From the Department of Emergency Medicine, University of Cincinnati College of Medicine (Drs Pancioli, Bullard, and Jauch), Institute for Health Policy and Health Services Research (Mr Perkis), and Department of Pediatrics, Children's Hospital Medical Center (Dr Grulee), Cincinnati, Ohio.
In 1994, the American Heart Association Stroke Council concluded that there were no data to support the routine use of supplemental oxygen in patients who had a stroke. More recently, supplemental oxygen has been suggested to be potentially detrimental. The purpose of this study was to determine the extent of oxygen use in ischemic stroke patients and whether patients receiving oxygen had indications for its use.
A literature search was performed to generate a comprehensive list of explicit criteria for supplemental oxygen use. When the literature disagreed, the criteria were included in the list to overestimate rather than underestimate the justification for oxygen use. A retrospective chart review of consecutive, nonintubated, ischemic stroke patients admitted to a university hospital was performed. Statistical tests and logistic regression models were constructed to identify the presence of unjustified oxygen use within the sample. Hospital charges were used to quantify opportunities for resource conservation.
A total of 167 patient charts were reviewed yielding a total of 600 inpatient days abstracted. One hundred two patients (61.1%) received oxygen during some portion of their hospitalization. Of the 322 days that patients received oxygen, 147 (45.6%) met at least 1 criterion for oxygen use. Of the 278 days that patients did not receive oxygen, 69 (24.8%) met at least 1 of the criteria for oxygen use. There were 384 days for which no criteria were met. Of these, a patient still received oxygen 45.6% of the time (175 days). Factors associated with oxygen use included the presence of at least 1 justifying criteria as well as increasing age and male sex. Withholding oxygen from those not medically justified by the criteria could produce resource savings of roughly 45%.
Using a literature-based list of criteria for supplemental oxygen use, only 45.6% of days of oxygen use were justified in our ischemic stroke population. This study demonstrates that oxygen therapy is commonly given to ischemic stroke patients without clear indication, and opportunities exist for substantial resource conservation.
IN 1994, the American Heart Association Stroke Council concluded that there were no data to support the routine use of supplemental oxygen in patients who had a stroke.1 While there are many explicit criteria for oxygen utilization among inpatients, a diagnosis of ischemic stroke alone is not adequate. More recently, investigators concluded that supplemental oxygen should not routinely be given to non–hypoxic stroke patients who experienced minor or moderate strokes. These authors state that further research is needed to determine specific indications for oxygen supplementation for patients who experienced severe strokes.2
In the United States, stroke is the third leading cause of death and a leading cause of adult disability.3 In 1993, Broderick et al4 estimated that there are 700 000 acute strokes per year in the United States. In 1993, it was determined that more than $30 billion per year was spent on the care of stroke patients.5 Dollars spent in the first 30 days after a stroke equal greater than 50% of the total costs in the first year after a stroke.6 As such, interventions designed to contain resource utilization for these patients represent potentially significant economic and resource savings.
With no proven rationale for supplemental oxygen use in ischemic stroke patients, and the potential for harm, the practice of routine supplemental oxygen therapy for these patients must be scrutinized. In addition, resource utilization may be decreased if unjustified oxygen therapy is curtailed. Several studies have shown that established guidelines for the use and discontinuation of respiratory therapy result in decreased expenditures without increased morbidity or mortality.7- 9
We performed an extensive MEDLINE search to review documented explicit criteria for supplemental oxygen utilization.7,10- 13 From this review, a table of explicit criteria for oxygen utilization was created (Table 1). When the literature disagreed, criteria were included in the list to overestimate rather than underestimate the justification for oxygen use.
We designed a retrospective chart review data collection instrument to gather demographic and oxygen utilization information including a daily check for any of the explicit criteria for supplemental oxygen utilization. The data instrument was tested independently in a series of 10 patients by 2 of the researchers to ensure interrater reliability and reproducibility. Results of this initial evaluation of the data collection tool led to nonsubstantial formatting changes for clarification only. The data instrument was retested independently in a series of 5 patients by 2 of the researchers, yielding excellent agreement. A computer interface for data entry was then developed.
All patients presenting to a large tertiary referral center with the diagnosis of acute ischemic stroke for the calendar year 1998, a total of 210 patient records, were reviewed. We excluded patients who were never admitted to the hospital, patients who were intubated, as well as 2 patients whose charts were unavailable. A total of 167 patient charts were reviewed. Each patient chart was abstracted for demographic and medical historical information. This was followed by an abstraction of each inpatient day, identifying documentation of oxygen utilization as well as any documentation relevant to any of the explicit criteria that justified oxygen utilization. In keeping with the goal of overestimating the justification of oxygen utilization, the explicit criteria were considered met if a single measure was documented that met the criteria. For example, if there was 1 measurement of a patient's respiratory rate greater than 24/min, then that explicit criterion was considered met for that day. For this study, prehospital documentation and supplemental oxygen use were not considered. A patient's evaluation period in the emergency department was considered part of the first hospital day.
