Traditionally, medical students have been taught basic science and introductory clinical material in the traditional lecture format. It has long been assumed that such lectures impart the greatest amount of information in the shortest period of time, forming the framework for future clinical learning. However, it is now appreciated that little of what is taught in a purely lecture format is actually retained by the students. It is estimated that students’ attention diminishes after only 10 minutes of passive listening, limiting what is learned.1 Additionally, lectures do little to challenge and stimulate the students to problem solve,2 a skill that is necessary for their clinical years as well as their medical careers.
Small-group discussions offer the opportunity for students to be more active participants, making them partners in the educational process. The theoretical benefits of small-group discussions include independent thinking, problem solving, interaction with peers and instructors, and reinforcement of knowledge through verbal recitation. It may further encourage the students to extend learning and problem solving beyond the core material presented in the classroom.
Previous articles have compared problem-based small-group learning with traditional lecture formats for medical students and found a greater satisfaction level with the former.2–6 However, group assignment was usually on a volunteer basis; hence student evaluations may have been biased by the self-selection process. Additionally, most studies applied problem-based learning during the preclinical years and allowed students independent study time outside of the classroom to research their answers. We were interested in prospectively comparing the value of small-group discussion with traditional lecture format by a single instructor to a group of third-year medical students in the obstetrics and gynecology clerkship, as determined by student evaluation and multiple-choice examination as measures of satisfaction and efficacy, respectively. Students were to be blinded to the nature of the study, and were not allowed additional time for nonclassroom study for those in the discussion group. We hypothesized that small-group discussion would be associated with greater student satisfaction as well as improved test scores immediately after the instructional sessions.
MATERIALS AND METHODS
Third-year medical students at the University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School at Camden rotate through the obstetrics and gynecology clerkship in 8-week blocks, usually in groups of 6 to 9. The first week of the rotation is “core week,” consisting of 8 hours of didactics per day for 5 consecutive days, typically taught in a lecture format. The division of maternal–fetal medicine teaches 1 day of core week, which includes 1 lecture on diabetes in pregnancy (gestational and pregestational) and another on hypertension in pregnancy (gestational and chronic). Over a 12-block period (February 2001 through February 2003), students were taught each of these 2 subjects by a single instructor (R.L.F.) in 1 of 2 formats, either by traditional lecture or by small-group discussion, with a maximum time of 1 hour allotted for each topic. The former consisted of a fact-filled outline that was reviewed in a traditional 1-way lecture, supported by slides containing more detailed information. For the small-group discussions, fact-based and problem-based questions were assigned to student pairs for independent review for 5 minutes, followed by discussions led by the students and facilitated by the instructor. Principles used by the instructor in small-group discussions included memorization of the students’ names in advance to promote familiarity, facilitation rather than lecturing, group discussions of clinical scenarios, and problem solving. To allow the students to have an informational basis for the group discussions,7 they were asked to review the topic outlines the evening before the discussion. Because this would have provided the discussion group students with an additional educational opportunity, students in the lecture group were also asked to review the outline the night before their lecture.
To avoid introducing bias in test scores based on differing educational levels and to ensure comparable numbers in each group, format assignment for the 12 blocks were nonrandomly divided into 6 alternating couplets (eg, lecture–small-group, then small-group–lecture). Students were blinded to the nature of the study.
At the conclusion of each educational session, the students were asked to anonymously complete a 20-question multiple-choice examination to assess their immediate postinstructional knowledge of the topics. The test questions were written by 1 of the investigators (S.L.J.), who was blinded to the outline material and who had prior experience writing multiple-choice questions.8 The examination consisted of 10 questions on diabetes and 10 on hypertension in pregnancy, with an equal proportion of fact-based and problem-based questions within each topic. The questions were reviewed and approved by 2 members of the full-time faculty not associated with the study, although formal psychometric analysis (ie, validity and reliability testing) of the individual questions was not assessed. The instructor was blinded to the test questions to prevent bias in the material taught in the 2 educational formats. After the examination, the students were asked to anonymously complete an evaluation form of the instructional format. Responses were marked on a 5-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree). At the end of each 8-week rotation on obstetrics and gynecology, the students were given the same 20-question multiple-choice examination to assess retention of the material. The results of the National Board of Medical Examiners “Shelf” examination, oral examination, clinical grade, and final clerkship grade were also compared. Demographic information on the students was obtained from the Dean's office after completion of the study. The chairman of our institutional review committee reviewed the study protocol and granted an exemption from formal committee review, because this purely educational project did not involve patient care.
Statistical tests included the Student t test for parametric data (examination scores), Mann–Whitney U test for ordinal data (Likert scores), χ2 test for proportions, and logistic regression to control for potentially confounding variables, all performed on SPSS 10 (SPSS Inc, Chicago, IL). Statistical significance was defined as P < .05.
