Secondary Logo

Journal Logo

Characteristics of Desaturation and Respiratory Rate in Postoperative Patients Breathing Room Air Versus Supplemental Oxygen: Are They Different?

Taenzer, Andreas H. MS, MD*; Perreard, Irina M. PhD*; MacKenzie, Todd PhD; McGrath, Susan P. PhD*

doi: 10.1213/ANE.0000000000002765
Technology, Computing, and Simulation: Original Clinical Research Report

BACKGROUND: Routine monitoring of postoperative patients with pulse oximetry–based surveillance monitoring has been shown to reduce adverse events. However, there is some concern that pulse oximetry is limited in its ability to detect deterioration quickly enough to allow for intervention in patients receiving supplemental oxygen. To address such concerns, this study expands on the current limited knowledge of differences in desaturation and respiratory rate characteristics between patients breathing room air and those receiving supplemental oxygen.

METHODS: Pulse oximetry–derived data and patient characteristics were used to examine overnight desaturation patterns of 67 postoperative patients who were receiving either supplemental oxygen or breathing room air. The 2 modalities with respect to the speed of desaturation, in addition to magnitude and duration of desaturation events, are compared. Night-time pulse rate, oxygen saturation, respiratory rate, and the transition times from normal oxygen saturation levels to desaturated states are also compared. The behavior of respiratory rate in proximity to desaturation events is described. Statistical methods included multivariable regression and inverse probability of treatment weighted to adjust for any imbalance in patient characteristics between the oxygen and room air patients and linear mixed effect models to account for clustering by patient.

RESULTS: The study included 33 patients on room air and 34 receiving supplemental oxygen. The speed of desaturation was not different for room air versus oxygen for 2 types of desaturation (adjusted % difference, 95% confidence interval [CI]: type I; 22.4%, −51.5% to 209%; P = .67, type II; −17.3%, −53.8% to 47.6%; P = .52). Patients receiving supplemental oxygen had a higher mean oxygen saturation (adjusted difference, 95% CI, 2.4 [0.7–4.0]; P = .006). No differences were found for the average overnight respiratory or pulse rate, or proportion of time in desaturation states between the 2 groups.

The time to transition from a normal oxygen saturation (92%) to 88% or below was not longer for supplemental oxygen patients (P = .42, adjusted difference 26.1%: 95% CI, −28.1% to 121%). Respiratory rates did not differ between the overall mean and desaturation or recovery phases or between the oxygen and room air group.

CONCLUSIONS: In this study, desaturation characteristics did not differ between patients receiving supplemental oxygen and breathing room air with regard to speed, depth, or duration of desaturation. Transition time for desaturations to reach low oxygen saturation alarms was not different, while respiratory rate remained in the normal range during these events. These findings suggest that pulse oximetry–based surveillance monitoring for deterioration detection can be used equally effectively for patients on supplemental oxygen and for those on room air.

From the *Department of Anesthesiology, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire

Department of Biomedical Data Science, Dartmouth College, Hanover, New Hampshire.

Published ahead of print December 29, 2017.

Accepted for publication November 9, 2017.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website.

Funding: This project was supported by Grant Number P30HS024403 from the Agency for Healthcare Research and Quality. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Agency for Healthcare Research and Quality.

The authors declare no conflicts of interest.

Reprints will not be available from the authors.

Address correspondence to Andreas H. Taenzer, MS, MD, Department of Anesthesiology, Dartmouth-Hitchcock Medical Center, One Medical Center Dr, Lebanon, NH 03756. Address e-mail to

Hypoxemic episodes are a common occurrence in the postoperative patient population. In 1 study of patients continuously monitored after surgery with pulse oximetry, 21% had desaturations to 90% for longer than 10 minutes per hour, 8% for longer than 20 minutes per hour, and 8% had saturations of <85% for more than 5 minutes per hour.1 While the long-term consequences of hypoxic episodes are unknown, hypoxemia indicates instability, with detrimental short-term consequences ranging from wound healing2 to myocardial ischemia.3 Hypoxic states can be reduced, but not eliminated, by supplemental oxygen,4 and such episodes are less likely to be detected by intermittent vital sign spot checks as compared to continuous pulse oximetry monitoring.5,6 Lee et al7 reported in a closed claims analysis of postoperative opioid-induced respiratory depression that 97% of episodes should be preventable with better monitoring and response. Patient surveillance via continuous monitoring has been introduced to assist in the early detection of low oxygen saturation states and prevention of failure-to-rescue events with promising results.8–10