The proportion of patients receiving oxygen without justifying criteria was tested to determine if it was statistically different than 0 to confirm or disprove the presence of resource savings opportunities. In addition, a logistic regression model was implemented to determine statistically significant relationships (α = .05) between the decision to give a patient oxygen therapy (outcome variable) and factors from the patient charts (covariates) including demographics, medical history, and whether the patient met at least 1 of the justifying criteria for oxygen use. The unit of analysis for the regression was individual patient days where the patients in our sample had varying hospital stay durations. The longest hospital stay measured was 17 days and the shortest was 1 day. Due to the unit of analysis, it was necessary to account for the correlation of the outcome between days within each patient. Generalized estimating equations were incorporated, assuming an autoregressive correlation structure between days for each patient, to account for the correlation.
Finally, in an attempt to quantify savings opportunities, the total charges along with the total respiratory charges for each patient were obtained from hospital billing records. Unfortunately, the usefulness of this billing data was questionable due to missing values and the likelihood of erroneous values. For example, several patients received respiratory services and yet had respiratory charges that were missing or equal to 0. Due to several discrepancies in the billing data, we decided to use the standard charges (fees) for both oxygen therapy setup and daily oxygen administration as proxies for actual costs to approximate proportional savings opportunities for respiratory resources due to oxygen therapy.
A total of 167 patient charts were abstracted. The geometric mean patient age was 61 years. The patients were 47% male and 53% female. The patients were 62% African American, 37% white, and 1% Asian/Pacific Islander.
From the 167 patient charts reviewed, a total of 600 inpatient days were abstracted. One hundred two (61.1%) of the 167 patients received oxygen during some portion of their hospitalization. Forty-five (44.1%) of these patients met at least 1 oxygen justification criterion on the day of oxygen initiation. Of the 322 total days that patients received oxygen, 147 (45.6%) met at least 1 of the criteria for oxygen use. Of the 278 days that patients did not receive oxygen, 69 (24.8%) met at least 1 of the criteria for oxygen use. Vital sign abnormalities such as a respiratory rate greater than 24/min and tachycardia greater than 100/min were the most common explicit criteria found. Table 2 lists the explicit criteria for oxygen therapy and the number and percentage of patients who met each criteria for hospital days during which at least 1 of the criteria was met. Table 2 also provides a breakout of this information by whether or not the patient was receiving supplemental oxygen at the time the criterion was met.
Of the 600 inpatient days abstracted, 384 (64.0%) did not meet any of the explicit criteria for oxygen use. Of the 384 days, patients still received oxygen therapy on 175 days (45.6%) of the days. The 99% confidence interval for this percentage is 39.7% to 51.5%. The confidence interval rejects any hypothesis that this percentage is a very small number close to 0 and confirms the presence of unjustified oxygen utilization based on the defined criteria.
While oxygen use was not limited to patients with a justifying criterion, logistic regression shows that meeting at least 1 criterion was significantly associated with the decision to administer oxygen (Table 3). We expanded the analysis by using logistic regression to simultaneously examine multiple factors. Factors associated with oxygen use included presence of at least 1 justifying criterion, increasing age, and male sex (Table 3).
At the time of the study, the standard charges for oxygen administration were a one-time setup fee of $29 for the visit and a daily fee of $34 for oxygen. Based on these fees, total oxygen therapy charges for our sample were $13 906. By administering oxygen to only patients in the sample for which oxygen is justified by the criteria, a savings of $6274, or 45.1%, would be realized based on 147 less days on oxygen and 44 less setup fees. Because these totals do not include charges for other respiratory services that are directly or indirectly related to a patient receiving oxygen, a savings of 45.1% would yield a higher absolute savings than the amounts reported herein.
While it seems to be intuitive that providing oxygen to a patient with an ischemic stroke would be beneficial, paradigms in clinical practice that are not evidence based should be questioned. Recent expert panel reports document the lack of utility for oxygen in patients without specified criteria.1 In addition, one recent report detailed the potential harm in this practice.2 We have generated a comprehensive list of explicit criteria that would tend to overestimate, rather than underestimate, the number of inpatient days for which oxygen use is justified. Even with very inclusive criteria, 54% of the days during which a patient received oxygen in our study did not meet a single one of the criteria.
