"Although our results should be confirmed with a randomized trial, clinicians should be aware that many patients who are borderline candidates for ICU admission may benefit from care provided in an ICU," Dr. Colin Cooke, from the University of Michigan, Ann Arbor, said.
To gauge the relationship between ICU admission and outcomes for elderly patients with pneumonia, the researchers took a look back at more than 1.1 million Medicare beneficiaries admitted to nearly 3,000 acute care hospitals with pneumonia from 2010 to 2012.
The team compared 30-day mortality between patients admitted to the ICU and those admitted solely to the hospital ward. They used instrumental variable analysis to account for selection bias inherent in this comparison.
Thirty percent of the cohort was admitted to the ICU. Patients living less than 3.3 miles of a hospital with high ICU admission were more apt to be admitted to the ICU than their peers living farther away, according to a report in the September 22/29 issue of JAMA.
For the 13 percent of patients whose ICU admission was deemed discretionary, ICU admission was associated with a significant reduction in 30-day mortality (14.8 percent for ICU care vs. 20.5 percent for general ward care, p<0.02).
There were no statistically significant differences in total costs or total Medicare payments between groups, "suggesting that ICU admission can save lives for lower-risk patients, and can do so at similar cost," note the authors of a linked editorial.
This study provides "important empirical evidence that ICU admission can benefit low-risk patients," write Dr. Ian J. Barbash and Dr. Jeremy M. Kahn, of the University of Pittsburgh School of Medicine in Pennsylvania.
The findings, they add, "argue against active efforts to reduce ICU admissions through triage guidelines or bed supply reductions, at least for older patients with pneumonia. In the current health care system, more judicious use of the ICU may well lead to higher mortality in some patient populations." Dr. Cooke said, "It would be premature for our study to warrant significant changes in clinical practice. Nevertheless, our results are intriguing and suggest that there indeed is a population of patients with pneumonia who may not be hospitalized in an ICU but who would otherwise benefit from ICU-level care. The sooner we can identify this group of patients clinically, the sooner we can begin the confirmatory studies."