a b s t r a c t

Objective: Out-of-hospital cardiac arrest (OHCA) is a leading cause of mortality in the United States. We sought to evaluate the accuracy of the patient care report (PCR) for detection of 2 clinically important events: return of spontaneous circulation (ROSC) and rearrest (RA).

Methods: We used defibrillator recordings and PCRs for Emergency Medical Services-treated OHCA collected by the Resuscitation Outcomes Consortium‘s Pittsburgh site from 2006 to 2008 and 2011 to 2012. Defibrillator data included electrocardiogram rhythm tracing, chest compression measurement, and audio voice recording. Sensi- tivity analysis was performed by comparing the accuracy of the PCR to detect the presence and number of ROSC and RA events to integrated defibrillator data.

Results: In the 158 OHCA cases, there were 163 ROSC events and 53 RA events. The sensitivity of PCRs to identify all ROSC events was 85% (confidence interval [CI], .795-.905); to identify primary ROSC events, it was 85% (CI,

.793-.907); and to identify secondary ROSC events, it was 78% (CI, .565-.995). The sensitivity of PCRs to identify the presence of all RA events was .60 (CI, .469-.731); to identify primary RA events, it was 71% (CI, .578-.842); and to identify secondary RA events, it was 0. Of the 32 RA incidents captured by the PCR, only 15 (47%) correctly identified the correct lethal arrhythmia.

Conclusions: We found that PCRs are not a reliable source of information for assessing the presence of ROSC and post-RA electrocardiogram rhythm. For quality control and research purposes, medical providers should consider augmenting data collection with continuous defibrillator recordings before making any conclusions about the oc- currence of critical resuscitation events.

(C) 2014

Introduction

Out-of-hospital cardiac arrest (OHCA) is a leading cause of mortality in the United States [1]. Patients who are treated with cardiopulmonary resuscitation (CPR) for cardiac arrest may achieve return of spontane- ous circulation (ROSC), which is characterized by the return of pulses after CPR and/or electrical defibrillation. However, patients may often lose pulses after ROSC, a clinically significant event termed rearrest (RA). Two independent studies have previously shown that RA occurs in approximately 38% to 39% of cases and that RA is negatively associat- ed with survival [2,3]. When ROSC occurs, it is very important for med- ical providers to detect the event without delay so that ongoing chest compressions do not occur while the patient has a pulse, and transport to the hospital can be expedited. The timely detection of RA is also crit- ical to minimizing any absence of blood flow by resuming chest com- pressions as soon as RA occurs. Reducing the time to identify RA is

? Prior presentations: none.

?? Funding sources/disclosures: 1R21HL10444-01A1, 2U01HL077863.

* Corresponding author. 3600 Forbes Avenue, Suite 400A, Pittsburgh, PA 15261.

E-mail address: [email protected] (M.L. Sundermann).

also a good indicator of adequate CPR quality, as unwanted pauses be- fore an intervention are avoided.

Medical providers in the out-of-hospital setting detect RA through the electrocardiogram (ECG) by identifying nonpulsatile rhythms such as ventricular fibrillation, ventricular tachycardia, pulseless electrical activity, or asystole as well as a physical check for pulse presence by pal- pating an artery. When patient care is transferred to the receiving hos- pital, medical providers that treated the patient then complete a patient care report (PCR). The PCR is a descriptive document that provides a narrative of patient assessment, medical interventions, and patient out- comes. Patient care reports are retained in EMS systems with the pur- pose of keeping a formal medical record of the prehospital patient care. Medical directors have the ability to use the PCR to assess for ROSC and RA presence on a case review basis for both quality improvement (QI) and research purposes. It is important that QI assessments for ROSC and RA in the PCR be as accurate as possible because the ability to accurately check for the presence of pulses is one of the most critical skills of the medical provider; return of pulses in the OHCA patient is vital for survival. Any deficiencies in ROSC and RA ascertainment by medical providers must be addressed in a timely manner by medical directors, and minimal error in the PCR is necessary to streamline the

QI process.

http://dx.doi.org/10.1016/j.ajem.2014.10.037

0735-6757/(C) 2014

ECG

compression force

Fig. 1. Custom graphical user interface of sample ECG and chest compression data.

