In-Hospital Defibrillation

Determining time to defibrillation (abstract)

July 3, 1996

Time has been on the minds of resuscitation researchers recently–specifically, the problem of obtaining accurate time intervals in the treatment of cardiac arrest. Accurate determination of time intervals to major interventions is clearly important to resuscitation research in both pre-hospital and in-hospital settings. The success of resuscitation is obviously time-dependent, and the impact of elapsed time on the effectiveness of various treatment options is important in determining treatment sequences and in cost/benefit analyses of measures intended to shorten response times.

In the hospital setting, obtaining accurate time data is important for an additional reason. Delays in defibrillation are inherent in the prevalent approach to cardiac resuscitation in hospitals–particularly for unmonitored arrests.  No matter how obvious to some, these delays (and their contribution to dismal in-hospital survival rates) can and will be denied until good data are presented showing that time intervals from arrest to defibrillation are unacceptably long. A reliable and valid timing method could do much to overcome this obstacle to improving survival from in-hospital cardiac arrests.

I will review some of the major obstacles to obtaining good time interval data in cardiac arrests, with particular emphasis on in-hospital arrests. I will go on to outline an idea for a simple, relatively inexpensive device for automatically timing intervals to defibrillation. This device could be used with any hospital's current equipment and could also be adapted--with greater expense and design time--to improve data gathering in the out-of-hospital setting.

To obtain reliable and valid time data in a code, data-gathering must be automatic, simply because human observers are unreliable in codes. The highly stressful environment of an unexpected cardiac arrest can reasonably be compared to the "fog of battle," in which "a great part of the information obtained...is contradictory, a still greater part is false, and by far the greatest part is of a doubtful character."

The time of arrest (or time of discovery) cannot be assumed to be noted accurately by a human observer, simply because witnesses to an arrest always have higher priorities than timekeeping. Intervention times could conceivably be recorded by a code team member starting shortly after the arrest, but those with much experience in codes know that a person directly involved in emergency care cannot simultaneously keep an accurate record of events.

If no one can be relied upon to mark accurate times in a code, it follows that an automatic device is highly desirable to record the essential treatment times. Full automation is essential: merely making recording easier–for example, by requiring a button to be pushed to mark the time a treatment is given–is inadequate, because in the "fog of battle" any nonessential recording actions are likely to be delayed (producing inaccurate data) or omitted entirely. It is also important, for obvious reasons, that a timing method be independent of specific treatments or devices being investigated.

The typical hospital protocol requires the "discovery" nurse to call the code and call for the crash cart while staying with the patient in order to perform basic CPR (though performing one-person CPR is problematic at best). Whether this consistently happens is unclear--frequently, the nurse may run to the nurses' station in order to call the code and fetch the crash cart. With either scenario, however, the crash cart is likely to be mobilized around the same time that the code is paged by the central operator, and often before.

Based on the author's clinical experience and a review of the literature, the most valid feasible starting point for timing an in-hospital code is the time that the crash cart is mobilized. This starting point is as good as, and possibly better than, the time when the code is paged. It would be relatively simple to design an external device, useable with all existing equipment, which would automatically start a clock when the defibrillator (or cart) is unplugged for transport to the site of the arrest.

This clock could also easily detect and record automatically the time of delivery of countershocks . The clock could display elapsed time or clock time or both (or neither, if used solely for research). Ideally, time to first monitoring would also be recorded automatically.

A big question, of course, is whether this start time would be valid and reliable. I contend it would be at least as good as any other feasible starting point (the obvious one being the time that the code is paged). I base this in part on my knowledge of formal hospital protocols, but more importantly on my experience of what really happens in codes on general units. It is very difficult to say with assurance what typically happens in codes, but there are a number of reasons to believe that the crash cart is consistently brought to the scene very early in the code.

Validity of the recorded time intervals could be verified by correlating the intervals with incidence of V-fib vs. asystole, or possibly with amplitude of V-fib. If defibrillators (manual or automatic) are placed in locations separate from the unit's crash cart, it will be important to have timers on both the defibrillator and the crash cart and adjust the start time to the earlier of the two times in order to get the most valid approximation of time interval from arrest to defibrillation.

With such a device, data from cardiac monitoring units could be used to determine a survival curve plotting survival rates against time to defibrillation. This would require synchronizing the device clock with the clock of the central monitor, which could be done manually once a week or so or (better) automatically by radio synchronization.

A greater engineering challenge would be involved in adapting this idea to the out-or-hospital setting. Since times from the 911 call to arrival of the emergency vehicle at the scene are typically recorded reliably, the difficulty comes in determining the time interval from curbside arrival to first shock (or first monitoring). An external device using piezoelectric touch sensors and/or motion detectors could be used to start a clock when the defibrillator is taken from the emergency vehicle.

Web Pages Created With AOLpress