Also known as ECOG Performance Status, ECOG PS, Zubrod score, performance status
Clinician-assigned functional capability grade, 0 to 4.
Clinically reviewed byJennifer J. Wu, MDNot an endorsement of any individual result, and no substitute for the treating clinician’s judgement.
Reviewed Oct 27, 2026Version 1.0.0Next review Oct 27, 2027
Applies to / doesn’t apply to5
Applies to: Adults with cancer. The scale was developed and is used in oncology, and it is the instrument trial protocols and prognostic models in that field are written against.
- Pediatric patients — a separate performance scale exists for children, and this one was not developed on them.
- Non-oncology disease. This is outside the population the scale was developed in: no validation study in another disease area could be sourced, and equally no publication declares it invalid there.
- As a determination of clinical-trial or treatment eligibility. Eligibility thresholds are protocol rules written against the grade, not the clinical meaning of the grade. The FDA’s 2026 guidance states that patients with ECOG PS 2 should be included unless there is a scientific or clinical rationale for exclusion.
- As a substitute for assessing comorbidity or independence in older adults. The FDA’s 2026 guidance states plainly that existing performance-status scales are suboptimal for most patients with cancer aged 65 or older.
- To derive a Karnofsky score. Several mappings circulate and ECOG-ACRIN describes the one it publishes as “one commonly used comparison”; transforming an individual patient’s grade carries a high level of error (Verger 1992).
Record the performance the patient is capable of, not what they happen to be doing — voluntary rest does not itself lower the grade.
What this score estimates
Records which of five described levels of functional capability a clinician judges a patient with cancer to be at, on the scale published by the Eastern Cooperative Oncology Group in 1982. There is no formula: the clinician selects one level and its grade, 0 to 4, is the result. The full scale also includes grade 5 (dead), which isn’t offered here. Higher grades indicate progressively greater restriction in work, ambulation and self-care. It describes observable capability at a single point in time — it is not a prognosis, not a measure of disease severity, and not a determination of eligibility for treatment or for a clinical trial.
Pearls & pitfalls
23
- The central limitation is that the same patient is assigned different grades by different assessors. Published inter-observer agreement spans kappa 0.19 to 0.914 across studies using different rating procedures, observers and case mixes. These cannot be reconciled into a single estimate, so no one figure should be read as the reliability of the scale. For context, one review found Karnofsky had better inter-rater reliability than ECOG in four studies.
- Grade 2 is the least reliably assigned grade — kappa 0.31, against 0.55, 0.48, 0.43 and 0.33 for grades 0, 1, 3 and 4, with all observers agreeing in only 40 of 100 cases (Sørensen 1993). It is also the grade most eligibility thresholds are written against.
- Clinicians rate function better than patients do. Across 1,636 physician–patient pairs the grades disagreed in 56.6% of cases, weighted kappa 0.35, and where they disagreed the physician rated the patient better 75.1% of the time — a mean grade of 0.91 against the patient’s 1.30 (Schnadig 2008). The direction reverses at the poor-function end, where patients with poor performance rated themselves better. Corroborated at kappa 0.41–0.53 in three further cohorts.
- A precision worth stating, because the opposite is often quoted: in that same study, whether clinician and patient agreed about ECOG grade was not significantly associated with survival. The “disagreement predicts death” findings are Karnofsky and nutritional-status results, not ECOG results.
- Assignment varies with who is asking and why. Palliative-care specialists rated patients 0.5 grades worse and nurses 1.0 grade worse than oncologists, both p<0.0001, and discordance was 2.4 times more likely when a treatment was available. That association is documented; any inference about deliberate up-coding is not.
- Two studies independently found women are rated worse — a vignette study of 72 clinicians, and Blagden 2003.
- The descriptors do not match measured behaviour. With accelerometry, 80% of patients graded 0 or 1 spent more than half their waking hours resting, and the correlation between measured activity and assigned grade was rho −0.26 (Broderick 2014).
- It misses what matters in older adults. Among 363 patients aged 65 or over, 74% graded below 2 — yet of those with “good” performance status, 37.7% had limitations in instrumental activities of daily living and 13.0% had two or more comorbidities (Repetto 2002).
