This article describes the preparation and internal and external evaluation of materials, critical issues in the external quality assessment (EQA) of point-of-care testing (POCT) devices for measuring blood glucose. A comparison was made between different materials, both of natural and synthetic origin and with and without stabilisers. The aims were to produce a material which was compatible with as many POCT-devices as possible and so reduce the number of materials sent out in each campaign as well as to optimise the precision and comparability of results between methods and devices. Although the use of near natural material--sterile-filtered plasma spiked with glucose--survived internal testing, this material proved to be unsuitable for EQA surveys. The study resulted in the reduction of materials for each survey to stabilised whole blood for one device, stabilised plasma for two devices and a synthetic material based on a polyethylene glycol matrix for all other devices. Samples were sent as pairs six times annually. The POCT-devices tested measured precisely but inaccurately in the synthetic material, when compared with the reference method (gas-chromatography coupled with isotope-dilution mass-spectrometry; GC-IDMS), so that the devices could only be evaluated for precision. The construction of ratios between the concentrations measured on the two samples distributed allowed an indirect assessment of accuracy. The need for surveillance of POCT devices is stressed in this publication, which combines theory and practice in setting up and running an EQA programme for blood glucose.