What Is the WHO Performance Status Scale?

The WHO Performance Status scale is a simple five-point scoring system used by oncologists to rate how well a cancer patient can carry out everyday activities. Also known as the ECOG Performance Status scale (after the Eastern Cooperative Oncology Group that developed it), the tool assigns a score from 0 (fully active with no restrictions) to 4 (completely bedridden and unable to perform any self-care). Despite its apparent simplicity, this single number carries enormous weight: it influences which treatments you are offered, whether you qualify for a clinical trial, and even how long you are expected to survive.

Where the Scale Came From

Before the WHO/ECOG scale existed, oncologists relied on the Karnofsky Performance Status (KPS) scale, introduced in 1948 for patients receiving chemotherapy for lung cancer. KPS uses a 100-point ladder in increments of ten, with 100 representing perfect health and 0 representing death. It works, but the fine gradations can feel arbitrary when you’re trying to make a quick bedside judgment. The ECOG scale was developed in 1960 as a stripped-down alternative, collapsing all those gradations into just five functional categories.¹ The World Health Organization later adopted it, which is why you’ll see the terms “WHO PS,” “ECOG PS,” and “Zubrod scale” used interchangeably in medical literature. They all refer to the same thing.

What Each Grade Means

The scale’s five grades describe a spectrum from full independence to total disability. Understanding what each one actually captures helps you see why the boundaries between them matter so much for treatment decisions.

Grade 0: You are fully active and able to carry on all your normal activities without restriction. This is the person who feels essentially the same as before their diagnosis, at least physically.
Grade 1: You are restricted in physically strenuous activity but still ambulatory and able to carry out light or sedentary work. Think of someone who can do office work or light housework but gets winded doing yard work.
Grade 2: You are up and about more than half the day and capable of all self-care, but you cannot carry out any work activities. This is the critical threshold grade: many clinical trials and aggressive treatment protocols draw the line here.
Grade 3: You are capable of only limited self-care and confined to a bed or chair for more than half of your waking hours.
Grade 4: You are completely disabled, cannot carry out any self-care, and are totally confined to bed or chair.

Some references include a Grade 5, which simply means the patient has died. In practice the scale is used to assess living patients, so the 0-to-4 range is what matters at the bedside.

How the Score Shapes Treatment Decisions

Performance status is not just a descriptive label. It directly determines which therapies oncologists will offer. A patient rated at 0 or 1 is typically considered fit for most standard chemotherapy regimens, surgery, or enrollment in a clinical trial. A patient rated at 2 is in a gray zone: some treatments may still be appropriate, but the risk-benefit calculus shifts. At 3 or 4, aggressive chemotherapy is rarely offered because the likelihood of benefit drops steeply while the risk of harm stays high or even increases. One study of small-cell lung cancer patients found that those with an ECOG score of 4 who received chemotherapy had extremely short survival and a high rate of early death, raising real questions about whether treatment was appropriate at all.²

Research on patients with good baseline performance status (ECOG 1) who were nearing the end of life found that receiving chemotherapy was actually associated with lower quality of dying compared with forgoing it. Among patients with moderate or poor status (ECOG 2 or 3), chemotherapy showed no meaningful relationship to quality of dying at all.³ Findings like these underscore why the score is not merely academic: it is supposed to help clinicians and patients decide when continuing aggressive treatment crosses the line from hopeful to harmful.

The Clinical Trial Gatekeeping Problem

Performance status also functions as a gatekeeper for clinical trials, and the gate has been narrowing. An analysis of cancer drug trials that led to FDA approval between 2009 and 2023 found that only about a quarter of trials allowed patients with poor performance status to enroll. Worse, that proportion dropped sharply over time: roughly 43% of trials permitted ECOG 2 or higher patients in the 2009–2013 window, but only about 18% did so by 2019–2023. When poor-PS patients were enrolled, they made up a small fraction of participants, with a median of just over 4%.⁴

