Objective: To identify the usefulness of critical-incident reports for reflective learning and for classifying the types of experiences that medical students regarded most memorable during clinical clerkship at general medical ward and clinic.
Design: Descriptive qualitative study using the critical-incident technique.
Setting: General medical ward and clinic of a university hospital.
Participants: Fifth-year medical students that took part in the clinical care of patients during the clerkship from October 1, 1996 to September 31, 1997.
Measurements and Main Results: Ninety-five reports were collected over one year and subjected to the analysis. Reports were classified for seven major themes, with a mean of 1.85 themes per narrative. As for the distribution of themes, 68 reports (72%) dealt with biomedical topics, 45 (47%) with communication with patients and families, 26 (27%) with personal feelings, 19 (20%) with the physician's role, 11 (12%) with ethics, 7 (7%) with psychosocial subjects, and 1 (1%) with recognition of alternative medicine. Reports from men and women contained a similar mean number (men 1.82; women 2.09) and distribution of themes.
Conclusions: Medical students have divergent meaningful experiences on non-biomedical themes as well as biomedical issues during their general medicine rotation.

OBJECTIVE: (1) To estimate the prevalence of cancer as an cause of abdominal pain, (2) to identify useful information of history and physical examination, (3) to assess performance of potential screening tests, and (4) to formulate an efficient algorithm for distinguishing patients with cancer.
DESIGN: Prospective observational study.
SETTING: General medical outpatient clinic in a university teaching hospital.
PARTICIPANTS: 470 outpatients (men, 216; women, 254; mean age±standard deviation, 44.6±16.5; age range, 16 to 89) complaining of abdominal pain at their first visits.
MEASUREMENTS AND MAIN RESULTS: A cancer causing abdominal pain was found in 18 patients (3.8%) . These 18 patients with cancer and the remaining 452 patients without cancer were compared regarding clinical findings. Findings significantly more common in patients with cancer were: age of 40 years or over; unexplained weight loss; insidious onset; and duration of pain longer than one preceding month without relief. Among recorded physical findings, countenance reflecting severe pain, fever, and abdominal tenderness were significantly associated with cancer. Elevation of LDH, WBC, CRP, ESR, and CEA were moderately useful discriminators for patients with and without cancer. The logistic regression analysis with complete clinical data set covering history, physical examination and laboratory tests showed that the only findings significantly associated with cancer were unexplained weight loss (odds ratio: 18.9, 95% CI 1.5-20.1), duration of pain over one month (odds ratio: 24.8, 95% CI 1.6-27.9), countenance of severe pain (odds ratio: 159.1, 95% CI 4.5-162.1), and WBC>10, 000 (odds ratio: 22.3, 95% CI 1.1-31.0) .
CONCLUSIONS: Our data support the value of the combined use of history, physical and selected laboratory findings in detecting cancer among patients with abdominal pain. Application of such selected criteria holds the promise of more efficient care with judicious and effective use of endoscopy or sonography without lowering the quality of care.