Objective. To study health care utilization in veterans with SpAs. Methods. In a postal survey of 70 508 veterans in Veterans Integrated Service Network (VISN)-13 from 1 October 1996 to 31 March 1998, demographics, smoking status and performance of activities of daily living (ADLs) were queried. Databases provided ICD-9 codes for AS, PsA and ReA; comorbidities; demographics; and health care utilization post-survey in respondents. Multivariable linear/logistic regression compared out- and inpatient health care utilization in SpA vs non-SpA, and its predictors in SpA. Results. A total of 1001 veteran respondents had diagnoses of SpA: AS, n = 154; PsA, n = 814; ReA, n = 33. Veterans with AS, PsA and ReA, respectively, had significantly higher adjusted annual medical specialty (2.8, 3.6 and 3 vs 1.5; P = 0.0001), surgical care (3.3, 2.7 and 3.2 vs 1.9; P < 0.0001) and primary care visits (3.4, 3.0 and 2.3 vs 2.7, P = 0.024). Multivariable-adjusted analyses showed that more ADL limitations and higher comorbidity were associated with higher in- and outpatient health care utilization in PsA and none of the predictors were significantly associated with utilization in AS. Conclusions. After adjustment for differences in demographics and comorbidities, more outpatient health resource utilization was observed in SpA patients. Further studies should focus on what leads to this increase in utilization, and whether any modifiable factors can be introduced to reduce health care utilization in PsA patients. © The Author 2009. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved.