To determine if patient ethnicity among patients with a hip fracture influences the type of fracture, surgical care, and outcome. This was an observational cohort study using a linked dataset combining data from the National Hip Fracture Database and Hospital Episode Statistics in England and Wales. Patients' odds of dying at one year were modelled using logistic regression with adjustment for ethnicity and clinically relevant covariates. A total of 563,640 patients were included between 1 April 2011 and 1 October 2020. Of these, 476,469 (85%) had a coded ethnicity for analysis. Non-white patients tended to be younger (mean 81.2 vs 83.0 years), and were more commonly male (34.9% vs 28.5%; p < 0.001). They were less likely to be admitted from institutional care (12.9% vs 21.8%; p < 0.001), to have normal cognition (53.3% vs 62.0%; p < 0.001), and to be free of comorbidities (22.0% vs 26.8%; p < 0.001), but were more likely to be from the most deprived areas (29.4% vs 17.3%; p < 0.001). Non-white patients were more likely to experience delay to surgery for medical reasons (14.8% vs 12.7%; p < 0.001), more likely to be treated with an intramedullary nail or a sliding hip screw (52.5% vs 45.1%; p < 0.001), and less likely to be mobilized by the day after surgery (74.3% vs 79.0%; p < 0.001). Mortality was higher among non-white inpatients (9.2% vs 8.4% for white), but was lower at one year after hip fracture (26.6% vs 30.3%). Our study identified serious problems with the quality of NHS ethnicity data. Despite this, we have shown that there is complex variation in case-mix and hip fracture morphology between ethnic groups. We have also identified variations in care received between ethnic groups, and that disparities in healthcare may contribute to poorer outcome. Inpatient mortality is higher in non-white patients, although this finding reverses after discharge, and one-year mortality is lower. Patients from ethnic minority backgrounds experience disparities in healthcare, however at one year postoperatively non-white patients have a statistically significant lower mortality after controlling for relevant clinical covariates.