Where Is Your Burnout Coming From? Design, Scoring, and Limitations of a Six-Domain Burnout Source Inventory

Abstract

Most established burnout instruments measure whether burnout is present and how severe it is. They were built for the workplace, and the World Health Organization's ICD-11 entry for burn-out is explicitly limited to the occupational context. The women this instrument was built for describe an exhaustion that does not stay at work: it moves between their beliefs, their relationships, their boundaries, their sense of purpose, their money, and their bodies. The Where Is Your Burnout Coming From? quiz is a 33-screen, self-administered inventory that asks a different question from the established scales. Rather than measuring how burned out a respondent is, its primary aim is attribution: which of six life domains (Beliefs, Belonging, Boundaries, Business, Budget, Body, referred to as the six B's) is carrying the most weight for this respondent right now. Thirty forced-choice paired statements produce a ranked profile across the six domains, and a separate three-item intensity block estimates how deep and how long-standing the burnout is. This paper documents the conceptual model, the item architecture, the exact scoring rules, the exclusions and assumptions the instrument depends on, its known limitations, and the validation work that has not yet been done. It is published so that readers, practitioners, and reviewers can evaluate the instrument on its actual design rather than on its marketing.

1. Purpose and scope

The quiz is a reflection and triage tool offered free to the public at quiz.iamieshiadanielle.com. It is used to help a respondent name the domain of life most likely to be driving her exhaustion, and to route her to the part of a coaching curriculum that addresses that domain. It is not a diagnostic instrument, it is not a clinical screen, and it produces no clinical cutoffs.

This document exists for three reasons. First, any instrument that tells a person something about herself owes her a plain account of how that conclusion was reached. Second, transparency about scoring, exclusions, and assumptions is the only honest basis on which a self-assessment can ask to be trusted. Third, the author intends to collect response data and publish validation results, and this document fixes the design in writing before that data exists, so that later claims can be checked against it.

2. Background: how burnout has been measured

The term burnout entered the professional literature through Herbert Freudenberger's 1974 description of exhaustion among volunteer staff in a free clinic (Freudenberger, 1974). Christina Maslach and Susan Jackson operationalized it with the Maslach Burnout Inventory (MBI), which measures three dimensions: emotional exhaustion, depersonalization (later generalized as cynicism), and reduced personal accomplishment (Maslach and Jackson, 1981). The MBI remains the most widely used measure, and its three-factor structure has been replicated across large international samples (Poghosyan, Aiken, and Sloane, 2009).

Two later instruments matter for the present design. The Oldenburg Burnout Inventory (OLBI) reduced the construct to two dimensions, exhaustion and disengagement, and deliberately included positively and negatively worded items to reduce response bias (Demerouti, Bakker, Vardakou, and Kantas, 2003; Reis, Xanthopoulou, and Tsaousis, 2015). The Burnout Assessment Tool (BAT) reframed burnout around four core dimensions, exhaustion, mental distance, cognitive impairment, and emotional impairment, and was built in part to address the MBI's limitations as a screening instrument (Schaufeli, Desart, and De Witte, 2020).

In 2019 the World Health Organization included burn-out in the eleventh revision of the International Classification of Diseases as an occupational phenomenon (code QD85), defined by energy depletion, increased mental distance or negativism toward one's job, and reduced professional efficacy. The WHO entry states that burn-out is not classified as a medical condition and that it "refers specifically to phenomena in the occupational context and should not be applied to describe experiences in other areas of life" (World Health Organization, 2019).

That occupational boundary is exactly where the present instrument departs from the established literature. Researchers have already extended the construct beyond paid work: the Parental Burnout Assessment measures exhaustion, contrast with a former parental self, feelings of being fed up, and emotional distancing among parents (Roskam, Brianda, and Mikolajczak, 2018). Maslach and Leiter's six areas of worklife model (workload, control, reward, community, fairness, and values) established that burnout is best understood as a mismatch between a person and her environment rather than a personal deficiency (Leiter and Maslach, 1999; Maslach and Leiter, 1997).

Finally, the overlap between burnout and depression is well documented and unresolved. Bianchi, Schonfeld, and Laurent's review found substantial overlap between the two constructs and cautioned against treating burnout as a clearly distinct entity (Bianchi, Schonfeld, and Laurent, 2015). This finding shapes the instrument's care language and referral rule, described in Section 8.