While continuous pulse oximetry monitoring has reduced unanticipated transfers to the intensive care unit and activations of rapid response teams9,10 by prompting early intervention, there is concern about the ability of pulse oximetry to detect respiratory deterioration with rapid desaturations in patients on supplemental oxygen with enough warning time to allow for effective intervention.11 The American Society of Anesthesiology12 and the Anesthesia Patient Safety Foundation13 recommend the assessment of respiratory status, including respiratory rate (RR), for postoperative patients on opioids. However, the role of RR in desaturation events in postoperative care is not well defined, although RR has been reported to remain unchanged in patients with desaturations who received intrathecal opioids and that most low saturation states occur with RRs in normal ranges.14,15

The objective of this study is to explore 2 specific areas related to the application of pulse oximetry–based continuous monitoring in postoperative inpatients. The first aim is to understand whether there are indeed differences in desaturation features such as duration, depth, and speed between patients receiving supplemental oxygen and those breathing room air (RA) that might affect the timeliness of intervention and rescue. The second aim is to understand RR variation during desaturations to highlight the impact that this parameter might have in patient surveillance in general, and for patients on supplemental oxygen, more specifically.

Back to Top | Article Outline


Setting and Data Collection

This was an observational study approved by the Center for the Protection of Human Subjects at Dartmouth College. The study was performed at Dartmouth-Hitchcock Medical Center, in Lebanon, NH, where surveillance monitoring using pulse oximetry is used for all patients residing in general care medical and surgical units, unless they are ambulating, in the presence of a caregiver, or whether use of the monitoring system presents a risk to the patient. For this study, surveillance system monitoring data were collected using Masimo Radical 87 (Irvine, CA) devices from patients in a 36-bed surgical unit, primarily serving orthopedic patients. Pulse oximetry–derived data includes peripheral oxygen saturation (Spo2, percent saturation) and pulse rate (PR, beats per minute). Acoustic RR monitoring (breaths per minute, Masimo sensor model RAS 125c) was also in place during the period of study. Surveillance data were collected from each patient at a rate of 1 Hz and stored permanently in a database for further analysis.

Each morning for a 2-week period, bedside nurses recorded whether the patient was breathing RA or receiving supplemental oxygen (O2) the previous night. Only adult patients who were either continuously on RA or on O2 during their first or second night postoperatively became part of the study cohort. Oxygen was administered via nasal cannula or simple facemask (no nonrebreathing masks) with an oxygen flow rate between 1 and 6 L·minute−1. Continuous physiologic data (Spo2, PR, RR) for the period between 11 pm and 5 am, when all patients were expected to be asleep and continuously monitored, were collected for each patient. For inclusion in the study, patients were required to have at least 70% data availability over the 6-hour period. Patients who transitioned from receiving supplemental O2 to breathing RA or vice versa were not included.

Back to Top | Article Outline

Dataset Description

Electronic medical records and administrative data were used to obtain age, gender, body mass index (BMI), history of respiratory disease, respiratory medications, diagnostic-related group (DRG), and DRG relative weight (indicates severity of illness). Records for each patient were reviewed and the history of respiratory disease and use of respiratory medications (such as bronchodilators) were assigned a “yes” or “no” flag.

Back to Top | Article Outline

Methods for Defining Desaturation



For each patient, segments of the Spo2 time series data that represent states of oxygen saturation for the 6-hour period of interest were segmented and labeled using an algorithm implemented with MATLAB (R2014a & R2016a; The MathWorks, Natick, MA). The algorithm identified 3 types of desaturation state, associated recovery states, and a normal state as described in Supplemental Digital Content, Table 1,, and illustrated in Figure A, based on established criteria.16–19 Desaturation type I was defined as a decrease in Spo2 of at least 4% over at least 10 seconds with a desaturation minimum value below 92%. This type typically represented a desaturation from a clinically normal state to a low saturation state. Recovery for desaturation type I represented an increase in oxygen saturation level from the minimum value of a desaturation to a value of at least 92%. Desaturation type II was defined as decrease in Spo2 of at least 4% over at least 10 seconds, with a minimum desaturation value above 92%. Type II represented a desaturation where Spo2 remains in a clinically normal range. Desaturation type III was a decrease in Spo2 of at least 4% and at least 10 seconds long, with a minimum value below 92% but where the recovery, if it exists, does not reach 92%. These typically represented desaturations that have an onset from a low saturation value or desaturations with only modest recovery. Recovery for desaturation types II and III was defined as an increase in oxygen saturation levels from the minimum desaturation point to at least within 2% of the desaturation onset value.

