Background: The Karnofsky Performance Score (KPS) is an important prognostic indicator in cancer care. Additionally, patient-reported outcomes are increasingly incorporated into clinical care and are validated in cancer patients with the NIH’s Patient-Reported Outcomes Measurement Information System (PROMIS). However, little is known about the concordance of the provider-rated KPS and patient-rated PROMIS measures. Methods: Retrospective review of patients with advanced cancer who underwent implantation of an intrathecal pump for refractory pain. We compared KPS and PROMIS scores recorded within 1 week of each other. PROMIS scores included an average score and 3 domains: physical function, fatigue, and pain interference. We divided each measure into 3 previously-described categories (Functionally Independent, Needs Limited Assistance, and Functionally Dependent) and assessed categorical concordance within each measure. The association of patient-related demographic and clinical characteristics were analyzed with KPS–PROMIS concordance using chi-square and Mann-Whitney U tests. Results: 47 patients were included. The provider-rated KPS score consistently indicated a higher functional status compared to the patient-rated PROMIS average score and across all PROMIS domain scores. 40% of patients were rated “Functionally Dependent” using the average PROMIS score, compared to just 19.1% using the KPS, a statistically significant difference (p=0.042). We noted categorical concordance of 55.3% for PROMIS average score, 57.5% for physical function, 44.7% for fatigue and 42.7% for pain interference compared to KPS. We found no statistically significant associations between concordance and patient demographics or characteristics. Conclusions: Our results revealed a notably low concordance in functional status in advanced cancer patients between provider-rated KPS and patient-rated PROMIS measures. Provider-rated scores consistently indicated a higher overall performance status, in particular with measures of the lowest functional status. Additional work is needed to better understand the reasons for this lack of concordance and its implications for ongoing clinical decision-making.