Abstract 166: Use of Risk Stratification for the Management of TIA in the Emergency Department: Focus Group Results.

Abstract

Background: Evaluation of TIA cases in the emergency department (ED) represents a clinical dilemma because no firm guidelines exist as to their disposition (hospitalization vs. out-patient care). The ABCD 2 clinical prediction rule risk stratifies patients but little is known about how Emergency Medicine physicians (EMPs) use the rule in clinical decision making. We undertook focus groups with EMPs to determine their attitudes and use of the ABCD 2 score, and to understand how information on baseline risk, costs, compliance, and feasibility affect their decision making. Methods: Physicians from 2 EM practice groups in Michigan were invited to attend a focus group meeting. Data were collected on their knowledge, attitudes, and use of the ABCD 2 clinical prediction rule in the evaluation of TIA cases. Using a case vignette of a moderate risk patient (ABCD 2 score = 4, 7-day stroke risk = 6%), physicians were asked to choose between hospitalization or discharge for out-patient care. We then changed several baseline conditions, including 7-day stroke risk, health care costs, and compliance with out-patient follow-up, to determine under what conditions they altered the initial disposition decision. Results: Twenty two EMPs participated; all worked in community-based hospitals, 91% were male, 95% were EM board certified with an average of 16.5 years of EM experience. Respondents reported seeing an average of 6.7 (SD= 4.6) TIA patients per month. Sixty four percent (14/22) were familiar with the ABCD 2 score, but only 9% (2/22) used it regularly. Almost 60% (13/22) initially chose to hospitalize the moderate risk patient. Increasing the cost of the episode of care (from $3000 to $9000) did not change the decision to hospitalize for the majority (8/13, 62%) of EMPs. Only when 7-day stroke risk was lowered from 6% to 1% did the majority of EMPs (11/13, 85%) change their decision from hospitalization to outpatient care. Forty percent (9/22) initially chose to manage the moderate risk patient as an out-patient. A small increase in the cost of care (from $3000 to $3500) resulted in 56% (5/9) EMPs switching their decision from out-patient care to hospitalization, while a modest increase in stroke risk (from 6% to 10%) resulted in 78% (7/9) switching their initial decision. The choice of out-patient care was also influenced by the likelihood that patients would complete testing in the out-patient setting; if compliance dropped from 100% to 80% then half of the EMPs switched their decision from out-patient care to hospitalization. Increasing the number of hours that a patient would need to complete testing (from 4 to 12 hours) only had a modest impact on physician decision making. Conclusions: The ABCD 2 score was rarely used in practice. The decision to hospitalize was relatively insensitive to cost of care; 7-day stroke risk only influenced the decision when reduced to virtual certainty (1%). The decision to use out-patient care was more sensitive to cost of care, increases in stroke risk, and compliance in the out-patient setting. These data suggest future studies should focus on acceptable outpatient risks and costs to increase adoption of clinical prediction rules and appropriate decision making for TIA cases.