Pain and deception: use of verbal pain measurement as a diagnostic aid in differentiating between clinical and simulated low-back pain.

The pain reports of 553 patients with low back pain and 347 healthy subjects faking low back pain were compared to determine if pain language can be employed as diagnostic aids in differentiating between clinical groups and groups attempting to feign a clinical condition. The best set of discriminating words correctly identified 90% of the subjects using 18 to 54 words in pain description, and 83% using 4 to 17 words in their pain report. The same pain words were used to categorize 366 new pain cases and 202 new fake cases. Cross-validation shrinkage was 8 and 7% respectively. Effectiveness in discriminating between pain and simulation was increased using selective cut-off scores. Patients in active treatment for back pain whose choice of pain words resemble those of the simulators reported 21% more clinical pain in a follow-up validation study suggesting that an accurate representation of malingering may have been achieved using a simulation research design.