Copenhagen Burnout Inventory

The Copenhagen Burnout Inventory (CBI) is a validated 19-item self-report instrument measuring burnout across three subscales: Personal Burnout (6 items), Work-Related Burnout (7 items), and Client-Related Burnout (6 items). Developed by Tage Kristensen, Marianne Borritz, Ebbe Villadsen, and Karl B. Christensen at the Danish National Institute of Occupational Health and published in Work & Stress in 2005, the CBI was designed and validated within the longitudinal PUMA cohort of 1,914 Danish human service workers. It is in the public domain and is the most widely used open-license alternative to the proprietary Maslach Burnout Inventory.

Unlike the MBI, which conceptualizes burnout as three-dimensional (exhaustion + depersonalization + reduced accomplishment), the CBI treats burnout primarily as physical and psychological exhaustion, partitioned by attribution context (personal, work, clients) rather than by symptom dimension. This conceptual choice was deliberate: Kristensen and colleagues argued that exhaustion is the core construct of burnout and that the depersonalization and inefficacy dimensions of the MBI are coping responses or consequences rather than burnout itself.

The CBI matters for two reasons: its open-license status and its conceptual reframe of burnout.

The MBI is licensed by Mind Garden and requires per-administration fees that make free public deployment impossible. The CBI's public-domain status removes this barrier and is the reason it is used in most free-to-use online burnout screens, government-sponsored workplace health surveys, and large-scale academic research where licensing budgets are constrained. The instrument has been translated into 30+ languages with substantial cross-cultural validation.

The conceptual contribution is equally important. The Maslach three-dimensional model has been challenged by empirical work showing that exhaustion is the dimension most predictive of clinically relevant outcomes (sickness absence, intent-to-leave, mental health), while depersonalization and reduced personal accomplishment correlate with exhaustion but contribute relatively little independent predictive variance (Kristensen 2005, Schaufeli & Taris 2005). The CBI operationalizes this empirical finding by focusing measurement on exhaustion and using attribution context to differentiate burnout types, rather than dispersing measurement across three loosely-correlated dimensions.

Predictive validity is well-established. Borritz et al. (2005, JOEM) followed 824 PUMA cohort members for 3 years and found that high baseline Personal Burnout predicted future sickness absence (OR = 2.1) and intent to leave the profession (OR = 3.5) at follow-up, controlling for demographic and job-characteristic covariates. Similar predictive associations have been reported in nurses, teachers, and general working populations across European, Asian, and Latin American samples.

The CBI was developed within the PUMA project (Project on Burnout, Motivation and Job Satisfaction), a longitudinal cohort study of 1,914 Danish human service workers conducted by the Danish National Institute of Occupational Health from 1999 to 2005. The project was led by Tage S. Kristensen, an occupational health psychologist, with co-investigators Marianne Borritz, Ebbe Villadsen, and Karl B. Christensen.

The motivation for developing a new instrument was twofold. First, the PUMA team needed an open-license measure they could administer freely across the cohort and disseminate to other researchers. The MBI’s proprietary licensing made it unsuitable for the cohort’s long-term, multi-wave design. Second, the team had theoretical disagreements with the Maslach three-dimensional model. Kristensen and colleagues argued that the depersonalization dimension was particularly problematic: it was originally framed as cynicism toward service recipients (patients, clients, students), but workers in non-service roles cannot meaningfully exhibit "depersonalization" toward an abstract job. The MBI-General Survey attempted to address this by reframing depersonalization as cynicism toward work itself, but this introduced its own conceptual difficulties.

The CBI’s solution was to treat exhaustion as the core burnout construct and to use attribution context (personal, work, clients) as the dimensional structure. This produces three subscales that are conceptually parallel and applicable across a range of occupational contexts, including non-service work and non-employment situations. Personal Burnout asks about generic exhaustion regardless of cause; Work-Related Burnout asks about exhaustion specifically attributed to work; Client-Related Burnout asks about exhaustion specifically attributed to working with clients.

The CBI was published in Work & Stress in 2005 (DOI 10.1080/02678370500297720) under public-domain terms. The paper presented baseline psychometric properties and the first wave of validation evidence; subsequent papers from the PUMA cohort (Borritz et al. 2005, 2006, 2010) presented longitudinal validation data and predictive validity findings for sickness absence, intent-to-leave, and mental health outcomes.