We sought to evaluate the accuracy of the PCR in the detection of 2 critical resuscitation events, ROSC and RA, and to specifically compare it to the capabilities of ECG and other signals recorded on the defibrilla- tor monitor, which we considered to be the gold standard determina- tion. We expected to observe a considerable amount of discrepancy between the PCRs and defibrillator recordings.

Methods

Study design

The study was approved by the University of Pittsburgh Institutional Review Board. The study was designed as a case review of EMS-treated OHCA patients where narrative descriptions of cardiac arrest obtained

in PCRs were compared to multichannel defibrillator data. Both PCR and defibrillator data were evaluated for the presence of ROSC and RA events.

Study setting and population

Case data for nontraumatic, EMS-treated OHCA were obtained from the Pittsburgh site of the Resuscitation Outcomes Consortium for the peri- od spanning January 2006 to December 2008 and January 2011 to Decem- ber 2012. Cardiopulmonary resuscitation process variables needed to be present for inclusion into the analysis, which meant that a full recording of the defibrillator data needed to have existed. The years 2009-2010 were excluded from the analyses because of an ongoing clinical trial dur- ing which data were embargoed (Resuscitation Outcomes Consortium prehospital resuscitation using an Impedance valve and Early versus Delayed (ROC PRIMED)).

Fig. 2. ROSC and RA classification technique from defibrillator data.

Fig. 3. ROSC sensitivity ascertainment from defibrillator and PCR data.

Measures

Defibrillator data for all cardiac arrests were downloaded and ana- lyzed in a custom MATLAB software that time-coded the signal traces and displayed them on a graphical user interface. A sample tracing of the graphical user interface can be seen in Fig. 1. Defibrillator recordings were obtained from a Philips MRX monitor (Phillips Healthcare, Andover, MA). To confirm ROSC, the defibrillator signal needed to have an audio recording verbalizing pulses or an ECG conversion to a normal sinus rhythm along with a greater-than-10-second cessation of chest compressions. Rearrest was determined by one or all of the following: paramedic verbalization of pulse absence on the audio recording or presence of ventricular fibrillation, ventricular tachycardia, or asystole along with a greater-than-10-second presence of chest compressions. Rearrest due to pulseless electrical activity was required to have audio confirmation of a “loss of pulses.” This approach is outlined in Fig. 2. Chest compression presence was determined by assessing for the pres- ence of a rhythmic signal in either the impedance signal trace measured by the defibrillation pads or a force sensor that was placed on all pa- tients within the study. Primary ROSC and RA were defined as the first event, whereas secondary ROSC and RA were defined as any successive event that followed the first event.

Data analysis

After the signal data were marked for ROSC or RA, the corresponding PCR was read to compare findings of presence of ROSC or RA. The narra- tive component consisting of a written description of the cardiac arrest call, as well as temporal listing of the interventions, was analyzed and read in the PCR. Sensitivities and 95% confidence intervals (CIs) were calculated using Stata software (Stata Statistical Software: Release 12; Stata, College Station, TX).

Results

We processed 1094 OHCA cases in our database. Of these cases, 158 had ROSC and all signals available. In these cases, there were 163 in- stances of ROSC and 53 instances of RA (34%). The sensitivity of PCRs to identify the presence of all ROSC events was .85 (CI, .795-.905); to identify primary ROSC events, it was .85 (CI, .793-.907); and to identify secondary ROSC events, it was .78 (CI, .565-.995). The sensitivity of PCRs to identify the presence of all RA events was .60 (CI, .469-.731); to identify primary RA events, it was .71 (CI, .578-.842); and to identify secondary RA events, it was 0 (no secondary RA events were captured). Of the 32 RA incidents captured by the PCR, 15 (47%) correctly identified

Fig. 4. RA sensitivity ascertainment from defibrillator and PCR data.