3. The conceptual model: six B's

The instrument is built on Ieshia Danielle's Becoming Methodology, a coaching framework developed over roughly a year of practice with Christian women, first in a small pilot intensive and then at a live event in Mississippi. The framework holds that the exhaustion its clients describe as burnout is not a single thing with a single source. It is sustained by whichever of six life domains is most out of order, and the domains are addressed in a fixed teaching order because earlier domains tend to feed later ones.

Three modeling decisions follow from the framework and are stated here as design commitments rather than findings.

The domains are treated as exhaustive for this population. The framework author reviewed the patterns arriving in her coaching inbox, including from pregnant women and women in ministry roles, and concluded that no seventh domain was needed. This is an assumption, not a result, and Section 7 lists it as one.

The domains are not treated as mutually exclusive. A respondent's burnout can be, and in the framework's language usually is, braided from more than one domain. The scoring therefore reports a ranked profile rather than a single label, and it has an explicit blend rule (Section 5.3).

Mapping decisions are fixed. Some statements could plausibly belong to more than one domain. The framework resolves these deliberately: "I don't know who I am outside of my roles" is scored as Boundaries (enmeshment), not Beliefs. Doing everything oneself with building resentment is scored as Belonging. Low self-esteem language is scored as Beliefs when the trigger is inner narrative and as Body when the trigger is self-image and upkeep. These rules are documented so that any future revision of item wording preserves the same mapping.

The instrument does not measure the MBI's or the BAT's dimensions and does not claim to. Its relationship to those instruments is complementary: they answer "how burned out, and in what form," and this instrument answers "from which part of life." A respondent who wants a validated measure of burnout severity should use one of the instruments cited in Section 2.

4. Instrument architecture

4.1 Why forced choice

Earlier versions of the quiz (versions 2 and 3) used agreement rating scales. In review they showed a predictable problem: a deeply exhausted respondent can endorse nearly everything at the high end, which yields a flat profile with every domain scored high and no usable signal about which domain was primary. This is a known behavior of Likert-type scales under acquiescence and extreme-response styles, and it is the standard motivation for multidimensional forced-choice formats, in which the respondent must choose between statements rather than rate each one (Brown and Maydeu-Olivares, 2011; Hontangas et al., 2015).

Forced choice trades one thing for another. It always differentiates within a person (the respondent cannot pick both), which is exactly the property the instrument needs for attribution. The cost is that the resulting scores are ipsative: they describe the relative standing of the six domains within one respondent and carry no information about how that respondent compares to anyone else (Brown and Maydeu-Olivares, 2013). Section 10 returns to this.

4.2 Pair construction

The six domains form 15 unique pairings. Each pairing appears exactly twice in the instrument, once under each of two stems, with different wording each time. This yields 30 choice screens. Each domain therefore appears in exactly 10 screens, and each domain's raw score has a fixed range of 0 to 10. The two appearances of each pairing are what allow the convergence check in Section 5.5.

Screen order was constrained so that no domain appears on two consecutive screens, to reduce carry-over and pattern-recognition effects. The domain behind each statement is never shown during the quiz; the six labels are revealed only in the report.

Each pair was written to a weight-matching rule: the two statements in a pair should be roughly equally comfortable, or equally uncomfortable, to admit. A pair in which one option is obviously more embarrassing would push respondents toward the safer option regardless of accuracy, which is the social-desirability problem forced choice is meant to reduce, not reintroduce.

4.3 The two stems

The 30 screens are split into two halves of 15, each under a different stem.

Stem One (screens 1 to 15): "Lately, this is more like me..." Each statement describes a lived symptom. Sample pairs:

• "I keep replaying the same story about why things can't change for me." (Beliefs) / "I'm tired in a way that sleep doesn't fix." (Body)
• "I say yes while everything in me is saying no." (Boundaries) / "My days are one loop: work, home, feed everybody, repeat." (Business)
• "I avoid looking at my bank statements because I already know." (Budget) / "Everybody's meals, health, and appointments come before mine." (Body)

Stem Two (screens 16 to 30): "Right now, this would help me more..." Each statement describes a desired relief. Sample pairs:

• "Permission to say no without feeling guilty." (Boundaries) / "One person who really sees me." (Belonging)
• "Waking up and actually feeling rested." (Body) / "Work that feels connected to my purpose again." (Business)
• "A plan for my money instead of a monthly scramble." (Budget) / "Fewer plates to spin, without feeling like I let everybody down." (Boundaries)

The two stems are the instrument's central design idea. Asking about symptoms alone invites the respondent to report whatever is most socially acceptable to be tired from. Asking about desired relief approaches the same domain from the other side: what a person most wants fixed is often a more honest signal of what is most broken than what she is willing to say is wrong. Where the two stems agree, the attribution is stronger; where they disagree, the report says so rather than hiding it.

