A simple strategy for improving patient contact after ED discharge.

The purpose of this study was to assess strategies to improve telephone contact with adult patients discharged from the emergency department (ED). The basic procedure was a prospective, randomized, interventional trial of a convenience sample of patients 18 years or older being discharged from the ED. Patients were excluded if they had altered mental status or were unable to communicate with the College Research Associates (RAs). RAs asked intervention subjects a set of scripted questions confirming patients' telephone numbers and times for a follow-up call. Control subjects received routine discharge instructions from the ED staff. Subjects were called back within 4 days of ED discharge. Eighty-seven control subjects and 76 intervention subjects were enrolled. There were no significant demographic differences between the 2 groups. Forty-seven (54%) control subjects were contacted versus 58 (77%) in the intervention group (P <.003; Chi-square test). A simple patient interview conducted immediately before discharge confirming the patient's telephone number and setting a time for a follow-up call significantly improved patient follow-up contact rates.

False-negative and false-positive errors in abdominal pain evaluation: failure to diagnose acute appendicitis and unnecessary surgery.

OBJECTIVES: To test the hypothesis that physician errors (failure to diagnose appendicitis at initial evaluation) correlate with adverse outcome. The authors also postulated that physician errors would correlate with delays in surgery, delays in surgery would correlate with adverse outcomes, and physician errors would occur on patients with atypical presentations. METHODS: This was a retrospective two-arm observational cohort study at 12 acute care hospitals: 1) consecutive patients who had an appendectomy for appendicitis and 2) consecutive emergency department abdominal pain patients. Outcome measures were adverse events (perforation, abscess) and physician diagnostic performance (false-positive decisions, false-negative decisions). RESULTS: The appendectomy arm of the study included 1, 026 patients with 110 (10.5%) false-positive decisions (range by hospital 4.7% to 19.5%). Of the 916 patients with appendicitis, 170 (18.6%) false-negative decisions were made (range by hospital 10.6% to 27.8%). Patients who had false-negative decisions had increased risks of perforation (r = 0.59, p = 0.058) and of abscess formation (r = 0.81, p = 0.002). For admitted patients, when the inhospital delay before surgery was >20 hours, the risk of perforation was increased [2.9 odds ratio (OR) 95% CI = 1.8 to 4.8]. The amount of delay from initial physician evaluation until surgery varied with physician diagnostic performance: 7.0 hours (95% CI = 6.7 to 7.4) if the initial physician made the diagnosis, 72.4 hours (95% CI = 51.2 to 93.7) if the initial office physician missed the diagnosis, and 63.1 hours (95% CI = 47.9 to 78.4) if the initial emergency physician missed the diagnosis. Patients whose diagnosis was initially missed by the physician had fewer signs and symptoms of appendicitis than patients whose diagnosis was made initially [appendicitis score 2.0 (95% CI = 1.6 to 2.3) vs 6.5 (95% CI = 6.4 to 6.7)]. Older patients (>41 years old) had more false-negative decisions and a higher risk of perforation or abscess (3.5 OR 95% CI = 2.4 to 5.1). False-positive decisions were made for patients who had signs and symptoms similar to those of appendicitis patients [appendicitis score 5.7 (95% CI = 5.2 to 6.1) vs 6.5 (95% CI = 6.4 to 6.7)]. Female patients had an increased risk of false-positive surgery (2.3 OR 95% CI = 1.5 to 3.4). The abdominal pain arm of the study included 1,118 consecutive patients submitted by eight hospitals, with 44 patients having appendicitis. Hospitals with observation units compared with hospitals without observation units had a higher "rule out appendicitis" evaluation rate [33.7% (95% CI = 27 to 38) vs 24.7% (95% CI = 23 to 27)] and a similar hospital admission rate (27.6% vs 24.7%, p = NS). There was a lower miss-diagnosis rate (15.1% vs 19.4%, p = NS power 0.02), lower perforation rate (19.0% vs 20.6%, p = NS power 0.05), and lower abscess rate (5.6% vs 6.9%, p = NS power 0.06), but these did not reach statistical significance. CONCLUSIONS: Errors in physician diagnostic decisions correlated with patient clinical findings, i.e., the missed diagnoses were on appendicitis patients with few clinical findings and unnecessary surgeries were on non-appendicitis patients with clinical findings similar to those of patients with appendicitis. Adverse events (perforation, abscess formation) correlated with physician false-negative decisions.

