Description of Abnormal Breathing Is Associated With Improved Outcomes and Delayed Telephone Cardiopulmonary Resuscitation Instructions
Background Emergency 9‐1‐1 callers use a wide range of terms to describe abnormal breathing in persons with out‐of‐hospital cardiac arrest (OHCA). These breathing descriptors can obstruct the telephone cardiopulmonary resuscitation (CPR) process.
Methods and Results We conducted an observational study of emergency call audio recordings linked to confirmed OHCAs in a statewide Utstein‐style database. Breathing descriptors fell into 1 of 8 groups (eg, gasping, snoring). We divided the study population into groups with and without descriptors for abnormal breathing to investigate the impact of these descriptors on patient outcomes and telephone CPR process. Callers used descriptors in 459 of 2411 cases (19.0%) between October 1, 2010, and December 31, 2014. Survival outcome was better when the caller used a breathing descriptor (19.6% versus 8.8%, P<0.0001), with an odds ratio of 1.63 (95% confidence interval, 1.17–2.25). After exclusions, 379 of 459 cases were eligible for process analysis. When callers described abnormal breathing, the rates of telecommunicator OHCA recognition, CPR instruction, and telephone CPR were lower than when callers did not use a breathing descriptor (79.7% versus 93.0%, P<0.0001; 65.4% versus 72.5%, P=0.0078; and 60.2% versus 66.9%, P=0.0123, respectively). The time interval between call receipt and OHCA recognition was longer when the caller used a breathing descriptor (118.5 versus 73.5 seconds, P<0.0001).
Conclusions Descriptors of abnormal breathing are associated with improved outcomes but also with delays in the identification of OHCA. Familiarizing telecommunicators with these descriptors may improve the telephone CPR process including OHCA recognition for patients with increased probability of survival.
What Is New?
Our study found that abnormal breathing descriptions by emergency 9‐1‐1 callers are independently associated with delays in telephone cardiopulmonary resuscitation provision of up to 45 seconds, but these out‐of‐hospital cardiac arrest patients were more likely to survive with favorable functional outcome.
What Are the Clinical Implications?
Our study emphasizes that early telecommunicator identification of abnormal breathing descriptions could help improve outcomes among persons experiencing out‐of‐hospital cardiac arrest.
Out‐of‐hospital cardiac arrest (OHCA) is an enormous healthcare problem in industrialized countries.1, 2, 3 More than 90% of OHCA patients die before reaching the hospital,1 and survival rates remain disappointing even after hospital arrival.4, 5, 6 Bystander cardiopulmonary resuscitation (BCPR) can more than double the chance of patient survival,7 but BCPR rates remain between 30% and 40%3, 4 in most communities. Emergency medical telecommunicators play a key role in the period before arrival of emergency medical services.8, 9 Telephone CPR (TCPR), in which telecommunicators guide callers in performing CPR, can double the frequency of BCPR,10 and recent guidelines emphasize the importance of TCPR for increasing rates and timeliness of BCPR.11 Many factors, however, can obstruct recognition of OHCA. A prime factor that can prevent or delay recognition of cardiac arrest is abnormal breathing, which presents frequently in the early stages of OHCA.12, 13, 14, 15, 16, 17 Callers use various terms to describe this abnormal breathing.13, 17, 18 These descriptors can confuse telecommunicators and prevent OHCA recognition.17, 19, 20 In this population‐based study, we evaluated patient outcomes when the caller used or did not use a breathing descriptor and audited OHCA audio recordings to identify the impact of caller descriptors of abnormal breathing on TCPR process measures.
Study Design and Population
We conducted an observational study of audio recordings from emergency 9‐1‐1 calls linked to confirmed OHCAs in an Utstein‐style database between October 1, 2010, and December 31, 2014. The calls were received at 8 regional dispatch centers participating in the Save Hearts in Arizona Registry and Education (SHARE) program, a collaboration of the Arizona Department of Health Services, the University of Arizona, and fire departments, police departments, and hospitals statewide. This program and its results have been reported previously.21, 22, 23, 24, 25, 26 Because OHCA has been designated a major public health problem in Arizona and the objective of SHARE is to improve resuscitation quality and patient outcomes, the data collected were exempt from the Health Insurance Portability and Accountability Act requirements. The Arizona Department of Health Services’ human subjects review board and the University of Arizona institutional review board approved publication of deidentified data.
