Back pain is a common illness and chiropractors provide a large proportion of back pain care in the United States. This is the first study to systematically compare chiropractic patients with those who saw other providers for back pain. The authors analyzed data from the RAND Health Insurance Experiment, a community-based study of the use of health services. Insurance claims forms were examined for all visits specified by the patient as occurring for back pain. Visits were grouped into episodes using decision rules and clinical judgment. The primary provider of back pain care was defined as the provider who delivered most of the services. Sociodemographic and health status and attitudes variables of patients were examined for association with the choice of chiropractor. Multivariate logistic regression models were constructed to calculate adjusted odds ratios for independent predictors. There were 1020 episodes of back pain care made by 686 different persons and encompassing 8825 visits. Results indicated that chiropractors were the primary provider for 40% of episodes, and retained as primary provider a greater percentage of their patients (92%) who had a second episode of back pain care than did medical doctors. Health insurance experiment site, white race, male sex, and high school education were independent predictors of choosing a chiropractor. Conclusions suggested that chiropractors were the choice of one third of all patients who sought back pain care, and provided care for 40% of all episodes of care. Geographic site, education, gender, and income were independent patient factors predicting chiropractic use.