Karnofsky Performance Scale: Clinical Purpose and Application
Primary Purpose
The Karnofsky Performance Scale (KPS) is an 11-level ordered scale (0-100) that assesses cancer patients' functional status based on activity level, work capacity, and self-care abilities, serving as both a prognostic indicator and a critical tool for treatment decision-making. 1
Scale Structure and Scoring
KPS ranges from 0 to 100 in increments of 10, with higher scores indicating better functional status and lower scores (particularly <60) associated with poor survival and serious illness 1
The scale evaluates three core domains: activity level, ability to work, and self-care capacity 1
KPS demonstrates moderately high interrater reliability (0.97) when used by trained observers, making it a reproducible assessment tool 2
Critical Clinical Applications
Treatment Eligibility Determination
Patients with KPS <60 should receive best supportive care only, without systemic chemotherapy 1
Patients with KPS ≥60 are candidates for systemic therapy with or without best supportive care, or enrollment in clinical trials 1
Three-drug chemotherapy regimens should be reserved exclusively for patients with good performance status (KPS ≥80), while two-drug regimens are preferred for those with lower scores 1
Prognostic Value
KPS accurately predicts early death in cancer patients, with low scores strongly correlating with shorter survival times 3
The scale demonstrates predictive validity for survival (correlation coefficient r = 0.30 in terminal cancer patients), though high initial scores do not guarantee long survival 3, 2
Rapidly declining KPS scores over serial assessments predict death within months, providing valuable prognostic information for care planning 3
Integration with Other Assessment Tools
KPS is incorporated into the Palliative Prognostic Score (PaP) alongside other clinical factors to predict 30-day survival probability 4
KPS correlates strongly with ECOG Performance Status (Spearman R = -0.869), with a reliable conversion: KPS 80-100 = ECOG 0-1; KPS 60-70 = ECOG 2; KPS <60 = ECOG 3-4 5
KPS can be mapped to the Functional Independence Measure for rehabilitation planning, helping identify cancer patients who would benefit from inpatient or outpatient rehabilitation services 6
Important Clinical Caveats
Assessment Limitations
KPS is inherently subjective, which can affect inter-rater reliability and introduce potential bias, particularly for borderline scores 1
Clinicians tend to assign higher (worse) KPS scores to patients aged >65 compared to younger patients, despite no objective difference in physical activity, representing age-related bias 1
KPS is less predictive of cancer-related outcomes in older adults compared to younger populations 1
The setting of evaluation matters: patients evaluated at home typically receive lower KPS scores compared to clinic-based assessments performed at the same time 3
Etiology Considerations
- Current KPS scoring does not differentiate the underlying cause of poor performance status (disease burden vs. comorbidities vs. treatment toxicity), which is clinically relevant since patients with disease-related poor performance may improve with effective treatment 1
Comparison with Alternative Scales
ECOG Performance Status demonstrates superior discriminatory ability compared to KPS in predicting patient prognosis and should be preferred when choosing between the two scales 5
The Palliative Performance Scale (PPS) expands beyond KPS by incorporating additional domains including ambulation, self-care, and oral intake, providing more comprehensive assessment in palliative care settings 4
Practical Implementation
Formal training in KPS assessment significantly improves reliability, with trained observers achieving interrater reliability of 0.97 compared to untrained assessors 2
Serial KPS assessments are more valuable than single measurements for tracking disease trajectory and predicting outcomes 3
KPS should be documented before, during, and after treatment to monitor functional changes and guide ongoing management decisions 5