This creates a self-reinforcing gap. Patients with lower performance status are excluded from the trials that generate evidence, so doctors then have little data to guide treatment decisions for those very patients. The American Society of Clinical Oncology (ASCO), along with Friends of Cancer Research, has pushed to broaden eligibility criteria, arguing that doing so would shorten recruitment timelines, increase diversity among trial participants, and produce results that better reflect the real-world population oncologists actually treat.⁵

A study of advanced non-small-cell lung cancer patients illustrated the stakes: when conventional trial-like criteria (ECOG 0–1, no brain metastases, normal kidney function) were applied, 73% of the real-world cohort would have been ineligible. Yet nearly half of those “ineligible” patients actually received treatment and survived about as long as the “eligible” group.⁶ It is hard to look at numbers like that and not wonder how many patients are being shut out of trials unnecessarily.

Why Clinicians Often Disagree on the Score

For a scale with only five levels, you might expect two doctors to assign the same grade most of the time. They frequently do not. A systematic review of inter-rater reliability studies found results all over the map: some studies reported poor agreement among clinicians, while others reported strong agreement, with reliability statistics ranging widely depending on the study.⁷ An updated meta-analysis pooling larger samples found that the ECOG scale’s inter-rater correlation was about 0.71, which sounds respectable in the abstract but still means clinicians frequently land on different grades for the same patient.⁸

Part of the problem is that the grade descriptions are short and somewhat vague. “Ambulatory and capable of all self-care but unable to carry out any work activities” leaves plenty of room for interpretation. A vignette study found that agreement with a gold-standard rating ranged from about 19% to 57% depending on the patient scenario, and that clinicians were influenced by factors that should be irrelevant. Patients described with socially desirable qualities received better performance scores, and female patients were scored as having worse performance status, independent of the clinical picture.⁹ The score, in other words, is not purely objective: a patient’s perceived likability and gender can nudge it.

The Gap Between Doctor and Patient Ratings

Disagreement doesn’t just happen between two clinicians looking at the same patient. It also happens between the clinician and the patient themselves. A study of patients with advanced cancer found that more than half disagreed with their physician’s ECOG rating. When there was a mismatch, physicians were three times more likely to rate the patient as better off than the patient was to rate themselves as better off. On average, physicians gave more optimistic scores than patients did.¹⁰

Interestingly, the direction of the disagreement flipped depending on how sick the patient was. Among patients who were relatively well, physicians tended to give rosier scores than the patients themselves chose. Among patients who were quite ill, the opposite happened: patients rated themselves better than their doctors did. The weighted agreement between patients and physicians was only fair, and the disagreement itself turned out to be clinically meaningful. When physicians and patients saw the patient’s status differently, outcomes tended to be worse.¹⁰ This is not just a measurement quirk: the mismatch itself seems to signal something about the patient’s trajectory.

Adding to the problem, performance status information often doesn’t even reach the team meetings where treatment plans are decided. One UK study found that only 14% of multidisciplinary team discussions included information about patient fitness.¹¹ A score that exists in principle but is absent from the room where decisions are made is not doing its job.

Prognostic Value and Its Limits

Despite all its measurement issues, the WHO/ECOG score does carry genuine prognostic power. A comprehensive analysis of nearly 27,000 patients with non-small-cell lung cancer identified good performance status as an independent favorable factor for survival, alongside early-stage disease and receiving treatment.¹² A change in performance status during treatment is also telling: among patients with locally advanced lung cancer undergoing chemoradiotherapy, those whose functional status deteriorated during treatment had worse progression-free and overall survival.¹³ Similarly, in pancreatic cancer patients undergoing treatment before surgery, about 8% experienced a decline in performance status, and their median survival was dramatically shorter than those who progressed to surgery.¹⁴

But the score does not capture everything about a patient’s condition. A study of advanced colorectal cancer patients compared WHO performance status against a more detailed quality-of-life physical-function measure. Among patients with the same WHO PS grade, those who scored well on the quality-of-life measure survived roughly twice as long as those who scored poorly. In a statistical model that included both, the quality-of-life measure was significantly prognostic for survival, while the WHO PS grade was not.¹⁵ In other words, the blunt categories of the WHO scale can lump together patients with very different outlooks. A single digit cannot summarize all the variation in how a person functions.

