WHO-5 Well-Being Index: scoring, cutoffs and licence

Three answers before the guide. Yes, the WHO-5 is free. No, you do not have to ask anyone’s permission to hand it to a client. And yes, there is a line: its Creative Commons Attribution-NonCommercial-ShareAlike 3.0 IGO licence restricts specific things, including what this blog may reproduce of it. The rest is scoring, the 13-versus-50 trap, and what its evidence base does and does not cover.

What it measures, and why every item is positively worded

WHO’s own description: the WHO-5 is a self-report instrument measuring mental well-being — five statements about the past two weeks, each on a 6-point scale, higher scores meaning better well-being. The five topics are mood, calm, energy, waking rested, and interest in daily life. Not one symptom question. The six options run from “all of the time” to “at no time” and score 5 down to 0 in an order that is not the intuitive one: “some of the time” scores 1, not 2.

That wording is deliberate. As the systematic review by Topp and colleagues (2015) puts it, WHO considers positive well-being another term for mental health; hence only positively phrased items. Nor is it a symptom scale in reverse: Bech and colleagues (2003), with 9,542 participants, concluded that it reflects aspects other than just the absence of depressive symptoms. Hence its niche: where the PHQ-9 asks about hopelessness and death, and the GAD-7 about uncontrollable worry, the WHO-5 asks about energy and good spirits, and it hands cleanly to a client who would bristle at a depression questionnaire.

Scoring, and the 13-versus-50 trap

Scale	Range	Suggested threshold
Raw (sum of the five items)	0 to 25	below 13
Percentage (raw × 4)	0 to 100	below 50

Multiplied by four, the raw score produces a percentage with the same meaning. WHO states the cutoff with both figures together: a percentage score below 50, or a raw score below 13, has been suggested as a cutoff for poor mental well-being and as an indication for further assessment for the possible presence of a mental health condition. Note the verbs — neither an established threshold nor a finding.

The trap comes from the official form. The one Spanish version WHO distributes carries a scoring block that contradicts itself: it instructs the user to multiply the sum by four to obtain “a score between 0 and 100,” and two lines later states that a score below 13 indicates low well-being. Read literally, that tells a clinician to compare a 0-to-100 score against 13, which would flag almost nobody. If you have seen that confusion, the sheet is the obvious candidate.

The correct reading, per WHO’s current English text: these are the same threshold on two scales; 13 raw equals 50 percent. Hence the practical rule: record which scale you used. “WHO-5: 48” and “WHO-5: 12” can be the same client, and without the label nobody will know in six months.

Two nuances. WHO writes “below 50”; Topp and colleagues recommend ≤ 50: nothing clinical turns on a single point, and the cutoff is not yours to move. And the Spanish sheet WHO distributes keeps a rule the current English version drops: administer a further depression measure not only below threshold but if any item is answered 0 or 1.

For follow-up there is a second number: Topp and colleagues put the threshold for a clinically relevant change at 10 points, without saying on which scale. The Spanish sheet WHO distributes does, applying its 10% to the percentage score. Below that, treat the movement as noise. We found no official standard for how often to repeat it: the floor is the instrument’s two-week window, and the rest is consistency — the framework in the guide on measuring client progress.

“WHO-5 percentage: 56 (previous: 40). A 16-point rise, above the clinically relevant change threshold and consistent with the resumed activities she reports.”

What a low score does not tell you

It is not a diagnosis. WHO speaks of an indication for further assessment, not a finding, and Topp and colleagues endorse the scale as a screening tool and as an outcome measure, with a diagnostic interview as the second step. The “never a diagnostic test” is ours, not theirs.

Specificity is the weak side. Across the 18 studies in that review, sensitivity was 0.86 and specificity 0.81, measured mostly against DSM-IV depression assessed by structured interviews. In practice it can fall further: in 431 primary care patients in Germany, against the CIDI, the WHO-5 produced sensitivity of 93% (95% CI: 85 to 98) and specificity of 64% (59 to 69). In a private caseload, where prevalence is lower, a substantial share of positives will not have a depressive disorder.

It measures well-being, not symptoms. A low WHO-5 in someone grieving, burned out or medically ill is real information about their life, not evidence of a disorder.

It does not ask about risk. Unlike the PHQ-9, it contains no suicidal ideation item, and a high score rules nothing out. Any risk disclosure calls for an immediate clinical risk assessment under your own protocol and training; this guide will not hand you those steps, and be wary of any blog that does. What you can take from here is how that assessment gets documented, in the guide on documenting a suicide risk assessment, plus a resource for the client: the crisis line serving their location, alongside local emergency services.

Translations and the evidence behind them

Three questions get conflated here, and the answers differ.

Translation. WHO’s page reports the WHO-5 in over 30 languages. For Spanish there is exactly one, and the PDF explains why that matters: the translation predates WHO accepting copyright in 2024, WHO is not responsible for its accuracy, and in any inconsistency the English version is the binding and authentic one. We located no separate Mexican or Latin American edition.

Reliability and factor structure. A single factor in all four Spanish-language samples we located, with alphas of 0.82 to 0.86: 199 older adults (Lucas-Carrasco, 2012) and 104 euthymic patients with bipolar disorder (Bonnín et al., 2018) in Spain; 1,462 adolescents (Espitia-Correa et al., 2020) and 105 patients at a dental implant service (Simancas-Pallares et al., 2016) in Colombia. Two caveats. The alphas are not uniform: Campo-Arias and colleagues got 0.67 and 0.70 across two administrations to 157 Colombian adolescents, a figure both Colombian papers cite. And those two measured a four-option variant scored 0 to 15, not the six-option 0-to-25 sheet WHO distributes: reliability of a different version.

Diagnostic accuracy. We located none. None of those studies establishes sensitivity, specificity or an optimal cutoff against a structured diagnostic interview; the Cartagena one concludes “questionable” construct validity and describes itself as a validation “without a reference standard.”

For Mexico we located two studies and no norms. A measurement-invariance analysis across 12 Latin American countries (Caycho-Rodríguez et al., 2023) reports its Mexican subsample — 300 people, alpha of 0.95, a single factor — again on the four-option variant. And 600 students in the State of México, alpha and omega both at 0.909 (Macias Becerril et al., 2025), three pages in a minor journal. Neither is a clinical-sample validation, neither supplies diagnostic accuracy, and neither produces norms: do not print a national reference value.

The lesson travels: translation is not validation, and reliability is not diagnostic accuracy. The 50 cutoff you apply comes from predominantly European samples. The same distinction runs through the guide on psychological test licensing.

Free, yes; unconditional, no

First, who grants the licence, because almost no summary has this current. The WHO-5 was developed by Per Bech at the Psychiatric Centre North Zealand in Copenhagen, out of the 1998 DepCare meeting report — but in 2024 that centre and WHO agreed to assign copyright in the WHO-5 to WHO. It is now a WHO publication, WHO/UCN/MSD/MHE/2024.1, with its own page; pages still crediting the Danish unit are out of date.

What the licence does not ask of you: a clinician who prints the official PDF, gives it to a client, scores it and files it is inside the licence, with no form and no fee, because WHO states plainly that permission is not required to use WHO materials issued under CC BY-NC-SA 3.0 IGO. Whether a fee-charging practice sits on the non-commercial side of that line, WHO does not say anywhere; our reading is that it does.

It does ask four things if you build the instrument into your own materials:

Attribution, using the citation WHO supplies. “World Health Organization. The World Health Organization-Five Well-Being Index (WHO-5). Geneva: World Health Organization; 2024. License: CC-BY-NC-SA 3.0 IGO”.
No logo, no implied endorsement. No suggestion that WHO endorses any organization, product or service, and no use of the WHO logo.
Same licence on anything you adapt. The legal code requires distributing any adaptation under this licence, a later version with the same elements, or an equivalent.
A disclaimer if you translate. “Adaptation” expressly includes translations, so rolling your own evades nothing; WHO asks you to add that the translation was not created by WHO and that the original English edition is binding and authentic.

And the boundary: commercial use does require permission, per WHO’s copyright page. The legal code pins down where that boundary falls — it bars exercising the granted rights “in any manner that is primarily intended for or directed toward commercial advantage or private monetary compensation.” A purpose test, not a test of whether anyone was charged.

Why we are not reprinting the five items

Explaining the decision beats announcing it: the reasoning transfers to your own site and your own software. First, that purpose test reaches us: this blog is published by a company that sells software to the people reading it, and although nobody pays to read it, it exists to build a product’s audience. Second, ShareAlike: laying the five items out in HTML is at minimum an arrangement of the work, and we would rather not litigate where that adaptation ends and the surrounding page begins — the legal code exempts a “Collection,” but a laid-out scoring box is not obviously just that. Third, there is no need: the official PDF is free and one click away.

To be precise: we are not saying that reprinting the items would be unlawful. We found no WHO statement on quotation for commentary, no de minimis threshold, and nothing resolving how this licence interacts with fair use or its equivalents. This is our reading of the licence and our editorial policy, and we apply it to every instrument — describe it, link the official source.

Where the score belongs in the record

A WHO-5 score is a test result in the client’s chart, and most jurisdictions regulate what a clinical record must contain. In Mexico, for example, NOM-004-SSA3-2012 files it under “relevant results of diagnostic and treatment support studies” (numeral 6.2.3) and requires date, time, and the full name and signature of whoever wrote the note (5.10) — though 6.2.3 sits inside the nota de evolución, which the norm says “deberá elaborarla el médico”, so an independent psychologist is not straightforwardly its addressee; that scope question is worked through in the guide on Mexico’s clinical-record standard. Wherever you practise, verify what your licensing board, state law and payers require.

Six elements keep a score useful two years later: instrument and version, date administered, raw and percentage score, the threshold you applied, the change since the previous administration, and the clinical interpretation in words. Never the number alone.

That kind of record is mechanical work that grows with every client. gesell.ai includes validated scales with automatic scoring — PHQ-9 and GAD-7 among them — and drops the score into the client’s structured chart, next to earlier measurements and ready for the note. The scale supplies the data point; the clinical reading stays yours.