To determine the association of a neurologist visit with headache health care utilization and costs.Methods
Utilizing a large privately insured health care claims database, we identified patients with an incident headache diagnosis (ICD-9 codes 339.xx, 784.0x, 306.81) with at least 5 years follow-up. Patients with a subsequent neurologist visit were matched to controls without a neurologist visit using propensity score matching, accounting for 54 potential confounders and regional variation in neurologist density. Co–primary outcomes were emergency department (ED) visits and hospitalizations for headache. Secondary outcomes were quality measures (abortive, prophylactic, and opioid prescriptions) and costs (total, headache-related, and non-headache-related). Generalized estimating equations assessed differences in longitudinal outcomes between cases and controls.Results
We identified 28,585 cases and 57,170 controls. ED visits did not differ between cases and controls (p = 0.05). Hospitalizations were more common in cases in year 0–1 (0.2%, 95% confidence interval [CI] 0.2%–0.3% vs 0.01%, 95% CI 0.01%–0.02%; p < 0.01), with minimal differences in subsequent years. Costs (including non-headache-related costs) and high-quality and low-quality medication utilization were higher in cases in the first year and decreased toward control costs in subsequent years with small differences persisting over 5 years. Opioid prescriptions increased over time in both cases and controls.Conclusion
Compared with those without a neurologist, headache patients who visit neurologists had a transient increase in hospitalizations, but the same ED utilization. Confounding by severity is the most likely explanation given the non-headache-related cost trajectory. Claims-based risk adjustment will likely underestimate disease severity of headache patients seen by neurologists.