The study population consisted of OHCAs of nontraumatic origin that were treated but not witnessed by emergency 9‐1‐1 responders. Audio recordings linked to these events were reviewed by program personnel according to a standardized process described previously.27 The terms callers used to describe abnormal breathing were recorded in a structured database. We excluded cardiac arrests (1) if they occurred in nursing homes, doctor's offices, jails, or unknown locations, to focus on how non–healthcare professionals would describe abnormal breathing; (2) if CPR was not indicated in the recording, because SHARE personnel could not confirm that the patient was not conscious and not breathing or not breathing normally; (3) if a language barrier obstructed the dispatch process; or (4) if the caller was not with the patient. For the TCPR process analysis, we excluded calls in which (1) bystander CPR was started before telecommunicator instructions and (2) the audio was incomplete or fragmented.
The TCPR protocol in the dispatch centers specifies (1) compression‐only CPR for adult arrests of presumed cardiac origin and (2) chest compression with rescue breathing for other causes of arrest. Emergency telecommunicators are expected to provide CPR instructions if the patient was reported as not conscious and not breathing normally. Roughly 80% of the audios investigated were from dispatch centers in Maricopa County that drafted their own protocols. The remainder were from dispatch centers outside Maricopa County using various versions of Medical Priority Dispatch or Association of Public Safety Communications Officials systems.
Measurements and Analysis
We divided the study population into 2 groups. The first group had at least 1 caller descriptor of abnormal breathing (Description YES group). The second group had no caller descriptors of abnormal breathing; the caller simply said the patient was “not breathing” or answered “no” when the telecommunicator asked if the patient was breathing normally (Description NO group). We investigated basic event characteristics such as patient age, sex, event location, witness status, BCPR status, shockable rhythm status, whether the patient had sustained return of spontaneous circulation, whether the patient survived, and whether the patient survived with good functional neurological outcome.
We considered TCPR to be provided if telecommunicators started CPR instructions that resulted in the start of bystander compressions. To determine whether caller descriptors effected TCPR process measures, we compared findings across 6 metrics between the 2 groups: (1) percentage of calls in which the telecommunicator recognized the need for TCPR, (2) percentage of calls in which the telecommunicator started TCPR instructions, (3) percentage of calls in which bystanders started TCPR, (4) time interval from call receipt until the telecommunicator recognized the need for TCPR, (5) interval from call receipt until the start of TCPR instructions, and (6) interval from call receipt until the bystander performed the first chest compression.
Continuous variables were summarized by median and range and were compared between the 2 groups of patients using the Wilcoxon rank sum test. Categorical variables were summarized by frequency and proportion with 95% Clopper–Pearson confidence intervals (CIs) and were compared between the 2 groups by either χ2 or Fisher exact test. The association between survival and description of abnormal breathing was also examined by logistic regression, adjusting for important risk adjusters and potential confounders including sex, age, location of arrest, presumed etiology of arrest, bystander‐witnessed arrest, bystander CPR, shockable initial rhythm, and whether the patient was transported to a cardiac receiving center. The effect of the continuous variable age was fitted nonparametrically using penalized thin plate regression splines through the generalized additive model.28 The backward elimination process was conducted to remove covariates from the model with a P value threshold of 0.05 while always keeping sex and age in the model. The process was then repeated to study the association between favorable functional outcome and description of abnormal breathing. In the process analysis, the proportions of calls with telecommunicator recognition of the need for TCPR, with TCPR instructions given or with TCPR started, were compared between the 2 groups of patients. To compare time to telecommunicator recognition of the need for TCPR, to start of TCPR instructions, and to first compression, the generalized log‐rank test29 was used to compare interval‐ and right‐censored data. The software environment R30 and the R packages interval31 and Icens32 were used for the time‐to‐event analysis. All tests were 2‐sided with α=0.05. No adjustment for multiple testing was made.
