Longitudinal knowledge assessments (LKAs) in United States medical certification have largely replaced the required decennial examinations in response to widespread physician concerns. Yet a decade later, the evidence base supporting LKAs for summative assessment remains comparatively limited. This session examines four measurement and policy questions in current LKAs: the relationship between current and historical passing standards, progress-report weighting that discounts early-cycle performance, proficiency estimation using cross-sectional IRT models, and security strategies that may not fully deter misconduct. We conclude with a reform agenda that emphasizes independent validity replication, transparent standard-setting, longitudinal growth-model scoring, explicit validity arguments, and clearer communication about security and score interpretation for boards, vendors, the physicians they certify, and the public.