ICU Admission or General Ward for Diabetic Ketoacidosis? The Answer Varies Dramatically in Different Hospitals
Abstract & Commentary
By David J. Pierson, MD
Professor Emeritus, Pulmonary and Critical Care Medicine, University of Washington, Seattle
This article originally appeared in the September 2012 issue of Critical Care Alert. It was peer reviewed by William Thompson, MD. Dr. Thompson is Associate Professor of Medicine, University of Washington, Seattle. Drs. Pierson and Thompson report no financial relationships relevant to this field of study.
Synopsis: In a large cohort of patients admitted to New York hospitals with diabetic ketoacidosis, about half were admitted to the ICU, with a range of 2% to 88% among individual hospitals. This large practice variation was unassociated with mortality or length of stay, and more than half of it remained unaccounted for after extensive adjustments for patient and institutional characteristics.
Source: Gershengorn HB, et al. Variation in use of intensive care for adults with diabetic ketoacidosis. Crit Care Med 2012;40:2009-2015.
Using a statewide administrative database and other sources, Gershengorn and colleagues examined data on all adult patients with a primary diagnosis of diabetic ketoacidosis (DKA) who were admitted to hospitals in the state of New York from 2005-2007. The investigators sought to determine what proportion of these patients were admitted to the ICU, and what associations with patient or hospital characteristics or other identifiable aspects of care might explain any observed differences.
During the study period, there were 15,994 patient admissions for DKA to 159 hospitals. Most of the hospitals were in urban settings and about half of them were teaching hospitals. Median hospital size was 190 beds with 9.4% of these being ICU beds. The DKA admissions represented 0.4% of all hospital and 1.4% of all ICU admissions to the study hospitals during the 3-year study period. Median reliability- and risk-adjusted hospital mortality was 0.7% (range, 0.4% to 3.4%), and median hospital length of stay was 3 days (range, 1 to 6 days).
Of the admissions for DKA, 52.6% were admitted to an ICU. These patients tended to be younger, white, privately insured, from a higher-income zip code, and admitted on the weekend, with all these differences being statistically significant. They were also more likely to have chronic illnesses and be admitted emergently. The proportion of DKA admissions receiving intensive care varied dramatically across hospitals, with an adjusted range of 2.1% to 87.7%. However, this variation was not associated with mortality or hospital length of stay. ICU admission occurred less often in hospitals that admitted larger numbers of patients with DKA (highest quartile vs lowest, odds ratio 0.40, P = 0.002), but more often in hospitals with higher rates of ICU admission for non-DKA admissions (odds ratio 1.31, P = 0.001, for each additional 10% increase). Using multilevel modeling to account for individual patient and hospital factors, the authors were able to explain less than half of the observed variation in ICU utilization for patients with DKA: 58% of the variability attributable to hospitals could not be explained.
Commentary
Like acute gastrointestinal bleeding without hypotension, DKA is a common reason for ICU admission that carries a low risk for mortality, and also has a thoroughly studied and highly protocolized management approach. The fact that this study found no differences in mortality or hospital length of stay in nearly 16,000 DKA admissions, only half of which included care in an ICU, reinforces the concept that a large proportion of such admissions can be handled safely and effectively on the acute-care wards. Despite the large number of patients included, this retrospective study based on administrative data cannot determine the reasons for the differences in ICU admission rates among the various hospitals. However, Gershengorn et al nicely demonstrate that the issue of where to manage DKA patients is currently being approached very differently in different institutions, with a rate of ICU admission varying from 1 in 50 to more than 4 out of every 5 such patients.
As the authors point out, their findings can be interpreted in various ways. It may be that patients admitted with DKA can be managed just fine without ICU admission. However, it is also possible that such patients are already being triaged appropriately at all the study hospitals, such that those who really need ICU care (and are thus having their outcomes improved accordingly) are getting it. The latter interpretation, while possible, seems less likely to me in view of the enormous practice variation documented across the 159 hospitals. I suspect that much of the observed variation relates to traditional practice patterns and other aspects of institutional culture in the different hospitals.
Practice variation has been identified as an important problem in health care,1 and its reduction is currently a major target for many quality and safety initiatives. However, as recently emphasized in a tri-society statement on the appropriate use of clinical research data and other types of knowledge in critical care, practice variation is inevitable in a high-stakes field with an incomplete and sometimes contradictory database.2 How much this assertion applies to DKA is uncertain. As the authors of the current study caution, further research is needed to clarify the most effective and cost-efficient use of the ICU for patients admitted with DKA.
References
1. Kennedy PJ, et al. Clinical practice variation. Med J Aust 2010; 193(8 Suppl):S97-99.
2. Tonelli MR, et al; ACCP/ATS/SCCM Working Group. An official multi-society statement: The role of clinical research results in the practice of critical care medicine. Am J Respir Crit Care Med 2012;185:1117-1124.
In a large cohort of patients admitted to New York hospitals with diabetic ketoacidosis, about half were admitted to the ICU, with a range of 2% to 88% among individual hospitals. This large practice variation was unassociated with mortality or length of stay, and more than half of it remained unaccounted for after extensive adjustments for patient and institutional characteristics.You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content