We reviewed 4298 audio recordings linked to OHCAs of nontraumatic origin that were treated but not witnessed by EMS. Among these, we excluded cases (1) at medical facilities, jails, or unknown locations (n=841); (2) in which CPR was not indicated (n=688); (3) with language barriers (n=57); and (4) in which callers were not with the patient (n=61). Overall, 2411 were eligible for the outcomes analysis (Description YES group, 459; Description NO group, 1952) and 1841 were eligible for the TCPR process analysis (Description YES group, 379; Description NO group, 1462; Figure 1). A summary of OHCA characteristics along with clinical outcomes in the 2 groups is shown in Table 1. The proportions of men in each group were not significantly different in the Description YES and NO groups (68.8% versus 64.2%, respectively; P=0.071). The 2 groups had similar age profiles, with a median of 62 years and range from 0 to 101 years in the combined sample. Patients in the YES group were more likely to be in a public location (14.4% versus 10.7%, P=0.032) and to have a witnessed arrest (49.2% versus 27.8%, P<0.0001) but were less likely to receive BCPR (50.3% versus 56.3%, P=0.022). Patients in the YES group were more likely to have a shockable initial rhythm (42% versus 19.1%, P<0.0001); to achieve sustained return of spontaneous circulation (27.5% versus 15.2%, P<0.0001); to be transported to a cardiac receiving center26 that could provide therapeutic hypothermia, prompt percutaneous coronary interventions, and other guideline‐based postarrest critical care (66.4% versus 56.1%, P=0.0001); to survive to discharge (19.6% versus 8.8%, P<0.0001); and to have favorable functional outcomes (14.8% versus 5.6%, P<0.0001). The logistic regression model analysis showed that the unadjusted odds ratio for description of abnormal breathing was 2.59 (95% CI, 1.95–3.46; P<0.0001) for the survival outcome and 3.06 (95% CI, 2.19–4.28; P<0.0001) for the favorable functional outcome (data are not shown). The adjusted odds ratio for description of abnormal breathing was 1.63 (95% CI, 1.17–2.25; P=0.003) for the survival outcome and 1.68 (95% CI, 1.15–2.46; P=0.008) for the favorable neurological outcome (Table. 2).
Figure 2 shows the frequencies of the 8 most common descriptors of abnormal breathing. The 3 categories with highest proportions were “gasping” (32%; 95% CI, 27.8–36.5), “snoring” or “snorting” (20.7%; 95% CI, 17.1–24.7) and “gurgling” (17%; 95% CI, 13.7–20.7). These were followed by “moaning” or “groaning,” “labored,” “heavy,” and “noisy,” each of which occurred in frequency <10%.
Results of the process analysis are shown in Tables 3 and 4. Table 3 summarizes the proportion of cases with each of 3 process events: (1) The telecommunicator recognized the OHCA, (2) CPR instructions were started, and (3) bystander chest compressions were started. The rate of OHCA recognition was lower in the YES group (79.7% versus 93.0%, P<0.0001), as were the rates of CPR instructions started and chest compressions started (65.4% versus 72.5% [P=0.0078] and 60.2% versus 66.9% [P=0.0123], respectively). The time between call receipt and OHCA recognition was longer in the YES group (118.5 versus 73.5 seconds, P<0.0001), as were the times to start of CPR instructions (203.5 versus 155.5 seconds, P<0.0001) and to start of chest compressions (242 versus 197.5 seconds, P<0.0005). A sensitivity analysis excluding cases in which callers could not get patients into position for CPR, had difficult access to patients, were in a dangerous environment, were severely distressed, refused CPR instructions, hung up or left the phone, were physically unable to do CPR, or thought the patient was dead demonstrated significant intergroup differences across all process measures and thus was consistent with the main process analysis (Tables S1 and S2).
The Description YES group comprised 19% of eligible calls. Callers used various terms to describe abnormal breathing. The terms “gasping,” “snoring,” or “snorting,” and “gurgling” accounted for roughly 70% of cases with descriptions. Consistent with previous studies' findings of improved survival among patients with agonal breathing, we found that patients for whom abnormal breathing was described had a higher chance of survival13, 15 than patients for whom it was not described. They also had a higher chance of favorable functional outcome. The rate of OHCA recognition was lower in the YES group and may in part explain the reduced proportion of cases in which CPR instructions and bystander compressions were started in this group. In addition, the time to OHCA recognition was 45 seconds longer in the YES group (118.5 versus 73.5 seconds, P<0.0001). This may in part explain the longer time to start of compressions in this group.
These findings are collective evidence that abnormal breathing descriptions can be a barrier to optimal patient outcomes. To overcome this barrier, telecommunicators need to be familiar with the ways in which callers describe abnormal breathing in OHCA patients. Education and training that highlight these descriptions could shorten the time to OHCA identification, increase TCPR provision, and enhance outcomes for patients with abnormal breathing.
