Do coursework summative assessments predict clinical performance? A systematic review

Rebecca Terry, Wayne Hing, Robin Orr, Nikki Milne

Two goals of summative assessment in health profession education programs are to ensure the robustness of high stakes decisions such as progression and licensing, and predict future performance. This systematic and critical review aims to investigate the ability of specific modes of summative assessment to predict the clinical performance of health profession education students. PubMed, CINAHL, SPORTDiscus, ERIC and EMBASE databases were searched using key terms with articles collected subjected to dedicated inclusion criteria. Rigorous exclusion criteria were applied to ensure a consistent interpretation of 'summative assessment' and 'clinical performance'. Data were extracted using a pre-determined format and papers were critically appraised by two independent reviewers using a modified Downs and Black checklist with level of agreement between reviewers determined through a Kappa analysis. Of the 4783 studies retrieved from the search strategy, 18 studies were included in the final review. Twelve were from the medical profession and there was one from each of physiotherapy, pharmacy, dietetics, speech pathology, dentistry and dental hygiene. Objective Structured Clinical Examinations featured in 15 papers, written assessments in four and problem based learning evaluations, case based learning evaluations and student portfolios each featured in one paper. Sixteen different measures of clinical performance were used. Two papers were identified as 'poor' quality and the remainder categorised as 'fair' with an almost perfect (k = 0.852) level of agreement between raters. Objective Structured Clinical Examination scores accounted for 1.4-39.7% of the variance in student performance; multiple choice/extended matching questions and short answer written examinations accounted for 3.2-29.2%; problem based or case based learning evaluations accounted for 4.4-16.6%; and student portfolios accounted for 12.1%. Objective structured clinical examinations and written examinations consisting of multiple choice/extended matching questions and short answer questions do have significant relationships with the clinical performance of health professional students. However, caution should be applied if using these assessments as predictive measures for clinical performance due to a small body of evidence and large variations in the predictive strength of the relationships identified. Based on the current evidence, the Objective Structured Clinical Examination may be the most appropriate summative assessment for educators to use to identify students that may be at risk of poor performance in a clinical workplace environment. Further research on this topic is needed to improve the strength of the predictive relationship.