Aims. This study aims to assess first, whether mutations in the epidermal growth factor receptor (EGFR) and Kirsten rat sarcoma (kRAS) genes are associated with overall survival (OS) in patients who present with symptomatic bone metastases from non-small cell lung cancer (NSCLC) and secondly, whether mutation status should be incorporated into prognostic models that are used when deciding on the appropriate palliative treatment for symptomatic bone metastases. Patients and Methods. We studied 139 patients with NSCLC treated between 2007 and 2014 for symptomatic bone metastases and whose mutation status was known. The association between mutation status and overall survival was analysed and the results applied to a recently published prognostic model to determine whether including the mutation status would improve its discriminatory power. Results. The median OS was 3.9 months (95% confidence interval (CI) 2.1 to 5.7). Patients with EGFR (15%) or kRAS mutations (34%) had a median OS of 17.3 months (95% CI 12.7 to 22.0) and 1.8 months (95% CI 1.0 to 2.7), respectively. Compared with EGFR-positive patients, EGFR-negative patients had a 2.5 times higher risk of death (95% CI 1.5 to 4.2). Incorporating EGFR mutation status in the prognostic model improved its discriminatory power. Conclusion. Survival prediction models for patients with symptomatic bone metastases are used to determine the most appropriate (surgical) treatment for painful or fractured lesions. This study shows that NSCLC should not be regarded as a single entity in such models. Cite this article: Bone Joint J 2017;99-B:516–21