Heart Murmur Clinical Trial
Official title:
Self-Directed Study Using MP3 Players Versus Multimedia Lecture to Improve Auscultation Proficiency of Physicians: A Randomized, Controlled Trial.
Our objective was to test two educational interventions for teaching cardiac auscultation: self-directed learning using portable audio files versus a single, massed multimedia lecture intervention similar to current best practice, to determine which would most effectively increase recognition of common cardiac sounds by physicians. The investigators hypothesized that doctors learning on their own time could better improve their skills.
Participants All non-cardiologist physicians were eligible for the study and responded to recruitment advertising. Participants were 83 non-cardiologist physician volunteers (57 practicing physicians, 26 residents in training) from one Academic Independent Medical Center in the Northeastern United States enrolled between January 11, 2011 and February 25, 2011. Participants were randomized by study investigators using a table of random numbers. All participants were given an MP3 player at the time of enrollment for compensation of their time, but received no grade for participation. Participants provided their informed consent before enrolling in the study. Educational Interventions Following the pretest, participants in the interventional arm of the study were given an MP3 player which contained eight, 15-second heart sound files (7 pathologic heart sound files and 1 normal heart sound file) alternated with nine leisure songs. MP3 players could play files in order or "shuffle" files in random order. Each heart sound file was introduced by the narrated phrase "Identify this sound,", followed by 10 beats of a heart sound recording, then the narrated identification of the heart sound. Participants were asked to use the MP3 player in order to improve their auscultation skills as often as they could and asked to record their activities while listening, but were given no further training. Following the pretest, participants in the control arm received a 1-hour multimedia lecture (Power Point) taught by the same faculty member (AGK), which reviewed the pathophysiology, exacerbating and relieving factors, as well as visual diagrams of the heart sounds. The normal and seven pathologic heart sounds were played for participants during the lecture for a total of 2 minutes through high-quality home stereo speakers, exposing learners to 1280 beats in the 1-hour session. Abnormal sounds were taught in groups of systolic, diastolic and extra sounds, and were first introduced by name, and then as unknowns using mixed practice in each of the three sections. At the end of the session, control group participants were given an MP3 player with no heart sound files as honoraria. Both the intervention and control group were told they would be retested at 12 weeks. Both were given a paper copy of the Power Point slides for self-study, but had no electronic access to the original slides. Both groups were asked not to discuss the study with the other participants. Content of the audio files The audio files of the heart sounds were obtained from actual patients and have validity evidence generated in prior studies. The pathologic sounds included aortic stenosis, aortic insufficiency, mitral stenosis, mitral insufficiency, mitral valve prolapse, S3, S4, and the normal heart sound. These heart sounds collectively represent the most important auscultatory findings according to a previously published national survey of residency program directors. Measurement All participants took a 15-question pretest in addition to a posttest at 12 weeks using identical heart sound files to those used in training. Tests consisted of 8-second clips of each heart sound file, with 10 beats of each sound. Participants were given 15 seconds after each question to choose the correct response from an extended match list that included 9 choices, one of which was not represented among the training materials ("benign flow murmur"). Each of the 7 pathologic heart sounds were represented twice, with normal heart sounds represented once. The order of questions for the test was randomly selected. The same speakers used in the lecture format were also employed utilized for testing. Study examiners tested the speaker sound quality before each test. All tests were independently hand-graded by two graders, and all discrepancies in handwriting and scoring were resolved by consensus of both investigators. Grades were not shared with participants. Both groups were asked to rate their prior experience with pathologic heart sounds, including observation and feedback they have received, satisfaction with the teaching modality and confidence in their auscultatory skill at the pretest and posttesting. Both groups audited their preference for learning style by the VARK© learning style assessment instrument Version 7.0, at the time of enrollment. Both groups were asked to record any other methods used to study auscultation during the study period. Intervention group exposure to heart sounds was determined by electronic interrogation of their MP3 player at 12 weeks. Play counts recorded by the device were reviewed by both study investigators and entered on an electronic spreadsheet. Data analysis Demographic comparisons, Likert scales of satisfaction and pathologic heart sound exposure between control and intervention groups were performed via Chi-square testing and Student's t-test as appropriate. Internal consistency of the pretest was performed using Cronbach's alpha. Total study time was recorded as 60 minutes for the control group, and was calculated by the total number of recorded plays multiplied by 15 seconds per play for the intervention group. Additional self-reported study methods and amount of exposure were recorded but not added to control and intervention study time calculations. The primary outcome was the difference in test score from pretest to posttest and was compared using paired t-tests. Effect sizes for both arms were computed using differences in means divided by pooled standard deviations. Correlations between exposure and improvement from pretest to posttest in the intervention group were calculated by linear regression. Subgroup analysis of intervention group participants who did not use their MP3 player between weeks 4 and 12 (defined as play count= 0 during time interval) was performed to approximate decay in that group, and was analyzed by paired t-test. The auditory learning preference of the VARK learning style inventory was characterized dichotomously as present or absent. A two-factor ANOVA was performed comparing intervention and control groups with presence of auditory learning style against the primary outcome as the dependent variable to determine the relative effects of learning style and instructional method. Power calculations were made for the observed differences between the control and intervention arms for both variables of the primary objectives. ;