Do-Not-Resuscitate Orders in Older Adults During Hospitalization: A Propensity Score-Matched Analysis

To explore the effect of the presence and timing of a do-not-resuscitate (DNR) order on short-term clinical outcomes, including mortality.

Retrospective cohort study with propensity score matching to enable direct comparison of DNR and no-DNR groups.

Large, academic tertiary-care center.

Hospitalized medical patients aged 65 and older.

Primary outcome was in-hospital mortality. Secondary outcomes included discharge disposition, length of stay, 30-day readmission, restraints, bladder catheters, and bedrest order.

Before propensity score matching, the DNR group (n=1,347) was significantly older (85.8 vs 79.6, p<.001) and had more comorbidities (3.0 vs 2.5, p<.001) than the no-DNR group (n=9,182). After propensity score matching, the DNR group had significantly longer stays (9.7 vs 6.0 days, p<.001), were more likely to be discharged to hospice (6.5% vs 0.7%, p<.001), and to die (12.2% vs 0.8%, p<.001). There was a significant difference in length of stay between those who had a DNR order written within 24 hours of admission (early DNR) and those who had a DNR order written more than 24 hours after admission (late DNR) (median 6 vs 10 days, p<.001). Individuals with early DNR were less likely to spend time in intensive care (10.6% vs 17.3%, p=.004), receive a palliative care consultation (8.2% vs 12.0%, p=.02), be restrained (5.8% vs 11.6%, p<.001), have an order for nothing by mouth (50.1% vs 56.0%, p=.03), have a bladder catheter (31.7% vs 40.9%, p<.001), or die in the hospital (10.2% vs 15.47%, p=.004) and more likely to be discharged home (65.5% vs 58.2%, p=.01).

Our study underscores the strong association between presence of a DNR order and mortality. Further studies are necessary to better understand the presence and timing of DNR orders in hospitalized older adults.