Back to Top | Article Outline

Outcome Measures

The primary analysis end point was the speed (Figure B), or rate of change, of desaturation, for both type I and II desaturations. Related but secondary end points are the transition times from normal levels of Spo2 (92% or above) to low Spo2 states. Three transition time end points were evaluated; the time to transition from 92% to >88%, from 92% to >84%, and from 92% to >80%. Other secondary end points were the proportion of time spent in the desaturation, recovery, and normal states, and the duration and magnitude (as illustrated in Figure B) of type I and II desaturations as defined above, as well as physiological characteristics, Spo2, RR, and PR during the entire night.

The end points based on type I and type II desaturations, including the primary end point, speed, magnitude, and duration, were aggregated to the patient level by calculating the mean across all of a patient’s desaturations (type I and type II calculated separately). Likewise, Spo2, RR, and PR are aggregated to the patient level by using the average value of each of these physiological characteristics during the entire night of their study follow-up. Unlike the former end points, transition times were analyzed at the patient-desaturation level as explained below.

Back to Top | Article Outline

Statistical Analysis

Descriptive statistics are reported for desaturation features and physiologic values for the O2 and RA groups. To estimate the effect of mode of oxygen delivery, multivariable linear regression to control for any imbalance in patient characteristics between the O2 and RA patients was used. In addition, to gauge the sensitivity of the findings to the statistical method used, inverse probability of treatment weighted estimates of the difference between O2 and RA outcomes was used. The latter is not reported as the findings were substantively equivalent.

The dependent variables, related to transition time from normal to desaturation alarm thresholds (92% to >88%, 92% to >84%, 92% to >80%) that occur with multiplicity in patients (ie, clustered by patient) were analyzed using a mixed-effects linear model to model these outcomes with a random intercept for patient and fixed effects for the primary exposure of interest, O2 versus RA, and the patient characteristics. As a sensitivity analysis inverse probability of treatment weighted in a linear mixed-effects model with the same random intercept of patient and a single fixed effect, mode of delivery was used.

The propensity to be exposed to oxygen versus RA was modeled in terms of the baseline characteristics reported in Table 1 using logistic regression. Inverse probability of treatment weighting (the probabilities derived from the logistic regression) achieves balance with respect to the baseline characteristics between the 2 exposure groups. To select the subset of baseline characteristics in the model for the propensity, we used logistic regression in combination with least absolute shrinkage and selection operator (penalized regression with a L1 penalty) to parse a subset of covariates. The penalty parameter that optimized jackknifed likelihood was used, which corresponded to 4 variables in the model: BMI, DRG relative weight, use of respiratory medication, and gender. These same 4 covariates identified above were controlled for in the multivariable linear regression. To confirm that inverse probability of treatment weighting achieved balance in baseline patient characteristics between the O2 and RA groups, we calculated standardized differences and tests of equality of weighted means and proportions. We ensured that the corresponding P values were >.10 and that no standardized differences exceeded 0.30.

Table 1.

Table 1.

The distributions of the desaturation speed, magnitude, and duration were skewed to the right. Therefore, geometric means were reported. Correspondingly, log transformations of these end points were set as the dependent variables in the regression models used to estimate the effect of mode of oxygen delivery. The use of log transforms facilitates reporting of the treatment effect as a percentage change (eg, multiplicatively).

A sample size of 30 patients in each of the O2 and RA groups was targeted. This allows detection of Cohen effect sizes of 0.7 with 80% power. For instance, for RR analysis during desaturations, which has a standard deviation between subjects of approximately 4, the study is powered to detect a difference of 3.

Statistical analysis was performed using R version 3.3.1 (R Foundation for Statistical Computing).

