Should time from cardiac arrest until call to emergency medical services (EMS) be collected in EMS research?

Because of the difficulty and cost of collecting the time from collapse to placing the 911 call in instances of out-of-hospital cardiac arrest and because of the potential noise and bias that might be inherent in such data, a simulation study was conducted to quantitate the impact that such data might have on estimates of the relationship between time from collapse to defibrillation and probability of survival. In the absence of bias, an underestimate of the slope on the order of 20% to 30% might be expected. However, in the presence of bias, the impact on the slope estimate is unpredictable. The most likely bias would tend to cause an overestimate of the slope. It is suggested that unless the time from collapse to placing the 911 call can be obtained accurately and without bias, it is probably not worthwhile to do so.