Fig. 5. RA rhythm capture by PCR and defibrillator data.

the correct lethal arrhythmia. The PCR detection outcomes are shown in Figs. 3 and 4. If RA was identified by the PCR, the rhythm documented was assessed against the defibrillator ECG data as shown in Fig. 5.

Discussion

In the OHCA cases that we analyzed, PCRs were insufficient in cap- turing ROSC and RA events. We speculate that this result is due to health care provider’s difficulty in recalling specific, temporal events when completing the PCR after the cardiac arrest resuscitation and/or patient transfer at the hospital has occurred. These findings are supportive of recent evidence of imprecise documentation in time-dependent events in OHCA [4]. Cardiac arrest calls can be especially hectic; and the after- the-fact narrative, when not supplemented with additional physiologic signals, may be an inaccurate tool when characterizing critical resuscita- tion events. Cardiac arrests with multiple ROSC and RA events may be especially difficult to recall, and we have shown evidence of this in that secondary RA events were never captured in the PCR report. Inac- curacy in reporting the post-RA ECG rhythm reflects the lack of texture that may be present in PCR data as well.

To provide context to our PCR sensitivity findings, we conducted a literature search of past studies investigating OHCA that reported ECG characteristics and outcomes. Search terms within Pubmed that ranged from 1980 to 2013 included patient care report, cardiac arrest documentation, out-of-hospital cardiac arrest, and resuscitation. Of the 17 studies we examined, 14 reported findings from PCRs alone [1,5-17], whereas 3 extracted their data by using a combina- tion of signal recordings and PCRs [18-20]. It should be noted how- ever that it was unclear if the “PCR only” studies included ECG recordings in their reports (many EMS systems do not). Therefore, to generalize our findings, it would have to be assumed that PCRs consisted of the narrative component only and excluded any printed or recorded ECG strips. It is also possible that ECGs may be in the PCR but not necessarily used.

We found that the ability to capture RA events was less sensitive than the ability to capture ROSC, especially secondary RA events. Successive RA events may be difficult for the medical provider to re- member after the cardiac arrest call has finished. In addition to the poor detection capabilities of the presence of RA, the PCR identified the correct lethal RA arrhythmia only 15 out of the 32 incidences of

RA. The signal record consisting of at least the ECG recording should be assessed before making ROSC/RA judgments solely based on the PCR.

Limitations

Limitations of the study include the small sample size as well as its retrospective nature. Patient care reports in other EMS systems may in fact include time-oriented resuscitation signals such as ECG, which would be more sensitive to ROSC or RA. The study was also limited in that only ROSC and RA were assessed. Perhaps other details of cardiac arrest treatment are documented more effectively in PCRs. The litera- ture search we conducted shows that determining whether these sig- nals exist is most times ambiguous.

Conclusions

We found PCRs to be an inaccurate source of information for assessing the presence of ROSC and RA. Relying on PCRs to accurately capture and identify an RA event is not as effective as defibrillator data. Those using narrative prehospital PCRs for quality improvement and research purposes should consider the limitations of this lone data source and should consider including ECG, audio, and compression data from the cardiac monitor.

Acknowledgments

City of Pittsburgh Emergency Medical Service.