Notably, of the 278 days that patients did not receive oxygen, 69 days (24.8%) met at least 1 of the criteria for oxygen use. It is impossible to determine from our review whether these patients would have benefited from supplemental oxygen. For this review, justification criteria for oxygen therapy were purposefully extremely liberal to define a lower boundary for the potential resource savings if oxygen therapy were based on explicit criteria. This study was not designed to evaluate the patient outcomes relative to oxygen utilization or the potential benefit of oxygen therapy for patients with ischemic stroke.
To confirm cost and resource savings opportunities, it must first be shown that a significant percentage of the inpatient days during which oxygen therapy is unjustified result in oxygen utilization. The confidence interval determined for this percentage shows with 99.5% confidence that there are at least 40% such cases in our study, thus demonstrating a significant percentage of unjustified oxygen utilization in one cohort of stroke patients.
While the presence of at least 1 of the justifying criteria was not an absolute determinant for oxygen therapy in our group of patients, a logistic regression reveals that it is a statistically significant factor. Our regression also determines statistically significant parameter estimates for age and male sex as factors. While this analysis does not imply that age and sex are direct inputs into the decision to administer oxygen, it does appear that they are directly or indirectly related to the decision in some manner. Criteria-based treatment guidelines might diminish the effect of such confounding factors.
Results of our study have revealed a significant potential for savings through the review of oxygen utilization in ischemic stroke patients. Using charge data, our study showed that 45% of oxygen therapy charges, as well as other related respiratory charges, could be saved. While there is clearly a disparity between charge data and potential cost savings with decreased resource utilization, the fact that only 46% of the days for which oxygen was utilized can be medically justified by the previously described criteria offers a significant hope of cost savings through decreased utilization of respiratory services.
Our study is limited because it is a retrospective chart review. While the documentation for each individual patient day was carefully reviewed for any criteria for oxygen use, it is possible that a patient met a criterion that was not recorded by a physician, nurse, or respiratory therapist.
There are also limitations to estimating the true potential cost savings or "cost-effectiveness" of decreasing inappropriate oxygen utilization for stroke patients. The definition of cost-effectiveness is "a change in health care resources used (ie, incremental cost) relative to a change in health outcomes (ie, incremental effectiveness) if a new intervention is adopted over the alternative."14 For this study, the true cost of oxygen utilization is an unknown; charges are often higher than actual cost, while some services are also underestimated by charges secondary to such procedures as cross-subsidization. When more than half of all oxygen utilization cannot be considered medically justified by a very broad sweeping and inclusive series of criteria, however, there is significant hope for a potential diminution of oxygen utilization, and the resources involved, for these patients.
Like so many disease processes, certain paradigms exist in clinical practice that may or may not have scientific or clinical validity. Automatic oxygen utilization for an ischemic process, while inherently pleasing, is unlikely to be medically justified for all patients. This study highlights one example where a careful look at clinical practice may offer a significant opportunity for resource utilization reduction for a disease of enormous proportion.
Accepted for publication April 18, 2001.
Corresponding author and reprints: Arthur M. Pancioli, MD, Department of Emergency Medicine, University of Cincinnati College of Medicine, 231 Albert Sabin Way, Cincinnati, OH 45267-0769 (e-mail: Arthur.Pancioli@uc.edu).
Thank you for submitting a comment on this article. It will be reviewed by JAMA Internal Medicine editors. You will be notified when your comment has been published. Comments should not exceed 500 words of text and 10 references.
Do not submit personal medical questions or information that could identify a specific patient, questions about a particular case, or general inquiries to an author. Only content that has not been published, posted, or submitted elsewhere should be submitted. By submitting this Comment, you and any coauthors transfer copyright to the journal if your Comment is posted.
* = Required Field
Disclosure of Any Conflicts of Interest*
Indicate all relevant conflicts of interest of each author below, including all relevant financial interests, activities, and relationships within the past 3 years including, but not limited to, employment, affiliation, grants or funding, consultancies, honoraria or payment, speakers’ bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued. If all authors have none, check "No potential conflicts or relevant financial interests" in the box below. Please also indicate any funding received in support of this work. The information will be posted with your response.
Register and get free email Table of Contents alerts, saved searches, PowerPoint downloads, CME quizzes, and more
Subscribe for full-text access to content from 1998 forward and a host of useful features
Activate your current subscription (AMA members and current subscribers)
Purchase Online Access to this article for 24 hours
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 4
Customize your page view by dragging & repositioning the boxes below.
and access these and other features:
Enter your username and email address. We'll send you a link to reset your password.
Enter your username and email address. We'll send instructions on how to reset your password to the email address we have on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.