- Age influences the assigned grade independently of measured activity, but the direction is unresolved: the FDA’s 2026 guidance states that clinicians assign higher scores to patients aged 65 or over, while Broderick 2014 reports lower. That paper could not be obtained, so the association is stated here without a direction.
- The FDA characterises the determination itself as inherently subjective, which can affect inter-rater reliability and invite potential bias, particularly for patients at the borderline between categories — and notes that current scales do not differentiate between the causes of a low performance status.
- No general table of outcomes by grade is displayable, and this is a gap in the evidence rather than in the search. Neither defining publication reports outcomes by grade, and trial populations are filtered: of 279 FDA-approval trials, 207 capped enrolment at performance status 1 or better, with median enrolment of patients at grade 2 or worse of 4.3%.
- A claim to be wary of in the secondary literature: Cheng 2017 is repeatedly cited, including in JCO, as showing no difference in clinical outcomes between grades 0, 1 and 2. It shows no difference in relative treatment-effect hazard ratios (0.65 against 0.67, p=0.68) — a different claim from no difference in absolute outcomes.
- The six-point scale is often credited to Zubrod 1960. That paper, read in full, defines five levels on a flow sheet and contains no grade 5; the descriptors in modern use are Oken 1982’s.
What the grade does not capture:
- Comorbidity
- Independence in instrumental activities of daily living
- Frailty
- Cognition
- Nutritional status
- Symptom burden, including pain and nausea — graded separately on Zubrod’s original flow sheet
- The cause of a low performance status — cancer, treatment toxicity, or an unrelated condition
- Change over time, or the rate of change
- The patient’s own assessment of their function, which differs systematically from the clinician’s
- Any laboratory value, vital sign, age or sex
What the guidelines say
3
States that performance-status determination is inherently subjective, which can affect inter-rater reliability and invite potential bias, particularly for patients at the borderline between categories, and that existing scales are suboptimal for most patients with cancer aged 65 or older. States that patients with ECOG PS 2 should be included in trials unless there is a scientific or clinical rationale for exclusion.
View guideline source ↗The International Prognostic Index for aggressive non-Hodgkin lymphoma uses a performance status of 2 or above as one of its five model factors — one of several instruments that take this grade as an input.
View guideline source ↗States that the scale circulates in the public domain, and that several ways exist to map the ECOG and Karnofsky scales — describing the comparison it publishes as one commonly used version rather than the definitive one.
View guideline source ↗Guideline positions are stated as published. Curie.MD Calc does not publish treatment direction.
Where this comes from
9
References
- Oken MM, Creech RH, Tormey DC, Horton J, Davis TE, McFadden ET, Carbone PP. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol. 1982;5(6):649-655. View source ↗
- ECOG-ACRIN Cancer Research Group. ECOG Performance Status Scale. Accessed 2026-09-17. View source ↗
- Zubrod CG, Schneiderman M, Frei E, et al. Appraisal of methods for the study of chemotherapy of cancer in man. J Chronic Dis. 1960;11(1):7-33. View source ↗
- Sørensen JB, Klee M, Palshof T, Hansen HH. Performance status assessment in cancer patients. An inter-observer variability study. Br J Cancer. 1993;67(4):773-5. View source ↗
- Schnadig ID, Fromme EK, Loprinzi CL, et al. Patient-physician disagreement regarding performance status is associated with worse survivorship in patients with advanced cancer. Cancer. 2008;113(8):2205-14. View source ↗
- Broderick JM, Hussey J, Kennedy MJ, O’Donnell DM. Patients over 65 years are assigned lower ECOG PS scores than younger patients, although objectively measured physical activity is no different. J Geriatr Oncol. 2014;5(1):49-56. View source ↗
- Repetto L, Fratino L, Audisio RA, et al. Comprehensive Geriatric Assessment Adds Information to ECOG Performance Status in Elderly Cancer Patients. J Clin Oncol. 2002;20(2):494-502. View source ↗
- US Food and Drug Administration. Performance status in oncology clinical trials. Guidance for industry. July 2026. View source ↗
- Verger E, Salamero M, Conill C. Can Karnofsky performance status be transformed to the Eastern Cooperative Oncology Group scoring scale and vice versa? Eur J Cancer. 1992;28(8-9):1328-30. View source ↗