Implementing emergency department observation units within a multihospital network.

BACKGROUND: The proportion of emergency department (ED) chest pain patients who undergo an extended "rule out MI (myocardial infarction)" evaluation beyond the ED determines both the quality and cost of patient care. The higher an organization's rate of such evaluations, the lower the average miss rate for MI. Five of the 13 hospitals in the Voluntary Hospital Association Northeast multihospital network implemented ED observation units by June 1997 for outpatient rule out MI evaluations. RESULTS: Compared with historical and case controls, the five hospitals with ED observation units had a higher observation rate (16% versus 0% [p < .001] and 2% [p < .001]) and a higher rule out MI evaluation rate (61% versus 46% [p < .01] and 45% [p < .01]), without a significantly higher admission rate (47% versus 46% and 45%). For the three hospitals with observation units that collected charge data during 1997 on a consecutive series of chest pain patients who had negative rule out MI evaluations, charges for patient services were lower for patients evaluated in the ED observation unit ($2,214.80 +/- $80.40) than in the hospital ($5,464.30 +/- $393.60). CONCLUSIONS: ED observation units represent a cost-effective restructuring of the diagnostic approach to patients with acute chest pain. In an improvement of quality of patient care, a larger proportion of ED chest pain patients receive an extended evaluation than is possible with hospital admission as the only ED disposition option.

The Goldman algorithm revisited: prospective evaluation of a computer-derived algorithm versus unaided physician judgment in suspected acute myocardial infarction.

BACKGROUND: It has been nearly a decade since Goldman's computer-driven algorithm to predict myocardial infarction was validated. Despite the potential to avoid admission of patients without acute myocardial infarction (AMI) to the coronary care unit (CCU), the routine use of computer-generated protocols has not been widely adopted. METHODS: Two hundred consecutive patients admitted to a university-affiliated community hospital with the suspected diagnosis of AMI as determined by physicians without the aid of the Goldman protocol underwent a blinded prospective evaluation to assess the performance of the Goldman algorithm in predicting the presence of AMI. Over the same time period, the Goldman algorithm was applied by retrospective chart review in 762 patients with non-AMI admitting diagnoses. Prospective history, physical examination, and electrocardiographic data were obtained within 24 hours of admission to the CCU by a physician blinded to each patient's clinical course. Retrospective chart reviews were conducted for 762 patients with chest pain given with non-AMI diagnoses. RESULTS: The diagnosis of AMI was confirmed in 68.5% (137/200) of patients with suspected AMI admitted to the CCU. In prospective parallel evaluations the Goldman algorithm predicted the presence of AMI in 167 (83.5%) of these 200 patients. All 137 confirmed patients with AMI were correctly identified by the Goldman algorithm. All major in-hospital complications occurred in the 137 patients who were diagnosed as having AMI. Of the 762 patients with chest pain with non-AMI diagnoses, only 27 (3.5%) sustained an AMI. The Goldman algorithm predicted the presence of AMI in 85% (23/27) of these patients. Adherence to the use of Goldman's algorithm in the triage of chest pain could have prevented 16.5% of CCU admissions for AMI. CONCLUSIONS: Routine adherence to the Goldman algorithm for the evaluation of patients with acute chest pain could have decreased the number of CCU admissions for suspected AMI by 16. 5%. Because major in-hospital complications occurred only in patients with AMI, this strategy would result in significant cost savings to our health care system without jeopardizing patient safety.