Uncertainty of calibrations at the accredited dosimetry calibration laboratories.

The American Association of Physicists in Medicine, through a subcommittee (formerly Task Group 3) of the Radiation Therapy Committee, has accredited five laboratories to perform calibrations of instruments used to calibrate therapeutic radiation beams. The role of the accredited dosimetry calibration laboratories (ADCLs) is to transfer a calibration factor from an instrument calibrated by the National Institute of Standards and Technology (NIST) to a customer's instrument. It is of importance to the subcommittee, to physicists using the services of the ADCLs, and to the ADCLs themselves, to know the uncertainty of instrument calibrations. The calibration uncertainty has been analyzed by asking the laboratories to provide information about their calibration procedures. Estimates of uncertainty by two procedures were requested: Type A are uncertainties derived as the standard deviations of repeated measurements, while type B are estimates of uncertainties obtained by other methods, again expressed as standard deviations. Data have been received describing the uncertainty of each parameter involved in calibrations, including those associated with measurements of charge, exposure time, and air density, among others. These figures were combined with the uncertainty of NIST calibrations, to arrive at an overall uncertainty which is expressed at the two-standard deviation level. For cable-connected instruments in gamma-ray and x-ray beams of HVL > 1 mm Al, the figure has an upper bound of approximately 1.2%.