A total of 91 students attended the didactic core week of the obstetrics and gynecology clerkship from February 2001 to February 2003. No protocol violations occurred, although 1 student in the discussion group did not complete the rotation. Forty-six students were taught in the lecture format, while 45 participated in the small-group discussions. As shown in Table 1, the lecture group had a slightly higher proportion of female students, although there were no significant differences with respect to educational experience or number of students per rotation.
As shown in Table 2, on a Likert scale of 1 to 5, the 2 groups were identical with regards to both pre- and postsession self-assessed knowledge of diabetes and hypertension. Although the students in the lecture group were more likely to have read the outline before the educational session, this did not reach statistical significance. However, there was a significantly higher enjoyment (median value 5 versus 4, P < .001) and sense of educational stimulation (median value 5 versus 4, P < .001) in the discussion group. Additionally, students inthe discussion group were more satisfied with their assigned educational format, as they were less desirous of the alternate instructional format than those in the lecture group (median value 2 versus 3, P < .001).
Students’ comments on the best and worst features of the instructional formats revealed generally favorable reviews for both. Positive comments about the lecture format included “High-yield material” and “straightforward,” while negative comments included “Rapidity of material covered,” “Boring, needs student participation,” and “I have a headache.” Students from the discussion group commented that they enjoyed the “interactive style,” which helped to “solidify and integrate knowledge.” It was also noted that everyone was “awake and interested.” Negative comments included the lack of time to exhaustively cover all the important topics, “embarrassment of ignorance,” and inadequate time to “review material independently before applying it to clinical scenarios.”
The results of test scores and clerkship performance grades are shown in Table 3. There were no significant differences in multiple choice test scores between the 2 instructional formats, either immediately after the educational session or at the end of the rotation. While there was overall improvement in test scores between the first and second examinations, there was no difference in the increase based on group assignment. Additionally, there were no significant differences in the students’ performances in the National Board of Medical Examiners “Shelf” examination, oral examination, clinical grade, or overall clerkship grade. After controlling for student gender, there were no changes in the comparisons of student evaluations or test scores between the 2 formats.
Our investigation demonstrated that small-group discussion was significantly favored over the traditional lecture format. However, immediate and delayed tests of material retention failed to demonstrate any significant difference based on instructional format. Previous studies have attempted to compare the value of traditional lecture with small-group discussion formats in medical education, usually using problem-based learning curricula. While student satisfaction was usually higher in the problem-based small groups,2–6,9,10 test performance was more variable. Some studies showed no difference in fact-based tests scores such as multiple-choice examinations,5,6 while others showed that students in problem-based preclinical curricula tended to score lower in the National Board of Medical Examiners Part I examination.3,11 In contrast, some studies showed improved clinical evaluations of students who had problem-based preclinical curricula, as well as trends toward higher test scores in short answer/essay questions and in the National Board of Medical Examiners Part II examination.3,4,6,9,11–13 Additionally, although students in the discussion groups generally gave positive reviews, many commented on inconsistent or inadequate amount of educational material covered.10 Discussion groups were also favored by most of the faculty, although it took additional time to teach the course material.4,10
In the majority of previous studies in which curriculum-wide changes were implemented during the preclinical years, students in the problem-based learning groups were allotted nonclassroom independent time for study and preparation of the clinical scenarios. Our study, however, sought to determine if a discussion-based format could successfully be applied by an individual educator in an otherwise lecture-based week of didactics during the student clerkship in obstetrics and gynecology, without the benefit of independent study time. Additionally, ours was not a purely problem-based approach; many of the assigned questions dealt with fact-based material. However, the students led the interactive discussions, with the instructor acting merely as a facilitator to redirect questions back to the group. Although the overall impact of discussion groups may be larger in a curriculum-wide approach, it has been observed that smaller studies such as ours can provide a well-controlled environment in which to evaluate various educational formats.3,4
The finding of no significant difference in test performance between the two groups suggests that there is no objectively measurable educational advantage to small-group discussion. Alternatively, our test score findings may be related to the type of examination the students were given, which consisted of an equal proportion of fact-based and problem-based multiple choice questions. This type of examination may simply be a poor instrument for assessing students’ knowledge, independent thinking, or problem-solving ability. Perhaps small-group advantages would be uncovered if the test questions were of a different variety, such as short-essay type.4,6
From an instructor's perspective, the small-group discussions were more challenging yet rewarding. Unlike a traditional 1-way lecture in which each lesson is virtually identical (and eventually monotonous), each small-group discussion is unique, with differing personalities of the individual students as well as the interactions of the group as a whole. Although the time of each instructional session was not recorded, the instructor observed that the group discussions took more time to complete than the lectures. It was much more difficult covering all of the educational material in the allotted time period, as the group often veered off the instructor's projected course. This should not necessarily be seen as a disadvantage, as allowing the students to pursue an area of interest in greater depth may help to spark interest in the topic and/or the educational process. Similarly, the time commitment to prepare for the educational sessions was greater in the discussion group. An outline, which incorporated the factual information, was created and utilized for both formats. For the lecture, the instructor had only to prepare a slide presentation based on this outline. In contrast, the initial preparation for the discussion group required careful consideration of a limited number of open- and closed-ended questions that would encompass the outline material. Once the questions were established, there was the additional time to memorize the students’ names from pictures provided by the clerkship director, as well as presession review of the principals of facilitating student-led group discussions to prevent reverting to the natural tendency to lecture.
