchmark features derived from accelerometer gyroscope two different datasets predict...
Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease where substantial heterogeneity in clinical presentation urgently requires better stratification of patients for the development drug trials and care. In this study we explored through crowdsourcing approach, DREAM Prize4Life ALS Stratification Challenge. Using data from >10,000 1479 community-based patient registers, more than 30 teams developed new approaches machine learning clustering, outperforming best current...
Open Science is encouraged by the European Union and many other political scientific institutions. However, practice proving slow to change. We propose, as early career researchers, that it our task change research into open commit principles.
The statistical analysis of health care cost data is often problematic because these are usually non-negative, right-skewed and have excess zeros for non-users. This prevents the use linear models based on Gaussian or Gamma distribution. A common way to counter this Two-part Tobit models, which makes interpretation results more difficult. In study, I explore a distribution from Tweedie family distributions that can simultaneously model probability zero outcome, i.e. being non-user...
Abstract. Mindfulness refers to a stance of nonjudgmental awareness present-moment experiences. A growing body research suggests that mindfulness may increase cognitive resources, thereby buffering stress. However, existing models have not achieved consensus on how should be operationalized. As the sound measurement is foundation needed before substantial hypotheses can supported, we propose novel way gauging psychometric quality instrument (the Freiburg Inventory; FMI). Specifically,...
Abstract Objective Because it is impossible to know which statistical learning algorithm performs best on a prediction task, common use stacking methods ensemble individual learners into more powerful single learner. Stacking algorithms are usually based linear models, may run problems, especially when predictions highly correlated. In this study, we develop greedy for model that overcomes issue while still being very fast and easy interpret. We evaluate our 7 different data sets from...
In response to the COVID-19 pandemic, Bavarian State government announced several mitigation measures beginning on March 16, 2020, which likely led a reduction in traffic and subsequent improvement air quality. this study, we evaluated short-term effect of NO2 concentrations Munich, Germany.We applied two quasi-experimental approaches, controlled interrupted time-series (c-ITS) approach synthetic control (SC) approach. Each compared changes occurring 2020 2014-2019, accounted for...
The increasing prevalence of obesity is a major public health problem in many countries. Built environment factors are known to be associated with obesity, which an important risk factor for type 2 diabetes. Online geocoding services could used identify regions high concentration obesogenic factors. aim our study was examine the feasibility integrating information from online assessment environments.
Body mass index (BMI) is an important parameter associated with mortality and health-related quality of life (HRQoL) in chronic obstructive pulmonary disease (COPD). However, informed guidance on stratified weight recommendations for COPD still lacking. This study aims to determine the association between BMI HRQoL across different severity grades support patient management.
This study aimed to assess the impact of using different weighting procedures for German Index Multiple Deprivation (GIMD) investigating their link mortality rates.In addition original (normative) GIMD domains, four alternative approaches were applied: equal weighting, linear regression, maximization algorithm and factor analysis. Correlation analyses quantify association between differently weighted versions based on district-level official data from Germany in 2010 applied (n=412...
Abstract Objective Applications of machine learning in healthcare are high interest and have the potential to improve patient care. Yet, real-world accuracy these models clinical practice on different subpopulations remains unclear. To address important questions, we hosted a community challenge evaluate methods that predict outcomes. We focused prediction all-cause mortality as question. Materials Using Model-to-Data framework, 345 registered participants, coalescing into 25 independent...
Surgical measures to combat obesity are very effective in terms of weight loss, recovery from diabetes, and improvement cardiovascular risk factors. However, previous studies found both positive negative results regarding the effect bariatric surgery on health care utilization. Using claims data largest insurance provider Germany, we estimated causal costs a time period ranging 2 years before 3 after intervention. Owing absence control group, employed Bayesian structural forecasting model...
OBJECTIVE To assess the independent causal effect of BMI and type 2 diabetes (T2D) on socioeconomic outcomes by applying two-sample Mendelian randomization (MR) analysis. RESEARCH DESIGN AND METHODS We performed univariable multivariable MR to jointly T2D outcomes. used overlapping genome-wide significant single nucleotide polymorphisms for as instrumental variables. Their impact household income regional deprivation was assessed using summary-level data from UK Biobank. RESULTS In analysis,...
Abstract Differential privacy promises to strike a balance between the protection of for individuals and generation insights from data. But how exactly does it work? Christoph F. Kurz works through simple hypothetical example
Inpatient care is a large share of total health spending, making analysis inpatient utilization patterns an important part understanding what drives spending growth. Common features measures such as length stay and include zero inflation, overdispersion, skewness, all which complicate statistical modeling. Moreover, latent subgroups patients may have distinct relationships between that observed covariates. In this work, we apply compare likelihood‐based parametric Bayesian mixtures negative...