Back to Top | Article Outline


Patient Characteristics

Data were analyzed from 34 O2 patients and 33 RA patients. The O2 group had a mean age of 64.1 ± 11.4 years with 56% being female; the RA group mean age was 62.6 ± 9.7 years with 39% being female (Table 1). BMI was higher (P = .01) in the O2 group (33.4 ± 7.9) as compared to the RA group (29.0 ± 5.4). The majority of patients in each group had orthopedic-related procedures (O2 64%, RA 76%) as opposed to nonorthopedic procedures. Patients in the O2 cohort were coded with higher (P = .045) illness severity (DRG relative weight) than the RA patients, with mean values of 3.3 ± 2.9 and 2.9 ± 0.9, respectively. Patients in the O2 group also had a greater history of respiratory disease diagnosis at 38.2% vs 9.1% for the RA group (P = .01), but not in use of respiratory medications.

Back to Top | Article Outline

Physiologic Parameters

Table 2.

Table 2.

Table 2 displays summary descriptive statistics for Spo2, PR, and RR based on the monitoring data analyzed for each patient from the 6-hour overnight period. Mean Spo2 for the O2 group was higher (adjusted difference: 2.4, 95% confidence interval [CI], 0.7–4.0; P = .008) than for RA patients. There was no significant difference in mean RR (adjusted difference: −0.2, 95% CI, −2.6 to 2.1; P = .75) or PR (adjusted difference: −5.6, 95% CI, −2.6 to 2.1; P = .15) between groups.

Back to Top | Article Outline

Proportion of State Types

Table 3.

Table 3.

Table 3 shows the average percentage of time patients spent in normal, desaturation, and recovery states as defined in the addendum. The table also includes proportion of total desaturation time spent in type I and II desaturation states. Desaturations of type III accounted for a very small fraction (0.8% in the O2 and 2.7% in the RA cohort) of the total time spent in a desaturation state (4 occurrences in 3 patients in the O2 group and 24 in 9 patients in the RA group, Fisher exact P = .13), and were therefore excluded from further analysis. Patients on supplemental oxygen spent more time in a normal state, (adjusted difference: 15.5, 95% CI, 3.1–27.9; P = .02).

Back to Top | Article Outline

Desaturation Features

Table 4.

Table 4.

Table 4 presents summary statistics for speed, duration, and magnitude of type I and II desaturations by group comparisons. No differences were identified with the exception that O2 patients had larger magnitudes (adjusted percentage difference: 9.8%, 95% CI, 11.4%–18.9%; P = .03). However, this is not a significant difference if the multiplicity of comparisons is considered (eg, Bonferonni correction).

Back to Top | Article Outline

Transition Times From Normal to Abnormal Spo2

Table 5.

Table 5.

Table 5 shows statistics for transitions between normal Spo2 (92% or above) and 3 thresholds for low Spo2 (88%, 84%, and 80%). For instance, 18 patients on supplemental oxygen transition from 92% to 88% and among them the median number of such transitions was 5. The geometric mean of these transitions was 1.0 minutes while it was 0.8 minutes in the RA patients. No differences in any of the 3 types of transition times (92% to >88%, 92% to >84%, or 92% to >80%) between patients on O2 and RA were identified.

Back to Top | Article Outline


This study examines key issues related to inpatient monitoring by providing insight into typical postoperative desaturation features and RRs associated with desaturations. In this study of 67 postoperative patients breathing RA or receiving supplemental oxygen in a general care unit setting, the main finding is that there was almost no difference in desaturation characteristics between the 2 cohorts. The magnitude, duration, and speed of desaturation were not different between patients breathing RA versus supplemental oxygen except for 1 outcome with limited clinical relevance. There were also no significant differences in the time spent in desaturated states or the time to transition from normal oxygen saturation rates to typical oxygen saturation alarm threshold values. Mean transition times for desaturations to reach 80% from 92% were 2.6 (RA) and 4.6 (O2) minutes, respectively, with a lower 95% CI of 1.4 and 2.5 minutes. These times highlight the need for rapid intervention once identified and emphasize the need for continuous monitoring in the general care setting, as intermittent vital sign checks lack the temporal resolution to detect these changes in time. RRs remained in the normal range during both desaturation and recovery periods. This finding is consistent with other analysis of RRs during desaturation events15 and suggests that oxygen saturation would provide an earlier indication of deterioration than RR in these circumstances. RR alone is likely not to be helpful as a sensor without using pulse oximetry to detect deterioration of respiratory function in many cases of deterioration.

