OBJECTIVES: Studies have indicated that approximately 80 to 95 percent of gabapentin use is off-label.  However, evidence in support of these off label uses is limited.  There is an increasing body of research on the efficacy of gabapentin for treating pain and mental conditions, but much of this work (especially that for mental conditions) is preliminary and suggestive at best. The objective of this study is to examine, using a nationally representative sample, the efficacy of gabapentin in treating the wide range of off-label conditions for which there is little or no evidence and the few conditions for which there is already some evidence.  METHODS: The study uses the 1997-2007 Medical Expenditure Panel Data set.  Users of gabapentin were identified and the top 20 ICD9 Codes, by age group, associated with prescriptions for gabapentin were used to select control subjects.  Sub-samples associated with pain, neurologic conditions, mental health, and other off-label conditions were created.  Ordinary least squares and propensity score matching following the regression method attributed to Hirano and Imbens were used to examine the association of gabapentin and self-reported health (physical and mental health)  with controls for lagged health, age, gender, race, income, education, insurance, and co-morbidities. RESULTS: Use of gabapentin was negatively associated with reporting of good physical or mental health in almost all cases.   The coefficients in the propensity score controlled models tended to be smaller but still suggested a significant relationship between gabapentin use and a 5-10% lower probability of reporting good physical or mental health.   CONCLUSIONS: There was little evidence that the use of gabapentin is associated with better self-reported physical or mental health and lower medical spending.  There is a strong case for improved efforts to promote evidence based use of medications as there is a potential for significant expenditures with little or no impact on health.