The CBI has three conceptually parallel subscales, each measuring exhaustion attributed to a different context:

The three subscales are conceptually independent. A person can score high on Personal Burnout without elevated Work-Related or Client-Related scores (suggesting non-occupational sources of exhaustion). A person can score high on Work-Related Burnout without elevated Personal Burnout (rare; typically indicates the work-context attribution is genuinely localized). A person can score high on Client-Related Burnout while their Work-Related score remains moderate (suggesting that client interactions specifically, rather than work generally, are the burnout source).

The CBI reports the three subscales separately, with no aggregate "CBI total" score. This is a deliberate design choice: aggregating across subscales obscures the diagnostic information they provide about the source of exhaustion. The LifeByLogic Stress & Burnout Index uses only the Personal Burnout subscale because it is the universal version applicable across all employment statuses.

The CBI Personal Burnout subscale items are reproduced verbatim from Kristensen, Borritz, Villadsen & Christensen (2005) under public-domain status. Items 1–3 use the question stem "How often do you feel..." and items 4–6 use the stem "How often do you feel that..." with appropriate grammar. The 5-point response scale is unique to the CBI:

• Always — 100
• Often — 75
• Sometimes — 50
• Seldom — 25
• Never / Almost never — 0

Note item 4 ("I can’t take it anymore"). Kristensen 2005 specifically flagged this as a sentinel item: a single response of 75 or 100 on this item indicates clinically meaningful exhaustion regardless of the overall subscale score. The LifeByLogic Stress & Burnout Index uses item 14 (corresponding to CBI Personal Burnout item 4 in the LBL combined screen) at the ≥75 threshold as one of three care-aware escalation criteria.

Scoring is straightforward: each item is rated on the 5-point Likert scale and the subscale score is the arithmetic mean of item responses. There is no overall CBI total; subscales are reported separately.

// Inputs: responses for each subscale, each item in {0, 25, 50, 75, 100}

// Personal Burnout (6 items)
personal_burnout = mean(personal[1..6])     // 0..100

// Work-Related Burnout (7 items, item 5 reverse-scored)
// Item 5: "Do you have enough energy for family and friends during leisure time?"
function work_item_value(i):
    if i == 5: return 100 - work[i]    // reverse-scored
    return work[i]

work_burnout = mean([work_item_value(i) for i in 1..7])      // 0..100

// Client-Related Burnout (6 items)
client_burnout = mean(client[1..6])         // 0..100

// Severity bands (Kristensen 2005)
function band(score):
    if score <= 24: return "Low"
    if score <= 49: return "Moderate"
    if score <= 74: return "High"
    return "Severe"

// Care-aware sentinel (Kristensen 2005)
if personal[4] >= 75:
    flag = "sentinel: 'I can't take it anymore' is endorsed at the action threshold"
  

A respondent who answers all 6 Personal Burnout items as "Always" (100) gets a Personal Burnout score of 100. A respondent who answers all items as "Never" (0) gets a score of 0. A respondent with mixed responses (e.g., 75, 50, 75, 25, 50, 0) gets a score of (75+50+75+25+50+0)/6 = 45.83.

The Work-Related Burnout subscale includes one reverse-scored item ("Do you have enough energy for family and friends during leisure time?"). Item 5 of this subscale must be reverse-coded (0 ↔ 100, 25 ↔ 75, 50 unchanged) before computing the mean. The Personal and Client-Related Burnout subscales have no reverse-scored items.

The four-band severity classification per Kristensen et al. (2005):

PUMA baseline distribution (n = 1,914 Danish human service workers): Personal Burnout mean ≈ 35 (SD ≈ 19), median ≈ 33. Work-Related Burnout mean ≈ 33 (SD ≈ 18). Client-Related Burnout mean ≈ 31 (SD ≈ 19). Cross-cultural samples replicate similar distributions, with notable elevation in healthcare workers during the 2020–2022 pandemic period (mean Personal Burnout shifted up by ~5–10 points across multiple national samples).