The domain labels shown in parentheses above are scoring metadata. The complete 30-pair item bank is not reproduced in this document; the instrument's questions, scoring, and result types are the intellectual property of the author, and full publication would also allow respondents to reverse-engineer their results, which would degrade the very signal the instrument exists to capture. Qualified researchers may request the full item bank for review purposes through the contact on the site.

4.4 The intensity block

Three additional screens follow the 30 pairs. They look different on purpose (a single question with stacked answer options, no stem) because they measure something different: the 30 pairs estimate where the burnout comes from, and these three estimate how deep it goes.

Screen 31, duration: "How long have you felt this way?" Four options scored 1 to 4, from "A few weeks. This feels like a season" to "So long I can't remember feeling any other way."

Screen 32, impact: "How much is it shaping your day-to-day right now?" Four options scored 1 to 4, from "It comes and goes. I have good weeks" to "I don't feel like myself at all anymore."

Screen 33, presentation: "Where does it show up the loudest?" Three options (body, emotions, spirit). This item is not scored. It selects which of three short passages appears in the report and has no effect on any number.

4.5 Administration

The quiz is mobile-first, one screen at a time, one tap per screen, with no typing. It takes roughly three to four minutes. The instruction screen tells the respondent: "Sometimes both will sound like you. Sometimes neither will. That's okay: pick the one that's CLOSER to your life lately, most of the time. Go with your first gut answer." This instruction is part of the instrument; it is the release valve that makes forced choice tolerable when neither option fits well, and it sets the reference period (lately, most of the time) that all items share.

A respondent may go back and change an answer. When she does, the earlier pick is removed and the new one recorded, so that the 30 screens always contribute exactly 30 points in total. No name or email is requested before or during the questions, and the primary result is shown before any contact information is collected.

5. Scoring

All rules below are the rules implemented in instrument version 4. Nothing is weighted, hidden, or adjusted.

5.1 Domain scores

Each tap awards one point to the domain behind the chosen statement. The unchosen statement receives nothing. Each domain's raw score is its count across the 30 screens, range 0 to 10. For display, the raw score is multiplied by 10 and shown as a percentage (7 picks = 70 percent). Because each domain is scored out of its own maximum of 10, the six percentages do not, and are not expected to, sum to 100. The six raw scores always sum to 30.

5.2 Primary and runner-up

The primary domain is the one with the highest raw score. The runner-up is the second highest. The report presents all six domains ranked from highest to lowest, not in framework order, because the ranking is the result.

5.3 Blend rule

If the runner-up is within two points of the primary (a difference of 0, 1, or 2), the report presents the result as a blend of two domains. The primary domain still leads the report. This rule exists because a two-point gap on a 10-point scale, produced by 30 binary choices, is not a confident separation, and the instrument should not pretend otherwise.

5.4 Tie rule

If two domains tie for the highest score, they are resolved by head-to-head record. Those two domains met on exactly two screens (once under each stem). If one of them won both, it is the primary. If they split one and one, the result is treated as a full blend and the report leads with whichever domain comes first in framework order (Beliefs, Belonging, Boundaries, Business, Budget, Body), because that is the domain addressed earlier in the curriculum. Version 4 does not specify a separate rule for a three-way or larger tie; in that case the head-to-head rule is applied where it is decisive and framework order where it is not, and this is noted as a design gap to be closed in the next revision.

5.5 Convergence indicator

If the primary domain earned at least one point under Stem One and at least one point under Stem Two, the report states that the domain appeared "both in what's been heavy and in what you're hoping for." If the primary's points all came from one stem, the report shows nothing. This is the only place the two-stem design produces a visible output beyond the scores themselves, and it is deliberately modest: convergence increases confidence in the attribution, but its absence is not reported as a failure.