Based on these observations, there is little reason to assume that patients receiving supplemental oxygen are at a disadvantage in terms of the time available for clinical intervention when pulse oximetry–based surveillance monitoring is used with the same alarm delays and thresholds9,10 used for patients breathing RA.

The only significant difference was in the mean decrease of saturations for desaturation processes that did not cross below 92% (type II). While the average oxygen saturation drop was a greater in the supplemental oxygen group for desaturations not crossing below 92%, the average baseline oxygen saturation was higher in the oxygen group and desaturations inside oxygen saturations states considered to be physiologic are of lesser clinical significance and normally not detected.

These data provide insight into the typical and most common desaturation events in a postoperative population. While the lack of tidal volume data makes it impossible to demonstrate with certainty, the preponderance of normal RRs while receiving supplemental oxygen suggest that these desaturation events may be explained by shunting of blood from atelectasis due to shallow breathing.20 Therefore, administration of supplemental oxygen does not alter the characteristics of the desaturations that are reflected in the results found within.

Clinical outcome–based studies comparing the use of pulse oximetry–based patient surveillance in patients breathing oxygen versus RA are difficult to conduct because of the large number of cohorts required because of the rarity of adverse events (as they are reduced by using surveillance). The description of desaturation and RR characteristics was chosen to provide some insight in possible differences between the 2 groups.

The primary limitation of the data presented in this study is that severe and prolonged desaturations or apnea phases are not contained in these data. This is due to the fact that pulse oximetry–based patient surveillance is standard of care at the study institution and desaturations below 80% trigger an alert to the covering nurse and prompt an intervention.10 It would be unethical to expose patients to risk while having the means to prevent deterioration. The 2 statistical approaches used to control for confounding, multivariable regression, and inverse probability of treatment weighting do not control for confounding variables that were not assessed (ie, those not reported in Table 1). Furthermore, the relatively small sample size limits the number of variables that can be controlled for in the multivariable regression and inverse probability of treatment weighting. Therefore any estimates of the difference in effect of oxygen versus RA may be biased due to unmeasured factors that influence the choice to use both supplemental oxygen and desaturations.

This study is not meant to suggest that the administration of supplemental oxygen to postoperative patients in general care units is or is not safe practice. But when monitoring patients with a pulse oximetry–based alarm notification system, desaturation characteristics—specifically the speed of the desaturation and the transition time to a desaturation alarm state—are not different between patients breathing RA versus supplemental oxygen. Therefore, these findings suggest that pulse oximetry–based surveillance monitoring for deterioration detection can be used equally effectively for patients on supplemental oxygen and for those on RA.

Back to Top | Article Outline


Name: Andreas H. Taenzer, MS, MD.

Contribution: This author helped with the design of the study; the acquisition, analysis, or interpretation of data for the study; drafted the study or revised it critically for important intellectual content; approved the final version to be published; and agreed to be accountable for all aspects of the study in ensuring that questions related to the accuracy or integrity of any part of the study are appropriately investigated and resolved.

Name: Irina M. Perreard, PhD.

Contribution: This author helped design of the study; the acquisition, analysis, or interpretation of data for the study; drafted the study or revised it critically for important intellectual content; approved the final version to be published; and agreed to be accountable for all aspects of the study in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Name: Todd MacKenzie, PhD.

Contribution: This author helped design of the study; the acquisition, analysis, or interpretation of data for the study; drafted the study or revised it critically for important intellectual content; approved the final version to be published; and agreed to be accountable for all aspects of the study in ensuring that questions related to the accuracy or integrity of any part of the study are appropriately investigated and resolved.

Name: Susan P. McGrath, PhD.

Contribution: This author helped design of the study; the acquisition, analysis, or interpretation of data for the study; drafted the study or revised it critically for important intellectual content; approved the final version to be published; and agreed to be accountable for all aspects of the study in ensuring that questions related to the accuracy or integrity of any part of the study are appropriately investigated and resolved.

This manuscript was handled by: Maxime Canneson, MD, PhD.

