How and why internal medicine clerkship directors use locally developed, faculty-written examinations: Results of a national survey

PURPOSE: To describe how and why internal medicine clerkship directors (CDs) use locally developed, faculty-written (LFW) examinations. METHOD: In 2009, the Clerkship Directors in Internal Medicine conducted an annual, online, confidential survey of its 107 U.S. and Canadian institutional members, including questions about LFW examinations. Data were analyzed using descriptive statistics and coding of free text. RESULTS: Sixty-nine of 107 members (64.5%) responded. The National Board of Medical Examiners (NBME) examination was administered by 93% (63/68), LFW examinations were used by 33% (22/67), and both types were used by 22% (17/67)-compared with 85%, 36%, and 12% in 2005. LFW examinations were frequently created by the CD alone (9/22; 41%) and consisted of one test (12/22; 52.2%), but some schools gave two (6/22; 26.1%), three (2/22; 8.6%), or four or more (3/22; 13%). Multiple-choice examinations were most common (26/38; 68.4%), followed by short-answer (8/38; 21.1%) and essay (4/38; 10.5%). Most were graded using preestablished criteria; half required a minimum passing score (60% most common). LFW exams were most commonly 5% to 10% of the total grade. Only a minority of CDs reported having reliability estimates or a control group for their exams. Most (70%) reported using LFW exams to cover content felt to be underrepresented by the NBME. CONCLUSIONS: Findings strongly suggest that a minority of internal medicine CDs use LFW examinations, mostly to measure achievement not assessed by the NBME. However, validity evidence is not consistently being gathered, which may limit judgments based on exam results.