Measuring health in patients with cervical and lumbosacral spinal disorders: is the 12-item short-form health survey a valid alternative for the 36-item short-form health survey?