Knowledge workers increasingly operate under conditions of persistent cognitive strain driven not by task volume alone, but by the accumulation of unresolved decisions, unprocessed inputs, and competing commitments. Existing measures of cognitive load and stress rely primarily on self-report instruments or laboratory-based methods, limiting their ability to capture continuous, real-world cognitive burden. This paper introduces the Decision Load Index (DLI), a conceptual framework for quantifying cognitive burden using system-level metadata generated by everyday productivity tools. Rather than assessing subjective experience directly, the DLI operationalizes cognitive load through observable indicators such as open loops, unprocessed inputs, overdue commitments, and active project scope. We situate the DLI within established cognitive psychology literature, including research on working memory, decision fatigue, and the Zeigarnik effect, and distinguish it from task-count-based productivity metrics. The framework is designed to enable continuous, ecological, and privacy-preserving measurement without additional survey burden or instrumentation. We outline a research agenda for empirical validation, including psychometric assessment, normative benchmarking, and intervention testing. By reframing cognitive load as an emergent property of system state rather than momentary task demand, the Decision Load Index offers a new foundation for studying and managing cognitive burden in knowledge work.