What are spikes?
Spikes are quality control measurements performed in order to assess bias, or in other words how closely a device's measurement is to the true or correct value. A spike is a radon measuring device that has been exposed in a chamber to a known and controlled radon concentration. The environmental parameters of relative humidity and temperature should also be known and controlled. A spike can be either a passive or active device, but usually the term "spike" refers to a passive device (alpha-track, charcoal or electret ion chamber device). Passive devices are not calibrated individually, but instead the manufacturer or analysis laboratory spikes sets of like devices and develops calibration curves or equations to apply to field devices. It is only through analyzing spiked samples periodically can it be demonstrated that the calibration curves or equations continue to be appropriate for the devices that are actually being deployed in the field.
What is a "blind" spike?
If the set of conditions to which the devices were exposed are not divulged to the entity performing the analysis prior to reporting of results, then these are called "blind" spikes. This is done for "proficiency" or "performance" tests to meet the requirements of certification programs or the licensing requirements of some regulated states, in which case the end user reports the results to the chamber facility. Also, end users who do not perform analyses may have the chamber conditions prior to the analysis, but may send the devices to the analysis laboratory in a blind manner.
How often must I do spikes?
EPA protocols (Protocols for Radon and Radon Decay Product Measurements in Homes, EPA 402-R-92-003, June, 1993; AARST has adopted this document as an Interim Standard) require that spikes be performed at the rate of 3%, with a minimum of three per year and a maximum of six per month. The intent of the protocols is that this pertains to passive devices only. This means that for every 100 passive devices that you use for field tests, you must send three devices to a radon chamber facility for spiking. This type of spiking is typically not done in a blind manner, but can be if the end user requests it be done that way. If you perform fifty real estate tests that are each duplicated, then you would use 100 devices and therefore you would need to do three spikes.
A user of a high volume of passive devices, such as a large charcoal analysis laboratory, would likely analyze at least 100 devices per month and therefore would need to spike the maximum of six devices per month to meet the requirements of the protocols. The user may wish to spike more than six devices per month, but only six per month are required by the protocols.
Users of lower volumes of passive devices should review with the chamber facility personnel their rate of usage and the facility's fee schedule to determine the best and most economic manner of performing the spikes. For users of moderate numbers of devices, it might make sense both technically and economically to perform spikes on a quarterly or semiannual basis.
For users of 100 or less devices in a year, three spikes must be done sometime during the year. The protocols do not state whether they must be done one at a time throughout the year, or whether they can be done all at once. From a technical view, it is desirable to assess the precision of the spiked devices as well as the bias, and for this purpose it is better to spike all three devices at the same time. Further, the chamber facility may charge a minimum base fee which would make it better economically to spike all three devices at once. However, the end user may wish to do them one at a time, separated by a few months, in order to have data throughout the year.
Since the analysis laboratory does spikes, why do I have to do them, too?
Ultimately, it is you that is reasonable for demonstrating that the results that you give to your clients are accurate. The quality control measures that the analysis laboratory performs are part of its QA program and are used to show that the laboratory is in control. However, just because the laboratory is in control does not necessarily mean that the end user is also in control. The laboratory has no control over the devices once they are sent to you. Therefore, you must also perform spikes to demonstrate that your handling and storing methods are not creating some problem that is affecting the response of the detectors. For example, charcoal devices that are stored in a damp area for long periods of time may not read accurately when finally deployed.
I read my own Eperms, do I need to do spikes?
In the case of electret ion chamber devices, the end user is often also the analysis laboratory, and thus spiking is an important part of the QA program. Just as with charcoal devices and alpha-track devices, the supplying laboratory performs calibrations and spikes on like devices but has no control over the devices once they are sent to you. In this case, however, not only can problems be introduced by the end user during storage and handling, but also during the analysis of the devices, that could affect the response of the devices. For example, if extreme care is not taken to keep the ion chambers and electrets absolutely clean, dust or dirt particles can discharge the electrets and produce false high measurements. Further, if the electret reader is not kept at the same temperature when reading the electret both before and after deployment, then this could affect the measurement. It is only through spikes of devices that you actually use and analyze can you ensure that your measurements are accurate.
Why would someone spike a continuous radon monitor?
The protocols do not require end users to spike continuous radon monitors; there are other requirements for monitors, such as annual calibrations and crosschecks. However, an end user that is applying for or renewing a certification with a credentialing organization, or a license in a regulated state, may be required to submit a continuous monitor for a "proficiency" or "performance" test. In this case, the monitor is sent to a qualified radon chamber facility where it is spiked in a blind manner and returned to the end user for analysis and reporting of results. Also, manufacturers of continuous radon monitors typically perform spiking for the purposes of research and development and for periodic checks to ensure that their calibration is correct.
Ok, I've done the spikes, and I have the results, but what do I do with them now?
Getting the devices spiked and analyzed is a good start, but it's not the whole job. The measurement result of each device must be compared with the value of radon concentration to which the device was exposed, or the "target value." This value is supplied by the radon chamber personnel.
First, the Relative Percent Error (RPE) is calculated for each device using the following equation:
RPE = [( MV - TV) / TV] x 100%
Where RPE = Relative percent error (%),
MV = Measurement value (pCi/L), and
TV = Target value (pCi/L) from the chamber.
Note that if the measurement value is less than the chamber's target value, the value of the RPE is negative.
The values of RPE should be tabulated and plotted on a control chart. A complete discussion of control charts is beyond the scope here. The protocol document referenced above (EPA 402-R-92-003) contains in Appendix B a discussion of precision control charts. A bias control chart can be constructed similar to Exhibit B-1 in that document, but with the Relative Percent Error (RPE) on the Y-axis. Because RPE can be both positive and negative, control limit lines must be placed both above and below zero on the Y-axis. Recommended values to use are: "In Control Level" at ±10%, "Warning Level" at ±20% and "Control Limits" at ±30%. The goal is for the overall average bias (average RPE) to be zero, and 95% of the RPE values to be within the limits of ±20%.
There are several Continuing Education courses available that discuss quality control measurements and control charting in more detail.
If you'd like more information about how we can help you with your radon laboratory needs, please let us know.
|
|
|
