Abstract Session
8:00 AM
Saturday, February 19, 2000
840 Five Year Documentation of the Association of Excellent Clinical Teaching with Better Student Examination Performance
C.H. Griffith, J.C. Georgesen, J.F. Wilson, University of Kentucky, College of Medicine, Lexington, KY
Background: Despite our cherished beliefs, little quantitative evidence exists that students learn more from better teachers. Preliminary research documenting an association of better teaching with more learning has been limited by small sample of students, teachers, and few teaching evaluations, making measurement of excellent teaching imprecise. We hypothesized that third-year medical students working in their medicine clerkship with our “best” teacher attending physicians and residents would perform better on clerkship examinations and on USMLE 2, and we proposed to document this with a larger sample of students and residents over many more academic years than in prior reports.
Methods: We collected data from 5 academic years 1993-1998 regarding 380 third-year medical students and their randomly assigned internal medicine attending physicians and supervising internal medicine residents. Outcomes of interest were student scores on post-clerkship NBME Subject Examination in Medicine, and the long-term outcome of USMLE 2. Attendings and residents were confidentially evaluated by students at the end of each month's rotation (prior to students assigned a clinical grade), on 16 and 9 item (respectively) 5 point Likert-scale evaluation forms (coefficient α both >0.90). One teaching rating for an instructor was the mean score across all 16 or 9 items of the form, and their overall teaching rating was the mean evaluation of all the students they precepted in the study period. “Best” teachers were defined as those in the top 20% of all student evaluations in the 5 year period (amounted to a score of >4.6 of 5). Instructors with less than 5 overall evaluations were excluded. “Worst” teachers were those in the bottom 20% of evaluations.
Results: Using multiple regression approaches, controlling for prior student academic achievement with USMLE I Score, independent predictors of better post-clerkship NBME Score were exposure to a “best” attending (adjusted mean score 477 vs. 455, [about 1/5 of a S.D], p=0.03) and “best” resident (adjusted mean of 478 vs. 454, [about 1/4 of a S.D.), p=0.01). Exposure to a “best” attending was associated with better USMLE 2 Scores (adjusted mean 204 vs. 198, [about 1/3 of a S.D.). p=0.02).
Conclusion: Excellent clinical teaching is associated with better student learning outcomes, both short-term (post-clerkship exams) and long-term (USMLE 2). As academic medical centers are pres-sured to justify the costs of the teaching mission, further studies of the outcomes of teaching are needed.
841 Smoking Status and Depressive Symptoms in the Kentucky Health Survey
P.F. BASS, J.T. Wilson, & C.H. Griffith, General Internal Medicine, University of Kentucky, Lexington, KY
Background Cigarette smoking is associated with lower education and socioeconomic status and smokers are more likely to have a history of major depression than are non-smokers. The purpose of this study was to examine differences in health related quality of life among smokers, former smokers, and non-smokers.
Methods The data are drawn from the 1997 Kentucky Health Survey administered by the University of Kentucky Survey Research Center. The survey included 677 Kentuckians who were 18 years of age or older. Participants were contacted during October 1997 by telephone using a random digit dialing method that gave every residential telephone line in Kentucky an equal probability of being called. The response rate was 56% of all eligible respondents. The margin of error for a sample of this size is less than ± 4% at the 95% confidence level. 92% of those surveyed were white and 5% were African-American; 56% of respondents were female; 45% had some education beyond high school and 18% did not complete high school. By self-report, 30% resided in a rural community and the median income was $30,000 to 40,0000. A mental health score as well as role functioning, social functioning, physical functioning, pain perception, and health outlook were calculated using the MOS-20 format and compared to self reported smoking status. The effects of self-reported smoking status on these MOS-20 scales were examined using multiple regression analysis controlling for age, sex, education, and rural status.
