OBJECTIVE: This study analyzed socioeconomic and clinical factors relating to both physicians and patients associated with physicians' choice of the treatment for sleep difficulties in a nationally representative sample of outpatient physician visits in the US. METHODS: A modified version of the model suggested by Eisenberg was used as a theoretical framework for this study to predict the factors influencing treatment of sleep difficulties. A multivariate logistic regression method was used to analyze the 1996-2001 National Ambulatory Medical Care Survey data to examine physician and patient related predictors of treatment variations for sleep difficulties RESULTS: From 1996 to 2001, about 4.8 billion visits were made to outpatient physician offices in the US, and 94.6 million of these were sleep-difficulty related visits. This study found that 32% of patients with sleep difficulties received no type of therapy during their visits and 5% of the patients received behavioral therapy only. Psychiatrist visits were 72% more likely (OR: 1.72, 95% CI: 1.08-2.61) to be associated with receipt of medication therapy and ten times more likely (OR: 10.19, 95% CI: 4.85-14.44) to be associated with behavioral therapy prescription only than visits to family practitioners and internists. Patient visits with public insurance as a primary payer source were more likely to be associated with benzodiazepine prescription among patients receiving at least some medication therapy (OR: 1.66, 95% CI: 1.13-2.45) than patient visits with private insurance as a primary payer source. CONCLUSIONS: The results of this study indicate that several patient and physician characteristics influence physician prescribing of treatments for sleep difficulties. This study suggests a need to develop better care management guidelines for sleep difficulties and a wider coverage of behavioral therapy in the US. The study also finds variations in quality of care and treatment for sleep difficulties in outpatient settings in the US.