Machine learning models have achieved strong predictive accuracy for 90-day outcome prediction in acute ischaemic stroke, but clinical adoption is limited by the misalignment of model explanations with clinicians' reasoning. A clinician user study called for clinical guideline-aligned cut-offs, prompting the question of whether continuous predictors can be replaced by clinically informed categorical encodings without sacrificing performance. On a multi-centre European registry, we compared standard and fully categorised gradient-boosted models, the latter using stroke guideline-aligned, treatment-specific thresholds. The fully categorised models were statistically indistinguishable from their continuous counterparts in two of the treatment cohorts, with a significant drop in predictive accuracy in one cohort. Global feature importance rankings remained consistent, suggesting that discretising continuous predictors into guideline-based categories preserves the core hierarchy of prognostic factors across all treatment groups. Thus, guideline-based categorisation is a viable design choice for stroke-outcome models. Blogger's Review: This study explores the potential of aligning machine learning models with clinical guidelines, providing important insights into improving model interpretability and adoption in clinical practice.