Results 29.8% were current smokers; 24.1% were former smokers; and 45.8% were non-smokers. In comparison to those who never smoked, current smokers were younger, less well educated, and predominantly urban dwellers. Current smokers had worse mental health scores (71.7 vs. 75.8 vs. 76.9 p=.01) when compared to former and non-smokers. Current smokers also had significantly poorer health outlook (62.9 vs. 65.1 vs. 69.1 p=.03) compared to former and non-smokers. No significant interactions were found when comparing smoking status to gender or rural living. There was no significant relationship between smoking status and the subscales for role function or physical function, but there was a non-significant trend for current smokers to have a greater pain perception and poorer social function than former or non-smokers.
Conclusion Current smokers report more depressive symptoms and poorer health outlook than non-smokers. Former smokers are more similar those who never smoked in health related they are to current smokers.
842 Literacy Skills, Numeracy Skills, and Cognitive Impairment among Patients Taking Warfarin
C.A. Estrada, B.T. Peek, M.M. Hryniewicz, C. Collins, V.B. Higgs, and J.C. Byrd, Section of General Internal Medicine, East Carolina University School of Medicine, Clinical Information and Support Office, University Health Systems, Greenville, NC, and Asheville Department of Veteran Affairs Medical Center, Asheville, NC.
BACKGROUND AND OBJECTIVES. Elderly patients with atrial fibrillation, prior stroke, and comorbidities may have difficulty taking warfarin. Our objectives were to determine literacy and numeracy skills, and prevalence of cognitive impairment among patients taking warfarin.
METHODS. Cross-sectional study of 143 consecutive patients attending two anticoagulation management units. Literacy is the ability to use printed material to function in society, while numeracy is the ability to use basic probability and mathematical concepts. Literacy was measured with the Rapid Estimate of Adult Literacy in Medicine (REALM). Numeracy was measured with a modified Schwartz scale (maximum score 6). We measured cognitive impairment with the Folstein Mini Mental Status Exam (MMSE).
RESULTS. The mean patient age was 65 years (SD 9.9), 62% were male, and 29% were non-white. Patients' measured literacy grade level was lower than reported grade completed (kappa 0.01), see table.
Numeracy questions were answered correctly by 17.5% (5-6 questions), 34.3% (3-4 questions), and 35.7% (1-2 questions) of patients. 12.6% of patients could not answer any of the questions. Cognitive impairment (MMSE ≤ 23 points) was detected in 20% of patients (95% CI 14% to 27%).
CONCLUSIONS. Low literacy, low numeracy, and cognitive impairment are prevalent among elderly patients taking warfarin.
843 Does Access to Medical Services Decrease Avoidable Admissions to Hospitals?
S.K. Sigworth, J.B. Perlin, W.R. Smith. Division of Quality Health Care, Department of Internal Medicine, Box 980306, MCV Hospitals, Virginia Commonwealth University, Richmond, VA
Background: Avoidable admissions, defined as ambulatory conditions treated in an inpatient setting, crowd emergency rooms and hospitals, utilizing healthcare resources inefficiently. Prior studies indicate that access to a primary care provider (PCP) reduces avoidable admissions. This study was designed to determine if patient identification of a PCP (a proxy for access) reduces avoidable admission rates.
Methods: A retrospective, cross sectional analysis of medical records, during the 4th quarter of 1998, was performed on all patients whose length-of-stay (LOS) was less than 48 hours admitted to the Internal Medicine (IM) service of a large university medical center. Exclusionary criteria included admissions to the Coronary Care Unit, admissions under clinical protocols, and elective admissions for various procedures. All admissions were classified as avoidable or unavoidable based on published criteria for appropriate use of acute inpatient services (InterQual®, 1997; Milliman & Robertson, 1998).
Principal Findings: 2427 patients were admitted to the Medicine service over the study period with 518 (21.3%) discharged within 48 hours. Ultimately, 279 admissions were included in this study. Among these, 215 (77.1%) identified a PCP and 64 (22.9%) reported no, or unknown, PCP. As well, 204 (73.1%) were deemed unavoidable and 75 (26.9%) were avoidable. Chi-square analysis revealed no statistical significance (p=.564) between the two groups. There were no statistically significant differences between payor status and avoidability (p=.433) between Indigent (30.9% avoidable), Medicare (26.2%), Medicaid (28.4%) and Commercially Insured (16.7%), however, the lower percent of avoidable admissions among commercially covered patients may be clinically important.
