Sheng Wang

Sheng Wang joined Mathematica in 2011 and serves as a principal data scientist with more than a decade of experience in statistical modeling, sampling design, program integrity, Bayesian analysis, quality measure development, dashboard design, and artificial intelligence (AI) development. He focuses on designing rigorous methodologies for program evaluation, improper payment measurement, and quality indicators for federal and state programs.

Since joining Mathematica, Sheng has served as a senior technical expert on key projects. He led the development of the sampling design, modeling, and rigorous estimation methodology for the evaluation of improper payment rates for the Federal-State Unemployment Insurance program as part of the Benefit Accuracy Measurement Methodology Evaluation. As the project director for the Value Incentives and Quality Reporting Center project, he led a team supporting efforts of the Centers for Medicare & Medicaid Services (CMS) to validate hospital-reported measures based on medical charts, claims data, and electronic health records data for the Inpatient Quality Reporting program, Hospital-Acquired Condition Reduction Program, Outpatient Quality Reporting program, and electronic Clinical Quality Measures program. He led all the tasks for designing the sampling methodology, conducting data analytics, and supporting the Annual Payment Update process, ensuring accuracy and reliability in hospital performance assessments for more than 10 years. He led the recalibration of Patient Safety Indicators (PSIs) for the CMS Hospital Reporting Programs, developing rigorous statistical methodologies for feature selection, modeling, and PSI refinement using Medicare fee-for-service claims data. As a senior statistician on the AHRQ Quality Indicators project, he developed statistical models, selected features, and supported the development of public-use AHRQ Quality Indicator software. In addition, as the project director for the Healthcare Cost and Utilization Project, he oversaw analytic tasks leveraging advanced sampling design and weighting methods to assess quality indicator impacts over seven years using all-payer discharge data based on the U.S. population.

In addition to his statistical expertise, Sheng has been at the forefront of AI and modeling innovations. In the Value Incentives and Quality Reporting Center project, he led the development of AI-based tools to significantly streamline the public rulemaking review process, analyzing thousands of pages across 10 years’ worth of regulatory documents. He also led the development of retrieval-augmented generation-based AI for CMS patient safety projects, extracting key insights about composite measures, such as PSI 90, from the final Inpatient Prospective Payment System rules to support CMS’s decision making. Furthermore, he has extensive experience designing AI-driven applications, including prototypes for chatbots that enable users to interact with large SQL databases using natural language and summarize complex data insights efficiently.

Sheng holds a Ph.D. in statistics and has dedicated his career to advancing statistical methodologies and AI-driven solutions to improve program integrity, quality measurement, and healthcare analytics.