OBJECTIVE: Synthesise evidence about the impact of family medicine/general practice (FM) clerkships on undergraduate medical students, teaching general/family practitioners (FPs) and/or their patients. DATA SOURCES: Medline, ERIC, PsycINFO, EMBASE and Web of Knowledge searched from 21 November to 17 December 2013. Primary, empirical, quantitative or qualitative studies, since 1990, with abstracts included. No country restrictions. Full text languages: English, French, Spanish, German, Dutch or Italian. REVIEW METHODS: Independent selection and data extraction by two authors using predefined data extraction fields, including Kirkpatrick's levels for educational intervention outcomes, study quality indicators and Best Evidence Medical Education (BEME) strength of findings' grades. Descriptive narrative synthesis applied. RESULTS: Sixty-four included articles: impact on students (48), teaching FPs (12) and patients (8). Sample sizes: 16-1095 students, 3-146 FPs and 94-2550 patients. Twenty-six studies evaluated at Kirkpatrick level 1, 26 at level 2 and 6 at level 3. Only one study achieved BEME's grade 5. The majority was assessed as grade 4 (27) and 3 (33). Students reported satisfaction with content and process of teaching as well as learning in FM clerkships. They enhanced previous learning, and provided unique learning on dealing with common acute and chronic conditions, health maintenance, disease prevention, communication and problem-solving skills. Students' attitudes towards FM were improved, but new or enhanced interest in FM careers did not persist without change after graduation. Teaching FPs reported increased job satisfaction and stimulation for professional development, but also increased workload and less productivity, depending on the setting. Overall, student's presence and participation did not have a negative impact on patients. CONCLUSIONS: Research quality on the impact of FM clerkships is still limited, yet across different settings and countries, positive impact is reported on students, FPs and patients. Future studies should involve different stakeholders, medical schools and countries, and use standardised and validated evaluation tools.