Conclusions: Our study revealed that having a PCP did not reduce avoidable admissions, therefore other factors concerning the admission process, such as social needs, remain dominant. This contradicts the belief that access to outpatient care provides alternatives to emergency rooms and inpatient services for ambulatory sensitive conditions. Further research into these factors is needed. Despite absence of statistical significance in payor status, commercially insured had 10% fewer avoidable admissions. This study of a large, urban academic center involves patients with significant socioeconomic disadvantage. A similar study of community hospitals may yield quite different results.
844 Readability of Anticoagulant Patient Information Material
C.A. Estrada, C. Collins, M. Hryniewicz, V.B. Higgs, J.C. Byrd, General Internal Medicine, East Carolina University School of Medicine, Greenville, NC.
BACKGROUND AND OBJECTIVES: The use of warfarin requires intensive monitoring and frequent dose adjustment. Patients generally receive brochures and other types of written material to explain the risks and benefits of anticoagulants. Older patients with atrial fibrillation, prior stroke, or other comorbidities have lower literacy skills and may have difficulty reading such brochures. In general, brochures written at the sixth grade level or below should be used for patient education. Our objective was to determine the readability of patient education materials offered to patients receiving anticoagulants.
METHODS: We measured readability with the SMOG readability formula. We obtained brochures from manufacturers of warfarin, patient advocacy groups, and the United States Pharmacopeia. We excluded brochures written in languages other than English.
RESULTS: We examined 50 brochures: 14% relating to atrial fibrillation, 28% concerning the use of warfarin, and 58% regarding other topics related to anticoagulation (valvular disease, stroke, etc.). The test-retest reliability was within one grade level 77.3% of the time. All brochures were written at the 7th grade education level or above:
The readability grade level was similar for short brochures (< 30 sentences) as compared to long brochures (> 60 sentences) (p = 0.3).
CONCLUSIONS: Patient education brochures related to the use of anticoagulants are written at a grade level beyond the understanding of many patients. Low-literacy brochures are needed for patients receiving anticoagulants.
845 Domestic Violence - Perceptions and Knowledge of Medical Students
E. Carr, J. Wilson PhD, D. Grigsby MD, and D. Kwolek MD. University of Kentucky College of Medicine, Lexington, Kentucky.
Domestic violence is an important issue in the medical profession. Studies indicate that almost One quarter of U.S. women will be abused at least once during their lifetime. Therefore, undergraduate medical education should reflect this need with curriculum dealing with the detection, triage, and treatment of domestic violence. Pre-clinical and clinical medical students were surveyed to assess their knowledge in different areas of women's health (N=296). To assess domestic violence knowledge, the survey asked the students to list factors related to the detection, triage, and treatment of domestic violence. The number of correct answers the students were able to identify in each category were tallied and analyzed by chi-square analysis. 203 surveys were returned.
Results were analyzed by comparing the number of correct responses by gender and the number of correct responses by level of clinical experience. Pre-clinical was defined as enrollment in medical school years one and two. Clinical was defined as enrollment in year four of medical school. Women students correctly identified more than one factor related to the detection, triage, and treatment of domestic violence more often then male students across all levels of clinical experience (p=.016). Clinical students listed more than one factor related to the detection, triage, and treatment of domestic violence more often than pre-clinical students (p=0.004). No significant difference was found in the ability of clinical verses pre-clinical students to list more than one factor related to triage or treatment of domestic violence (p-0.956). This survey demonstrates the need for increased educational emphasis about domestic violence, especially in the areas of triage and treatment. The reasons why male students were less knowledgeable about domestic violence should be studied and remedied. Poor performance by clinical as well as pre-clinical students is cause for concern among medical educators. Changes incorporating triage and treatment information on domestic violence into the clinical years of the undergraduate curriculum are needed for better care of patients who are victims of domestic violence.
