Abstract T P293: Determinants of Early Withdrawal of Care in Patients Who Die During Index Hospitalization for Acute Ischemic Stroke
Introduction: CMS recently announced plans to report risk-standardized stroke mortality. Decisions to withdraw medical care after acute ischemic stroke are based on numerous factors, not just the severity of the incident stroke. Withdrawal of care decisions have the potential to confound these adjusted mortality rates. We investigated factors associated with early withdrawal of care in patients hospitalized with AIS.
Methods: We used our institutional GWTG database (1474 consecutive stroke inpatients from 12/2009 - 12/2012) to identify 176 consecutive in-hospital mortalities; of these, 100 patients were transitioned to comfort care during hospitalization. Chart review was performed on these 100 patients to identify factors associated with early (< 5 days) withdrawal of care.
Results: Seventy-eight patients (78%) had early withdrawal of care. Table 1 shows the comparison of patients with early (<5 days) versus late (≥ 5 days) withdrawal of care decisions. As shown, Early-CMO patients were younger, and less often were Hispanic, had a college level or higher education, and annual income >$100,000. They had more often expressed a DNR request, had placed limits on life sustaining measures prior to admission or had a known health care proxy (84% vs. 68%). They had pre-existing comorbidities of prior MI, atrial fibrillation, cancer and dementia. They had slightly higher NIHSS scores, and lower rates of thrombolysis. Early-CMO patients had lower rates of post-stroke pneumonia.
Conclusion: Several non-clinical variables appear to be associated with early withdrawal of care decisions and contribute substantially to in-hospital stroke mortality. Many of these factors are non-modifiable, and the association with pneumonia may be confounded by the reduced opportunity to develop pneumonia in cases of very early death. Given the vigorous debate about preventable stroke mortality and public reporting, these data suggest the need for more complex mortality risk models.
Author Disclosures: N. Wingkun: None. S.F. Ali: None. T.K. Thankachan: None. L.H. Schwamm: Consultant/Advisory Board; Modest; Joint Commission; MA Dept of Public Health. Other; Modest; AHA Get With the Guidelines (unpaid); Coverdell Registry Advisor (unpaid); MGH provides telehealth services under contract to hospitals in northern new england. Employment; Significant; Director, MassGeneral TeleHealth. A.B. Singhal: None.
- © 2014 by American Heart Association, Inc.