References

1. Nichol G, Thomas E, Callaway CW, Hedges J, Powell J, Aufderheide T, et al. Regional variation in out-of-hospital cardiac arrest incidence and outcome. JAMA 2008;300 (12):1423-31.
2. Salcido DD, Stephenson AM, Condle JP, Callaway CW, Menegazzi JJ. Incidence of re- arrest after return of spontaneous circulation in out-of-hospital cardiac arrest. Prehosp Emerg Care 2010;14:413-8.
3. Brooke E, Lerner EB, O”Connell M, Pirrallo RG. Rearrest after prehospital resuscita- tion. Prehosp Emerg Care 2011;15(1):50-4.
4. Frisch A, Reynolds JC, Condle J, Gruen D, Callaway CW. Documentation discrepancies of time-dependent critical events in out of hospital cardiac arrest. Resuscitation 2014;85(8):111-4.
5. Atkins G, Everson-Stewart S, Sears G, Daya M, Osmond M, Warden C, et al. Epidemiology and outcomes from out-of-hospital cardiac arrest in children. Circulation 2009;119: 1484-91.
6. Garza A, Gratton M, Salomone J, Lindholm D, McElroy J, Archer R. Improved patient survival using a modified Resuscitation protocol for out-of-hospital cardiac arrest. Circulation 2009;119:2587-605.
7. Kuisma M, Maatta T. Out-of-hospital cardiac arrests in Helsinki: Utstein style reporting. Heart 1996;76:18-23.
8. Holmberg M, Holmberg S, Herlitz J. Effect of bystander cardiopulmonary resuscitation in out-of-hospital cardiac arrest patients in Sweden. Resuscitation 2000;47(1):59-70.
9. Eckstein M, Stratton S, Chan L. Cardiac arrest resuscitation evaluation in Los Angeles: CARE-LA. Ann Emerg Med 2005;45(5):504-9.
10. Kellum M, Kennedy K, Ewy G. Cardiocerebral resuscitation improves survival of pa- tients with out-of-hospital cardiac arrest. Am J Med 2006;119:335-40.
11. McNally B, Robb R, Mehta M, Vellano K, Valderrama A, Yoon P, et al. Out-of-hospital cardiac arrest surveillance-Cardiac Arrest Registry to Enhance Survival (CARES). MMWR Morb Mortal Wkly Rep 2011;60(SS08):1-19.
12. Sirbaugh P, Pepe P, Shook J, Kimball K, Goldman M, Ward M, et al. A prospective, population-based study of the demographics, epidemiology, management, and out- come of out-of-hospital pediatric cardiopulmonary arrest. Ann Emerg Med 1999;33 (2):174-84.
13. Finn J, Jacobs I, Holman C, Oxer H. Outcomes of out-of-hospital cardiac arrest pa- tients in Perth, Western Australia, 1996-1999. Resuscitation 2001;51(3):247-55.
14. Holmberg M, Holmberg S, Herlitz J. Incidence, duration and survival of ventricular fibrilla- tion in out-of-hospital cardiac arrest patients in Sweden. Resuscitation 2000;44(1):7-17.
15. de Vreede-Swagemakers I, Gorgels A, Dubois-Arbouw W, van Ree J, Daemen M, Houben L, et al. Out-of-hospital cardiac arrest in the 1990s: a population-based study in the Maastricht area on incidence, characteristics and survival. J Am Coll Cardiol 1997;30(6):1500-5.
16. Fairbanks R, Shah M, Lerner EB, Ilangovan K, Pennington E, Schneider S. Epidemiology and outcomes of out-of-hospital cardiac arrest in Rochester, New York. Resuscitation 2007;72:415-24.
17. Eschmann M, Pirrallo R, Aufderheide T, Lerner EB. The association between emer- gency medical services staffing patterns and out-of-hospital Cardiac arrest survival. Prehosp Emerg Care 2010;14:71-7.
18. Jacobs I, Finn J, Oxer H, Jelinek G. CPR before defibrillation in out-of-hospital cardiac arrest: a randomized trial. Emerg Med Australas 2005;17(1):39-45.
19. Wilk L, Kramer-Johansen J, Myklebust H. Quality of cardiopulmonary resuscitation during out-of-hospital cardiac arrest. JAMA 2005;293(3):299-304.
20. Iwami R, Kawamura T, Hiraide A, Berg R, Hayashi Y, Nishiuchi T, et al. Effectiveness of bystander-initiated cardiac-only resuscitation for patients with out-of-hospital cardiac arrest. Circulation 2007;116:2900-7.