The strengths of this study were multiple. First, unlike previous studies, instructional group assignment was not self-selected. It has been shown that students who select problem-based curricula may be brighter based on higher college grades and Medical College Admission Test scores.13 In fact, our students were unaware of the nature of this comparative study, thereby limiting bias based on preconceived preferences of educational format. Second, the use of educational materials (ie, outlines) was equal in the 2 groups, allowing us to directly compare the value of the instructional formats alone. Third, the instructor and examination writer were both blinded to the other's material, eliminating bias of the information taught and the questions on the multiple-choice test. Finally, our sample size was adequate; a post hoc analysis revealed that, with an α level of .05, our study had a 90% power to detect a 15% difference in postinstructional test scores.
Despite the blinded study design, there were a few limitations to this investigation. The sole instructor was obviously unable to be blinded to format assignment. Therefore, if he had a bias toward one of the pedagogical techniques, it might have unconsciously affected his instructional performance. Additionally, the same multiple-choice examination was given throughout the entire 2-year study to the students in this small medical school. As the students were not aware of the nature of this study, they may have discussed test questions with classmates who subsequently rotated through obstetrics and gynecology. However, test scores were not observed to have a higher trend over the course of the study.
In summary, third-year medical students learning about hypertension and diabetes in pregnancy during their core week in obstetrics and gynecology clearly preferred the small-group discussion over the traditional lecture format. However, this preference did not translate into improved test scores. The small-group discussion format, while more labor intensive for the educator, helped minimize instructor boredom with frequently repeated material. However, for the busy practitioners who are responsible for medical student clerkship education, they should not feel guilty that student learning is being compromised by continuing to offer traditional lectures. Further studies are needed to elucidate the educational value of these different instructional formats, including additional means of assessing student knowledge and longer-term educational stimulation.
1. Davis BG. Tools for teaching. San Francisco (CA): Jossey-Bass Publishers; 1993.
2. Dunnington G, Witzke D, Rubeck R, Beck A, Mohr J, Putnam C. A comparison of the teaching effectiveness of the didactic lecture and the problem-oriented small group session: a prospective study. Surgery 1987;102:291–6.
3. Vernon DTA, Blake RL. Does problem-based learning work? A meta-analysis of evaluative research. Acad Med 1993;68:550–63.
4. Albanese MA, Mitchell S. Problem-based learning: a review of literature on its outcomes and implementation issues. Acad Med 1993;68:52–81.
5. Wendelberger KJ, Simpson DE, Biernat KA. Problem-based learning in a third-year pediatric clerkship. Teach Learn Med 1996;8:28–32.
6. Antepohl W, Herzig S. Problem-based learning versus lecture-based learning in a course of basic pharmacology: a controlled, randomized study. Med Educ 1999;33:106–13.
7. Stein M, Neill P, Houston S. Case discussion in clinical pharmacology: application of small group teaching methods to a large group. Med Teach 1990;12:193–6.
8. Jacobs SL, Goldberg JS. Obstetrics and gynecology. In: Goldberg JS, editor. Appleton & Lange's outline review for the USMLE step 2. 3rd ed. Stamford (CT): Appleton & Lange, 1999. p. 439–79.
9. Curtis JA, Indyk D, Taylor B. Successful use of problem-based learning in a third-year pediatric clerkship. Ambul Pediatr 2001;1:132–5.
10. Brinton DA, Jarvis JQ, Harris DL. A small-group instructional experiment in medical education. J Med Educ 1984;59:13–8.
11. Kaufman A, Mennin S, Waterman R, Duban S, Hansbarger C, Silverblatt H, et al. The New Mexico Experiment: educational innovation and institutional change. Acad Med 1989;64:285–94.
12. Richards BF, Ober P, Cariaga-Lo L, Camp MG, Philp J, McFarlane M, Rupp R, Zaccaro DJ. Ratings of students’ performances in a third-year internal medicine clerkship: a comparison between problem-based and lecture-based curricula. Acad Med 1996;71:187–9.
13. Colliver JA. Effectiveness of problem-based learning curricula: research and theory. Acad Med 2000;75:259–66.