Back to Top | Article Outline


1. Sun Z, Sessler DI, Dalton JE, et al. Postoperative hypoxemia is common and persistent: a prospective blinded observational study. Anesth Analg. 2015;121:709–715.
2. Greif R, Akça O, Horn EP, Kurz A, Sessler DI; Outcomes Research Group. Supplemental perioperative oxygen to reduce the incidence of surgical-wound infection. N Engl J Med. 2000;342:161–167.
3. Rosenberg J, Rasmussen V, von Jessen F, Ullstad T, Kehlet H. Late postoperative episodic and constant hypoxaemia and associated ECG abnormalities. Br J Anaesth. 1990;65:684–691.
4. Stone JG, Cozine KA, Wald A. Nocturnal oxygenation during patient-controlled analgesia. Anesth Analg. 1999;89:104–110.
5. Taenzer AH, Pyke J, Herrick MD, Dodds TM, McGrath SP. A comparison of oxygen saturation data in inpatients with low oxygen saturation using automated continuous monitoring and intermittent manual data charting. Anesth Analg. 2014;118:326–331.
6. Stoelting RK. Continuous postoperative electronic monitoring and the will to require it. Anesth Analg. 2015;121:579–581.
7. Lee LA, Caplan RA, Stephens LS, et al. Postoperative opioid-induced respiratory depression: a closed claims analysis. Anesthesiology. 2015;122:659–665.
8. Abenstein JP, Narr BJ. An ounce of prevention may equate to a pound of cure. Anesthesiology. 2010; 112:272–273.
9. Taenzer AH, Blike GT. Postoperative monitoring - the Dartmouth experience. Anesth Patient Saf Found Newsl. 2012;27:1, 3–4, 21.
10. Taenzer AH, Pyke JB, McGrath SP, Blike GT. Impact of pulse oximetry surveillance on rescue events and intensive care unit transfers: a before-and-after concurrence study. Anesthesiology. 2010;112:282–287.
11. Fu ES, Downs JB, Schweiger JW, Miguel RV, Smith RA. Supplemental oxygen impairs detection of hypoventilation by pulse oximetry. Chest. 2004;126:1552–1558.
12. Practice Guidelines for the Prevention, Detection, and Management of Respiratory Depression Associated with Neuraxial Opioid Administration: An Updated Report by the American Society of Anesthesiologists Task Force on Neuraxial Opioids and the American Society of Regional Anesthesia and Pain Medicine. Anesthesiology. 2016; 124:535–552.
13. Anesthesia Patient Safety Foundation. Available at: Accessed September 14, 2017.
14. Bailey PL, Rhondeau S, Schafer PG, et al. Dose-response pharmacology of intrathecal morphine in human volunteers. Anesthesiology. 1993;79:49–59.
15. McGrath SP, Pyke J, Taenzer AH. Assessment of continuous acoustic respiratory rate monitoring as an addition to a pulse oximetry-based patient surveillance system. J Clin Monit Comput. 2017;31:561–569.
16. Berry RB, Budhiraja R, Gottlieb DJ, et al. Rules for scoring respiratory events in sleep: update of the 2007 AASM Manual for the Scoring of Sleep and Associated Events. J Clin Sleep Med.2012;15:597–619.
17. Sirian R, Wills J. Physiology of apnoea and the benefits of preoxygenation. Contin Educ Anaesth Crit Care Pain. 2009;9:105–108.
18. Elmoaqet H, Tilbury DM, Ramachandran SK. Evaluating predictions of critical oxygen desaturation events. Physiol Meas. 2014;35:639–655.
19. Davis DP, Hwang JQ, Dunford JV. Rate of decline in oxygen saturation at various pulse oximetry values with prehospital rapid sequence intubation. Prehosp Emerg Care. 2008;12:46–51.
20. West JB. Pulmonary Pathophysiology: The Essentials. 2012:8th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 56–76.
21. Anderson TW, Darling DA. Asymptotic theory of certain “Goodness of Fit” criteria based on stochastic processes. Ann Math Stat.1952;23:193–212.
22. Cohen J. Statistical Power Analysis for the Behavioral Sciences1988.2nd ed. Hillsdale, NJ: Lawrence Earlbaum Associates.

    Supplemental Digital Content

    Back to Top | Article Outline
    Copyright © 2017 International Anesthesia Research Society