Clinical practice guidelines provide a means to enhance physician performance. This investigation was undertaken in an attempt to understand the nature of impact of guideline use on physician performance. We investigated the impact of (a) algorithmic-based and (b) text-based practice guidelines on clinical decision-making by physicians at varying levels of expertise. Data were collected using clinical scenarios and a think-aloud paradigm, both with (primed) and without (spontaneous) the use of the guidelines. The two guidelines used in the study were management of diabetes and screening for thyroid disease. The results show that guidelines were used as reminders for both experts and non-experts. Guidelines acted as an educational tool for non-experts by assisting in knowledge reorganization, particularly for the non-experts. Text and algorithmic guideline formats were both useful to physician performance depending on the purpose of use: solving clinical problems or learning. These results provide insights into how guidelines can be fine-tuned for different users and for different purposes. Empirical research, coupled with design principles from the cognitive sciences, can form an essential component of guideline design and development.