The 50-point threshold is the most commonly cited "concerning elevation" anchor. The 75-point threshold is a care-aware threshold; the LifeByLogic Stress & Burnout Index uses Personal Burnout ≥75 as one of three care-aware escalation criteria, alongside PSS ≥ 33 and item 14 ≥ 75.

Internal consistency

Cronbach’s alpha for the three subscales in Kristensen 2005 PUMA cohort baseline (n = 1,914):

• Personal Burnout: α = 0.87
• Work-Related Burnout: α = 0.87
• Client-Related Burnout: α = 0.85

Subsequent validation studies have replicated alphas in the 0.83–0.92 range across multiple language versions and occupational contexts. The PUMA team specifically reported that the three-subscale structure was robust to confirmatory factor analysis with adequate model fit.

Predictive validity

Borritz et al. (2005) followed 824 PUMA cohort members from baseline to 3-year follow-up. High baseline Personal Burnout predicted:

• Sickness absence days: OR = 2.1 (95% CI 1.5–3.0) for any sickness absence in the follow-up period
• Intent to leave the profession: OR = 3.5 (95% CI 2.4–5.1)
• Self-reported physical and mental health symptoms at follow-up

Effects survived adjustment for age, sex, occupational role, and baseline self-reported health. The findings established the CBI as a clinically meaningful predictor of work-relevant outcomes, not just a cross-sectional symptom inventory.

Concurrent validity

The CBI Personal Burnout subscale correlates with: MBI emotional exhaustion (r ≈ 0.71–0.79), depression symptoms (PHQ-9, r ≈ 0.55–0.65), perceived stress (PSS-10, r ≈ 0.55–0.65), insomnia (Membrive-Jiménez 2022, r ≈ 0.39 burnout-insomnia correlation), and life satisfaction (negative correlation, r ≈ -0.45). The pattern is consistent with the CBI capturing exhaustion-spectrum constructs rather than dimensions specific to work cynicism or efficacy.

Cross-cultural invariance

The CBI has been translated into 30+ languages with formal psychometric validation in English, German, Mandarin, Spanish, Brazilian Portuguese, Japanese, French, Italian, Greek, Arabic, Korean, and others. Three-factor structure replicates in most contexts; subscale alphas range from 0.78 to 0.92. The Personal Burnout subscale is the most robust to cross-cultural translation, reflecting its generic non-occupational item content.

The MBI and CBI are the two dominant burnout instruments. Choosing between them depends on conceptual fit, deployment context, and licensing constraints.

The two instruments correlate r ≈ 0.55–0.75 at the exhaustion dimension, suggesting they measure overlapping but not identical constructs. The MBI’s depersonalization and accomplishment dimensions have weaker counterparts in the CBI, consistent with the CBI’s deliberate exhaustion-focused conceptualization.

For free public deployment, the CBI is the standard choice. The Personal Burnout subscale specifically is the version used in most consumer-facing burnout screens, including the LifeByLogic Stress & Burnout Index.

1. Exhaustion-only conceptualization may miss aspects of burnout

The CBI’s deliberate focus on exhaustion is its strength but also its limitation. Workers experiencing high cynicism without high exhaustion (a recognized "dropping-out" pattern in burnout literature) may not be detected by the CBI. Workers with high inefficacy ("Why bother?") feelings without exhaustion may also screen as low. Researchers wanting full coverage of the Maslach three-dimensional construct should pair the CBI with separate measures of cynicism and inefficacy, or use the MBI directly.

2. Sentinel item (item 4) is sensitive but not specific

Item 4 ("I can’t take it anymore") functions as a clinical sentinel: endorsement at ≥75 indicates concerning exhaustion regardless of the overall subscale. But it can also be triggered by acute crisis states (recent bereavement, acute medical illness, situational overload) that are not chronic burnout. The sentinel should prompt further evaluation, not standalone diagnosis.

3. Self-report dependency

Like all burnout instruments, the CBI is self-report and inherits the standard self-report concerns: insight limitations, social desirability bias, and demographic variation in reporting thresholds. Workers in highly demanding roles may underreport due to professional identity ("I’m fine"); workers with high-functioning anxiety may overreport.