A study that reviewed OHCA dispatch recordings in the United States showed that “barely breathing” was the most frequent description, followed by “heavy or labored breathing” and “problems in breathing.”13 However, this study evaluated a small number of cases, and 2 other studies of caller descriptions were non–English‐language reports.17, 18
It was reported that these callers' descriptions for abnormal breathing can interfere with the TCPR process because callers tend to perceive abnormal breathing as a sign of life.17, 18 Hauff et al reviewed 404 OHCA audio recordings and reported that the TCPR process was impeded in 51 cases in which callers described signs of life such as “breathing.” Even if telecommunicators identify cardiac arrest when callers describe abnormal breathing, the whole TCPR process can be obstructed.22 Our study adds to previous work in quantifying the time to recognition for cases in which callers described abnormal breathing. Because the probability for survival decreases by roughly 7% to 10% every minute BCPR is not performed,33 the 45‐second delay in recognition reported can be of great consequence in the period before arrival of emergency medical services.
Although the presence of descriptions appears to obstruct OHCA recognition, education and training on caller descriptors could turn these obstacles into opportunities to identify OHCA more quickly and comprehensively. Detailed regional analyses of callers' descriptions for abnormal breathing can be applied to local training and education to enhance the provision of TCPR across communities. Further studies across populations can help identify any universal descriptors that could be linked to survival and favorable functional outcomes of persons experiencing sudden cardiac arrest.
This study has limitations. First, we excluded 841 cases occurring in medical facilities, jails, and unknown locations in an effort to limit our catalog to terms lay rescuers use to describe abnormal breathing. Ultimately, however, the characteristics of emergency 9‐1‐1 callers are extremely difficult to assess, and we cannot rule out the possibility that some callers had medical backgrounds. Second, descriptions of abnormal breathing and their impact on telecommunicator instruction for CPR will vary by language and culture from one region to another; therefore, the applicability of our results to other countries and emergency medical services systems is unknown. Third, other unidentified barriers could also have affected our TCPR process measurements. Fourth, the majority of emergency 9‐1‐1 calls were received at dispatch centers in Maricopa County that draft their own protocols. This may limit the degree to which we can generalize of our findings. Finally, despite using a structured evaluation technique and data format, evaluating the TCPR process is not an exact science and requires some level of subjectivity and interpretation.
In this statewide study, we found that the identification of cardiac arrest and the start of CPR can be obstructed when the caller used a breathing descriptor. Caller descriptions of abnormal breathing in OHCA patients are associated with improved survival and functional outcome. Familiarizing telecommunicators with the most common descriptors may enhance cardiac arrest recognition, shorten the time to starting CPR, and improve patient outcomes.
Drs. Bobrow and Spaite disclose that the University of Arizona received funding from the Medtronic Foundation through the HeartRescue Grant to support community‐based translation of resuscitation science.
Table S1. Analysis of Process Measures: Proportion of Patients With Certain Events
Table S2. Process Analysis: Time to Events After Excluding Cases With Barriers to Telephone Cardiopulmonary Resuscitation
- ↵Graham R, MaCoy MA, Schultz AM, eds. Committee in the Treatment of Cardiac Arrest: Current Status and Future Directions, Board on Health Science Policy, Institute of Medicine. Strategies to Improve Cardiac Arrest Survival: A Time to Act. Washington, DC: National Academies Press; 2015.
- ↵Kitamura T, Iwami T, Kawamura T, Nitta M, Nagao K, Nonogi H, Yonemoto N, Kimura T. Nationwide improvements in survival from out‐of‐hospital cardiac arrest in Japan. Circulation. 2012;126:2834–2843.
- ↵Wissenberg M, Lippert FK, Folke F, Weeke P, Hansen CM, Christensen EF, Jans H, Hansen PA, Lang‐Jensen T, Olesen JB, Lindhardsen J, Fosbol EL, Nielsen SL, Gislason GH, Kober L, Torp‐Pedersen C. Association of national initiatives to improve cardiac arrest management with rates of bystander intervention and patient survival after out‐of‐hospital cardiac arrest. JAMA. 2013;310:1377–1384.
- ↵Bougouin W, Lamhaut L, Marijon E, Jost D, Dumas F, Deye N, Beganton F, Empana JP, Chazelle E, Cariou A, Jouven X. Characteristics and prognosis of sudden cardiac death in Greater Paris: population‐based approach from the Paris Sudden Death Expertise Center (Paris‐SDEC). Intensive Care Med. 2014;40:846–854.
