Of concern in the problem-based learning process in medical student education is the effect of allowing students to generate their own learning issues while discussing cases and problems. This study considered learning issues in problem-based learning, studying faculty and student overlap and consistency across learning groups while adding the perspective of analyzing learning issues developed by students without the assistance of a tutor. Lists of students' learning issues, interviews with the course director and course writer, and final examination scores were collected for 4 of 12 tutorial groups, representing 52 students, and 4 cases prepared by students. The mean case overlap in issues and content covered was somewhat lower for this study than for previous studies using staff tutors. Not having a tutor present during the initial case discussion may slightly reduce the amount of case overlap between faculty-intended learning objectives and the learning issues generated by students for the purposes of self-study. Only one-quarter of the 77 faculty objectives were identified by all 4 groups as learning issues. Results suggest that reliance on a tutorless format may not be appropriate when other sources of structure are absent from the curriculum. (Contains three tables and seven references.) (SLD)