How It Compares to the Karnofsky Scale

The KPS scale remains in wide use, especially in palliative care and brain tumor settings, so understanding how the two tools map onto each other matters. An empirical analysis found that the best conversion groups KPS scores of 80–90 as equivalent to ECOG 1, KPS 60–70 as ECOG 2, KPS 40–50 as ECOG 3, and KPS 10–30 as ECOG 4. This mapping achieved about a 75% hit rate when tested against patients who had been scored on both scales.¹⁶

The two scales do not always agree at the patient level, and the disagreements are not random. A study of brain tumor patients found that discordant scoring between KPS and WHO PS increased as the patient’s deficits became more severe. In most discordant cases, the KPS would have denied trial access while the WHO PS would have allowed it.¹⁷ This means the choice of scale can directly affect whether a patient gets into a trial, particularly for patients with neurological deficits whose physical function is hard to summarize in a single number.

Wearable Technology as an Alternative

Given the subjectivity baked into the current approach, researchers have been exploring whether wearable devices like fitness trackers could offer a more objective measure of functional status. A systematic review found moderate evidence that physical activity metrics from wearable monitors correlate positively with performance status, while sedentary behavior metrics correlate negatively with it.¹⁸ The associations were modest, but they point toward a future where performance status is measured continuously rather than guessed at during a clinic visit.

A pilot multicenter study from the Alliance for Clinical Trials in Oncology tested this idea directly. Patients wore Fitbit devices, and researchers built models using daily steps, heart rate, and activity minutes to predict patient-reported physical function. The models explained a substantial portion of the variation in how patients rated their own function, and heart rate, activity, and symptom burden were all meaningful contributors.¹⁹ This is still early-stage work, and no sensor can replace a clinician’s overall assessment of the patient. But a sensor that continuously tracks how much a patient actually moves through their week captures something a five-minute clinic conversation does not.

Biases That Distort the Score

Beyond the inherent subjectivity of the scale’s wording, specific biases warp the score in consistent directions. The vignette study mentioned earlier found that social desirability and patient gender both swayed ratings. A patient described as married, employed, and cheerful might score better than an identical clinical profile presented as unemployed, socially isolated, or depressed.⁹ These biases are especially dangerous around the ECOG 2 boundary, where a one-grade shift can mean the difference between being offered treatment or being told nothing more can be done.

Overestimation of performance status is also common. When clinicians rate a patient more favorably than the patient’s actual function warrants, and there is a mismatch between the clinician’s view and the patient’s own assessment, outcomes tend to be worse.¹¹ One possible explanation is that an overly optimistic score leads to more aggressive treatment than the patient can tolerate. Another is that the discordance itself reflects a clinician who hasn’t fully recognized the patient’s decline. Either way, the direction of the error matters: an inflated score is not a neutral mistake.

What the Scale Does Not Measure

The WHO/ECOG scale was designed to be quick and universal, and that simplicity comes at a cost. It focuses on physical function and mobility but says nothing about cognitive function, emotional well-being, nutritional status, or symptom burden. A patient with severe cancer-related fatigue and a patient with severe pain may both land on ECOG 2, even though their treatment needs and likely trajectories differ. The colorectal cancer study that found quality-of-life physical function outperformed WHO PS as a survival predictor illustrates the gap: within the same performance status grade, patients can have vastly different levels of function and prognosis.¹⁵

Brain tumor patients present a particularly clear example. A patient with a high-grade glioma may have severe cognitive or speech deficits but still be physically mobile, making the usual ECOG criteria awkward to apply. The finding that KPS and WHO PS frequently disagree in this population, often in ways that affect trial eligibility, suggests the scale was not designed with neurological deficits in mind.¹⁷ For these patients, a single mobility-centered score may miss the most important aspects of their daily function.

None of this means the scale is useless. It remains one of the strongest single predictors of cancer survival available, it is fast to administer, and everyone in oncology speaks its language. But it works best as a starting point for conversation rather than a final answer. When a score of 0, 1, 2, 3, or 4 is the only piece of information driving a treatment decision, something important is almost certainly being left out.