- ↵Ro YS, Shin SD, Kitamura T, Lee EJ, Kajino K, Song KJ, Nishiyama C, Kong SY, Sakai T, Nishiuchi T, Hayashi Y, Iwami T; Seoul‐Osaka Resuscitation Study G . Temporal trends in out‐of‐hospital cardiac arrest survival outcomes between two metropolitan communities: Seoul‐Osaka resuscitation study. BMJ Open. 2015;5:e007626.
- ↵Sasson C, Rogers MA, Dahl J, Kellermann AL. Predictors of survival from out‐of‐hospital cardiac arrest: a systematic review and meta‐analysis. Circ Cardiovasc Qual Outcomes. 2010;3:63–81.
- ↵Lerner EB, Rea TD, Bobrow BJ, Acker JE III., Berg RA, Brooks SC, Cone DC, Gay M, Gent LM, Mears G, Nadkarni VM, O'Connor RE, Potts J, Sayre MR, Swor RA, Travers AH. Emergency medical service dispatch cardiopulmonary resuscitation prearrival instructions to improve survival from out‐of‐hospital cardiac arrest: a scientific statement from the American Heart Association. Circulation. 2012;125:648–655.
- ↵Rea TD, Eisenberg MS, Becker LJ, Murray JA, Hearne T. Temporal trends in sudden cardiac arrest: a 25‐year emergency medical services perspective. Circulation. 2003;107:2780–2785.
- ↵Kronick SL, Kurz MC, Lin S, Edelson DP, Berg RA, Billi JE, Cabanas JG, Cone DC, Diercks DB, Foster JJ, Meeks RA, Travers AH, Welsford M. Part 4: system of care and continuous quality improvement: 2015 American Heart Association guideline update for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation. 2015;132:s397–s413.
- ↵Bobrow BJ, Zuercher M, Ewy GA, Clark L, Chikani V, Donahue D, Sanders AB, Hilwig RW, Berg RA, Kern KB. Gasping during cardiac arrest in humans is frequent and associated with improved survival. Circulation. 2008;118:2550–2554.
- ↵Fukushima H, Imanishi M, Iwami T, Seki T, Kawai Y, Norimoto K, Urisono Y, Hata M, Nishio K, Saeki K, Kurumatani N, Okuchi K. Abnormal breathing of sudden cardiac arrest victims described by laypersons and its association with emergency medical service dispatcher‐assisted cardiopulmonary resuscitation instruction. Emerg Med J. 2015;32:314–317.
- ↵Bobrow BJ, Spaite DW, Berg RA, Stolz U, Sanders AB, Kern KB, Vadeboncoeur TF, Clark LL, Gallagher JV, Stapczynski JS, LoVecchio F, Mullins TJ, Humble WO, Ewy GA. Chest compression‐only CPR by lay rescuers and survival from out‐of‐hospital cardiac arrest. JAMA. 2010;304:1447–1454.
- ↵Bobrow BJ, Vadeboncoeur TF, Stolz U, Silver AE, Tobin JM, Crawford SA, Mason TK, Schirmer J, Smith GA, Spaite DW. The influence of scenario‐based training and real‐time audiovisual feedback on out‐of‐hospital cardiopulmonary resuscitation quality and survival from out‐of‐hospital cardiac arrest. Ann Emerg Med. 2013;62:47e1–56.e1.
- ↵Spaite DW, Bobrow BJ, Stolz U, Berg RA, Sanders AB, Kern KB, Chikani V, Humble W, Mullins T, Stapczynski JS, Ewy GA; Arizona Cardiac Receiving Center C . Statewide regionalization of postarrest care for out‐of‐hospital cardiac arrest: association with survival and neurologic outcome. Ann Emerg Med. 2014;64:496–506.e1.
- ↵Dameff C, Vadeboncoeur T, Tully J, Panczyk M, Dunham A, Murphy R, Stolz U, Chikani V, Spaite D, Bobrow B. A standardized template for measuring and reporting telephone pre‐arrival cardiopulmonary resuscitation instructions. Resuscitation. 2014;85:869–873.
- ↵Wood SN. Generalized Additive Models: An Introduction With R. Boca Raton, FL: Chapman & Hall/CRC; 2006.
- ↵R Core Team . R: A Language and Environment for Statistical Computing [Computer Program]. Vienna, Austria: R Foundation for Statistical Computing; 2015.
- ↵Gentlemen R, Vandal A. Icens: NPMLE for Censored and Truncated Data [computer program]. Version R package 1.42.0.