5.6 Pattern bands

Each domain's percentage is also assigned to one of three descriptive bands: 0 to 33 percent, "quiet for now"; 34 to 66 percent, "active"; 67 to 100 percent, "loud." The bands are a reading aid. They are not cutoffs and carry no clinical meaning.

5.7 Severity

Severity is the sum of the duration and impact items, range 2 to 8. A score of 2 to 4 is reported as Early-Stage; 5 to 8 is reported as Deep. The two bands determine which next step the report recommends: Early-Stage results are routed to a self-paced workbook, Deep results to a facilitated six-week program. Severity is independent of the domain scores; a respondent can be Early-Stage with a strongly differentiated profile, or Deep with a flat one.

5.8 Care flag

If the impact item is answered at its highest level ("I don't feel like myself at all anymore"), a care flag is set. The care flag does not change any score. It moves the care note (Section 8) from the bottom of the report to a highlighted position directly under the greeting.

5.9 Worked example

A respondent chooses Boundaries on 7 screens, Belonging on 6, Beliefs on 6, Body on 5, Budget on 4, and Business on 2 (total 30). Boundaries is primary at 70 percent. Belonging and Beliefs are tied for runner-up at 60 percent; they met on two screens and Belonging won both, so Belonging is the runner-up. The runner-up is within two points of the primary, so the blend rule fires and the report leads with Boundaries, then explains how Belonging feeds it. The respondent answered duration 3 ("Over a year") and impact 2 ("It shows up every week"): severity 5, Deep, and the report recommends the facilitated program. Impact was not 4, so the care flag is not set and the care note sits at the bottom of the report.

6. Exclusions

The instrument's credibility depends as much on what it refuses to do as on what it does. The following are explicit exclusions.

Not a diagnosis. The quiz does not diagnose burnout, depression, anxiety, or any other condition, and it does not produce a score that can be compared against a clinical threshold. No cutoff in this document should be read as one.

Not a depression screen. Given the documented overlap between burnout and depression (Bianchi, Schonfeld, and Laurent, 2015), the instrument does not attempt to distinguish the two. Its care note names the distinction the framework draws in practice, coaching for burnout and licensed therapy for depression, and directs respondents who feel hopeless, numb, or unable to function to a doctor or licensed counselor. The care flag (Section 5.8) is a placement rule for that note, not a screening result.

Not an occupational burnout measure. The instrument deliberately does not restrict itself to the workplace and therefore does not measure ICD-11 QD85 burn-out as the WHO defines it. Readers looking for an occupational measure should use the MBI, the OLBI, or the BAT.

Not a personality or trait measure. Every item is anchored to "lately" and "right now." The instrument is designed to describe a season, and a respondent's profile is expected to change as her circumstances change.

Excluded content. The item bank contains no items about trauma history, abuse, substance use, self-harm, or suicidal thinking. This is deliberate. A three-minute unsupervised quiz is not an appropriate place to elicit those disclosures, and the instrument does not pretend to respond to them.

Excluded populations. The instrument was written for adult women, with language drawn from the lives of Christian women who carry roles at home, at work, and in church. It has not been written for, tested with, or adapted to men, adolescents, or respondents outside that cultural and faith context. It is not offered to minors.

The unscored item. Screen 33 (body, emotions, spirit) is excluded from all scoring. It personalizes one paragraph of the report and nothing else.

7. Assumptions

The scoring in Section 5 is only as sound as the assumptions beneath it. They are listed here so they can be tested rather than taken on faith.

Equal item weight. Every choice screen contributes exactly one point. No item is assumed to be more diagnostic than another. This is the simplest defensible choice in the absence of data, and it is the first assumption a validation study should test.

Domain balance. Because each domain appears in exactly 10 screens against every other domain exactly twice, the design assumes that no domain has an unfair number of opportunities to be chosen. That balance is a property of the design and was verified by count.

Weight matching within pairs. The design assumes the two statements in each pair are roughly equally comfortable to admit. This was pursued by writing and review, not measured. If a pair is unbalanced, the safer statement's domain will be over-selected.

Honest, first-instinct responding. The instrument assumes the respondent answers about her own life, lately, most of the time, and without deliberating. There is no lie scale and no consistency check beyond the two-stem convergence indicator.

Shared reference period. All items assume the respondent applies the same window ("lately") throughout. The intensity block then asks about duration separately.

