The aim of this paper is to report on a three-year follow-up evaluation of an experimental problem-based learning (PBL) integrated curriculum directed to second-year engineering students. The PBL curriculum brought together the contents of physics, mathematics and computer science courses into a single course. Instead of the students having to enroll in each course separately, they had the opportunity to enroll in a single course in which they learned the contents of those subjects by solving real-life engineering problems. In order to evaluate the effects of the program on the students' academic achievement, three data sources were taken: pre-test--post-test scores on two physics tests: students' grade point average (GPA); and students' grades in advanced engineering courses. With respect to the first source of data, results showed that while PBL students' improvements in scores were significantly higher than control students' improvements on the Mechanics Baseline Test, improvements were similar in both groups on the Force Concept Inventory. With regard to the second source of data, PBL students' GPA was significantly higher than control students' GPA. Finally, when PBL and control students' performance on advanced courses were compared, PBL students graded higher than control students on the five courses assessed. However, on only two of them ('Probability and Statistics' and 'Oral Communication') were the differences significant.