4. Recovery trajectory not specified

The CBI provides cross-sectional measurement but does not specify recovery trajectories. Borritz 2005 showed that approximately 50% of high-burnout workers normalized over 3-year follow-up while ~50% remained elevated, but the boundary between "recovered" and "still vulnerable" is not operationalized in the instrument itself. Serial CBI administration over months provides recovery information; single-point measurement does not.

5. Construct overlap with depression

Personal Burnout correlates with depression (r ≈ 0.55–0.65) at levels that, after disattenuation, raise the same discriminant-validity questions Bianchi 2021 raised for the MBI exhaustion dimension. High CBI Personal Burnout warrants depression screening (PHQ-9 or equivalent); the two constructs cannot be reliably distinguished by self-report alone.

6. Limited adolescent validation

The CBI was developed and validated in adult populations (ages 20–65). Validation in adolescents is limited; for youth populations, instruments developed specifically for that age range (e.g., the School Burnout Inventory; Salmela-Aro 2009) are generally preferred.

The full methodology page documents the implementation choices in detail: instrument selection rationale, scoring algorithm with reverse-coding, severity-band derivation, archetype thresholds, care-aware logic, validation evidence, population norms, and limitations.

What does CBI stand for?

CBI stands for Copenhagen Burnout Inventory. The instrument was developed at the Danish National Institute of Occupational Health in Copenhagen by Kristensen, Borritz, Villadsen, and Christensen, and published in Work & Stress in 2005.

Is the CBI free to use?

Yes. The CBI is in the public domain. Kristensen and colleagues released the instrument freely; users are asked to cite the original 2005 publication when reproducing or referencing the items. This applies to clinical use, research, and consumer-facing applications. The open-license status was a deliberate choice, motivated by the proprietary status of the Maslach Burnout Inventory.

How long does the CBI take to complete?

The full CBI (19 items) takes approximately 5–7 minutes. The Personal Burnout subscale alone (6 items) takes 2–3 minutes. Total time including review of subscale scores and severity bands is 8–10 minutes for the full instrument or 3–5 minutes for Personal Burnout only.

What is a concerning CBI score?

Per Kristensen 2005 anchors: scores ≤ 24 are Low, 25–49 Moderate, 50–74 High, and ≥ 75 Severe Burnout. The 50-point threshold is the most commonly cited "concerning elevation" anchor. The 75-point threshold is the care-aware threshold; persistent burnout at this level warrants professional consultation regardless of demographic context.

Can I have burnout without high CBI scores?

Yes. The CBI focuses on exhaustion. Workers experiencing burnout primarily as cynicism (the "I’ve checked out" pattern) or inefficacy (the "Why bother?" pattern) may have moderate CBI scores while still meeting clinical criteria for burnout. The MBI’s three-dimensional model captures these patterns more completely. If exhaustion is not your primary symptom but you suspect burnout, a clinical interview is more reliable than self-report alone.

How does the CBI compare to the MBI?

The MBI conceptualizes burnout as three-dimensional (exhaustion, cynicism/depersonalization, reduced personal accomplishment). The CBI focuses on exhaustion and partitions it by attribution (personal, work, clients). The MBI is proprietary; the CBI is public domain. The two instruments correlate r ≈ 0.55–0.75 at the exhaustion dimension. For free public deployment, the CBI is the standard choice; for research where the three-dimensional Maslach model is theoretically central, the MBI is preferred.

What is the PUMA cohort?

PUMA (Project on Burnout, Motivation and Job Satisfaction) was a longitudinal cohort study of 1,914 Danish human service workers conducted by the Danish National Institute of Occupational Health from 1999 to 2005. The CBI was developed within this study. Borritz et al. (2005) published the canonical 3-year follow-up validation paper showing predictive validity for sickness absence (OR = 2.1) and intent to leave the profession (OR = 3.5).

Is item 4 really a "sentinel" item?

Yes. Kristensen 2005 specifically flagged Personal Burnout item 4 ("I can’t take it anymore") as a sentinel: a single response of 75 ("Often") or 100 ("Always") indicates clinically meaningful exhaustion regardless of the overall subscale score. The LifeByLogic Stress & Burnout Index uses this item at the ≥75 threshold as one of three care-aware escalation criteria, ensuring that someone endorsing this item at the action level is flagged for support resources even if their overall scores are otherwise moderate.