Exhaustiveness of the six domains. The design assumes that the source of a respondent's burnout falls within the six B's. A respondent whose actual source lies outside them will still receive a ranked profile, because forced choice always produces one. That profile would be an artifact of the design, not a finding.

Additive severity. Duration and impact are assumed to contribute equally and additively to severity, and the Early-Stage/Deep boundary between 4 and 5 is a judgment, not a derived threshold.

Language and reading. The instrument assumes fluency in the American English idiom its statements are written in (for example, "the car line," "the group chat"). It has not been translated or adapted.

Faith content does not affect scoring. The report uses scripture and ministry language. None of that language is present in the items, and none of it enters the scoring. A respondent who does not share the author's faith receives the same profile from the same answers; only the interpretive text around it is faith-based.

8. Care language and referral

Every full report carries the following note verbatim: "This quiz is a reflection tool, not a medical or psychological diagnosis. Ieshia will tell you this herself: coaching is for burnout, and licensed therapists are for depression. If you've been feeling hopeless, numb, or unable to function, please also talk to a doctor or licensed counselor. That's not weakness. That's stewardship of you."

Two referral rules are built into the framework and appear in the report where relevant. First, when the impact item is answered at its highest level, the care note is moved to the top of the report as a highlighted card. Second, when Belonging is the primary domain, the report advises the respondent to work with a licensed therapist alongside any coaching program, because in the author's practice this domain most often traces to a relational wound that coaching is not designed to treat.

9. Data handling

No personal information is collected before or during the 33 screens. The primary result and its explanation are shown without any form. A respondent who wants the full ranked report provides a first name and email address, after which the full report is displayed and emailed. Stored fields are the six domain percentages, the primary and runner-up domains, the severity score, the unscored presentation item, the care flag, the instrument version, and the date taken. A respondent who retakes the quiz overwrites her previous scores; earlier version tags are kept for record-keeping only. Data handling is governed by the site's privacy policy.

10. Limitations

Ipsative scores. This is the instrument's most important limitation and it is inherent to its design. A domain score of 70 percent means the respondent chose that domain on 7 of the 10 screens where it appeared. It does not mean she is "70 percent burned out" in that domain, and it cannot be compared to another respondent's 70 percent. The scores describe the shape of one person's profile and nothing else. Forced-choice instruments can be scored to recover normative information using item response models (Brown and Maydeu-Olivares, 2011, 2013), but that requires calibration data the instrument does not yet have.

No psychometric evidence. As of this version there is no published evidence of reliability (internal consistency, test-retest) or validity (structural, convergent, predictive) for this instrument. The domain structure comes from the author's coaching practice and has not been confirmed by factor analysis or any other statistical method. Claims that the quiz "works" are, at present, claims about design intent and practitioner experience, not measurement.

Coarse resolution. With 10 binary opportunities per domain, scores move in 10-point steps and the primary and runner-up are frequently within the two-point blend window. The instrument is built to accept this and report blends honestly rather than to force a single label, but a longer instrument would resolve more finely.

Two-item severity. Duration and impact are single items each. The severity score is a rough sort, not a measure.

Single-author item pool. All items were written and mapped by the framework author with editorial support. There has been no independent expert review of item-to-domain mapping and no cognitive interviewing with respondents.

Self-selected respondents. People who take a free burnout quiz on a coaching site are, by definition, people who already suspect they are burned out and who found this site. Any future response data will describe that population, not women in general.

Response effects not yet tested. Order effects, fatigue across 33 screens, and the effect of the halfway transition screen have not been measured.

Convergence indicator is weak by design. The two-stem check reports only whether the primary earned at least one point under each stem. It is a floor, not a statistic.

Retake behavior. Because a retake overwrites prior scores, the instrument cannot currently support within-person change analysis without a separate data design.

Previous versions. Versions 2 and 3 used rating scales and different item sets. Results from those versions are not comparable to version 4 and are not included in any future analysis.

11. Validation roadmap

The author intends to publish the following when the data exist, in this order, and to update this document's version number when she does.

• Descriptive distribution: the distribution of primary domains, blend frequency, severity bands, and care-flag rate across all version 4 completions, with the completion count stated.
• Design checks: the rate at which the primary domain converges across both stems; the frequency of ties; the frequency with which each of the 30 pairs is answered in each direction, which is the direct test of the weight-matching assumption in Section 7.
• Reliability: test-retest agreement for primary domain and severity band over a short interval, using a separate retake path that preserves prior scores.
• Convergent validity: correlation between the severity score and an established, freely available burnout measure such as the OLBI or the BAT, administered to a volunteer subsample. The MBI is a licensed instrument and would require a separate arrangement.
• Structural checks: whether the six-domain structure is supported by the pattern of choices, using methods appropriate to forced-choice data.
• Expert review: independent review of the item-to-domain mapping by at least one practitioner outside the author's organization.

