The recent downward trend in reading proficiency for U.S. students underscores the urgent need for tools that can identify and support students at risk of falling behind. This study examined the extent to which diagnostic assessments of foundational reading skills can predict subsequent changes in performance on a state-level reading assessment, the Minnesota Comprehensive Assessments (MCA). Using ReadBasix diagnostic scores across six sub-skills, we trained a multi-class classification model to predict shifts in students’ MCA proficiency levels. Our results show that diagnostic measures of foundational skills, such as morphology and reading efficiency, can provide meaningful insight into students’ future reading trajectories. In addition, there appears to be a threshold for these key reading skills that needs to be reached before significant improvements in reading efficiency can occur. These findings highlight the importance of assessing and monitoring specific skill development, rather than relying solely on broad outcome measures, to guide instructional decisions. By linking sub-skill diagnostics to state assessment outcomes, this work suggests a path forward for more targeted, data-driven interventions aimed at reversing national trends in declining reading proficiency.