Spinal devices/instrumentation are used to augment the stability of a decompressed spinal segment during surgery. Like any other mechanical component, the device can fail. A standard in vitro test protocol, was developed to determine load vs number of cycles to failure curve for a pedicle screw-plate/rod type spinal device. The protocol based on the use of an 'artificial spine' model, is clinically relevant. The protocol was used to characterize the load-carrying capacities and failure modes of a specific pedicle screw-rod type fixation device to demonstrate its appropriateness. The devices (Kaneda) were tested in the quasi-static as well as fatigue bending modes. In the bending fatigue mode, the devices failed at loads significantly smaller than the corresponding quasi-static failure load magnitude (806 N). The device exhibited an endurance limit in the fatigue bending mode. The device is not likely to exhibit failure if subjected to cyclic loads which cause less than 380 N axial compression (and an accompanying bending moment relative to the device of less than 13.57 Nm). The failures observed in specimens subjected to the fatigue tests ranged from complete to partial breakage of the paraspinal rods as opposed to failure due to permanent deformation (yielding) of the rods in the quasi-static bending test specimens. The protocol developed can be used for any other screw-plate/rod type spinal instrumentation. The use of a standard protocol by researchers would enable a comparison of various devices currently available in the market. Such comparative data would be useful for the scientific community, and agencies such as the FDA and ASTM.(ABSTRACT TRUNCATED AT 250 WORDS)