Until the first of these is published, the correct description of this instrument is "a structured reflection tool with a documented design," and that is the only description the author will make.

12. How to cite this document

McArthur, I. (2026). Where Is Your Burnout Coming From? Design, Scoring, and Limitations of a Six-Domain Burnout Source Inventory (Version 1.0). iamieshiadanielle.com. Retrieved from https://iamieshiadanielle.com/burnout-quiz-methodology

Readers are welcome to link to, quote from, and critique this document. Corrections and challenges to the design are invited through the contact page on the site and will be acknowledged in subsequent versions.

References

• Bianchi, R., Schonfeld, I. S., and Laurent, E. (2015). Burnout-depression overlap: A review. Clinical Psychology Review, 36, 28-41. https://doi.org/10.1016/j.cpr.2015.01.004
• Brown, A., and Maydeu-Olivares, A. (2011). Item response modeling of forced-choice questionnaires. Educational and Psychological Measurement, 71(3), 460-502. https://doi.org/10.1177/0013164410375112
• Brown, A., and Maydeu-Olivares, A. (2013). How IRT can solve problems of ipsative data in forced-choice questionnaires. Psychological Methods, 18(1), 36-52. https://doi.org/10.1037/a0030641
• Demerouti, E., Bakker, A. B., Vardakou, I., and Kantas, A. (2003). The convergent validity of two burnout instruments: A multitrait-multimethod analysis. European Journal of Psychological Assessment, 19(1), 12-23. https://doi.org/10.1027/1015-5759.19.1.12
• Freudenberger, H. J. (1974). Staff burn-out. Journal of Social Issues, 30(1), 159-165. https://doi.org/10.1111/j.1540-4560.1974.tb00706.x
• Hontangas, P. M., de la Torre, J., Ponsoda, V., Leenen, I., Morillo, D., and Abad, F. J. (2015). Comparing traditional and IRT scoring of forced-choice tests. Applied Psychological Measurement, 39(8), 598-612. https://doi.org/10.1177/0146621615585851
• Leiter, M. P., and Maslach, C. (1999). Six areas of worklife: A model of the organizational context of burnout. Journal of Health and Human Services Administration, 21(4), 472-489.
• Maslach, C., and Jackson, S. E. (1981). The measurement of experienced burnout. Journal of Occupational Behaviour, 2(2), 99-113. https://doi.org/10.1002/job.4030020205
• Maslach, C., and Leiter, M. P. (1997). The truth about burnout: How organizations cause personal stress and what to do about it. Jossey-Bass.
• Poghosyan, L., Aiken, L. H., and Sloane, D. M. (2009). Factor structure of the Maslach Burnout Inventory: An analysis of data from large scale cross-sectional surveys of nurses from eight countries. International Journal of Nursing Studies, 46(7), 894-902. https://doi.org/10.1016/j.ijnurstu.2009.03.004
• Reis, D., Xanthopoulou, D., and Tsaousis, I. (2015). Measuring job and academic burnout with the Oldenburg Burnout Inventory (OLBI): Factorial invariance across samples and countries. Burnout Research, 2(1), 8-18. https://doi.org/10.1016/j.burn.2014.11.001
• Roskam, I., Brianda, M.-E., and Mikolajczak, M. (2018). A step forward in the conceptualization and measurement of parental burnout: The Parental Burnout Assessment (PBA). Frontiers in Psychology, 9, 758. https://doi.org/10.3389/fpsyg.2018.00758
• Schaufeli, W. B., Desart, S., and De Witte, H. (2020). Burnout Assessment Tool (BAT): Development, validity, and reliability. International Journal of Environmental Research and Public Health, 17(24), 9495. https://doi.org/10.3390/ijerph17249495
• World Health Organization. (2019, May 28). Burn-out an "occupational phenomenon": International Classification of Diseases. https://www.who.int/news/item/28-05-2019-burn-out-an-occupational-phenomenon-international-classification-of-diseases