Nottingham Health Profile (NHP)

1. Abstract

The Nottingham Health Profile (NHP) is an influential, patient-reported outcome measure (PRO) specifically designed to assess perceived physical, emotional, and social health problems. Developed in the late 1970s and early 1980s by Sonja M. Hunt, James McEwen, and Stephen P. McKenna at the University of Nottingham, the NHP emerged in response to the growing recognition that traditional objective clinical indicators (such as laboratory values, physiological markers, and radiographic findings) fail to capture the subjective burden of disease experienced by individuals in their daily lives. The instrument comprises two distinct sections totaling 45 items evaluated through a dichotomous (Yes/No) response format. Part I includes 38 subjective statements categorized across six fundamental dimensions of health-related quality of life (HRQoL): Energy Level (3 items), Pain (8 items), Emotional Reactions (9 items), Sleep (5 items), Social Isolation (5 items), and Physical Abilities (8 items). Part II contains 7 items evaluating the specific impact of health status on everyday life domains: paid employment, home maintenance, social life, home life/personal relationships, sex life, interests and hobbies, and holidays.

Psychometrically, the NHP relies on item weights derived through Thurstone’s method of paired comparisons, reflecting the relative perceived severity of each problem within its respective domain. Domain scores range from 0 (indicating no perceived health problems or distress) to 100 (indicating maximum distress and severe functional limitation). Extensive psychometric evaluations demonstrate robust internal consistency across subscales, high test-retest reliability (Pearson’s r and ICC commonly exceeding .75 to .88), clear convergent validity with clinical morbidity indices and other multidimensional instruments such as the Short Form-36 (SF-36) and Sickness Impact Profile (SIP), as well as distinct discriminant validity across healthy and clinical populations. The NHP remains one of the pioneer generic instruments used internationally across clinical trials, public health surveys, epidemiologic surveillance, and health economics evaluations.

2. Keywords

Nottingham Health Profile, NHP, health-related quality of life, patient-reported outcome measures, subjective health status, functional status, physical disability, emotional distress, Thurstone scaling, psychometrics.

3. Authors

The Nottingham Health Profile was developed by an interdisciplinary team of medical sociologists, epidemiologists, and community medicine specialists based at the Department of Community Health, University of Nottingham Medical School, United Kingdom:

• Sonja M. Hunt, PhD: Medical Sociologist and Research Fellow, Department of Community Health, University of Nottingham; later associated with the Research Unit in Health and Behavioural Change, University of Edinburgh, Scotland. Dr. Hunt served as the lead investigator in conceptualizing perceived health status and designing the initial qualitative elicitation protocols.
• James McEwen, MB ChB, FFCM: Professor and Consultant in Public Health Medicine, Department of Community Health, University of Nottingham; subsequently Professor of Public Health at the University of Glasgow. Dr. McEwen contributed epidemiological methodology and clinical integration frameworks.
• Stephen P. McKenna, PhD: Psychometrician and Researcher, Department of Community Health, University of Nottingham; later Director of Galen Research and Consultancy, Manchester, United Kingdom. Dr. McKenna spearheaded the Thurstonian paired-comparison weighting methodologies and international adaptations.
• Janice Williams: Statistician and Research Officer, Department of Community Health, University of Nottingham.

4. Purpose

The fundamental purpose of the Nottingham Health Profile is to provide a standardized, scientifically rigorous, and clinically meaningful metric of subjective health status that captures the patient’s own perspective of their health impairments. In mid-twentieth-century medicine, evaluation of clinical interventions relied almost entirely on biomedical endpoints such as survival rates, biological markers, blood pressure values, and radiographic assessments. However, these physiological indicators frequently correlated poorly with a patient’s capacity to function in daily life, their psychological well-being, or their subjective suffering. The NHP was designed to bridge this gap by quantifying perceived distress and disability across multiple dimensions of human experience.

The instrument was engineered to serve multiple distinct research and clinical functions:

• Epidemiological Surveillance and Needs Assessment: The NHP was originally designed as a brief population survey instrument to assess the prevalence of subjective health problems within community-dwelling populations, enabling public health officials to identify unmet healthcare needs, monitor socio-demographic health disparities, and allocate regional resources effectively.
• Clinical Outcome Evaluation: In clinical settings and randomized clinical trials, the NHP measures treatment outcomes from the patient’s vantage point. It detects changes following major medical and surgical interventions, such as total hip replacement, cardiac revascularization, renal transplantation, and chronic disease management programs.
• Medical Auditing and General Practice: The profile aids general practitioners and outpatient clinicians in systematically identifying hidden functional and psychological difficulties that patients might not spontaneously articulate during routine medical consultations.
• Health Economics and Program Evaluation: By assessing both physical and psychosocial distress, the NHP provides comprehensive profile data that inform comparative effectiveness research, health service planning, and resource utilization evaluations.

The theoretical rationale of the NHP emphasizes that perceived health is fundamentally distinct from physiological pathology. Two patients with identical clinical grades of rheumatoid arthritis, for example, may experience vastly different levels of pain, emotional distress, and functional limitation depending on their coping mechanisms, social support, and environment. The NHP focuses on these experiential consequences of health problems, ensuring that healthcare evaluations reflect genuine human suffering and impairment rather than biological measurements alone.

5. Psychological Construct

The Nottingham Health Profile operationalizes health-related quality of life as a multidimensional construct encompassing subjective physical, emotional, and social limitations. Rather than measuring objective performance capacity in an ideal environment, the NHP measures perceived dysfunction and distress in everyday life. Part I evaluates six primary dimensions:

Energy Level (EL)

This subscale assesses subjective vitality, vigor, and the presence of debilitating physical exhaustion. It captures persistent lethargy and asthenia that are disproportionate to exertion. It includes 3 items assessing whether the individual is constantly fatigued (“I’m tired all the time”), perceives daily tasks as overwhelming (“Everything is an effort”), or experiences rapid depletion of physical stamina (“I soon run out of energy”). In clinical research, this dimension is sensitive to systemic illness, cardiac insufficiency, chronic inflammatory conditions, and depressive states.

Pain (P)

The Pain dimension comprises 8 items evaluating both the subjective intensity of bodily discomfort and its interference with basic static and dynamic postures. It differentiates between nocturnal pain (“I have pain at night”), extreme severity (“I have unbearable pain”), unremitting distress (“I’m in constant pain”), and postural pain provoked by walking, standing, sitting, changing position, or climbing stairs. This dimension is central to evaluating orthopedic, musculoskeletal, and oncological disorders.

Emotional Reactions (ER)

Reflecting the psychological domain, this 9-item subscale measures feelings of demoralization, psychological vulnerability, dysphoria, affective lability, and existential despair. Items probe depressive mood (“Things are getting me down”, “I wake up feeling depressed”, “I feel that life is not worth living”), irritability and anger (“I lose my temper easily these days”), anhedonia (“I’ve forgotten what it’s like to enjoy myself”), psychic tension (“I’m feeling on edge”), cognitive temporality (“The days seem to drag”), and perceived loss of self-regulation (“I feel as if I’m losing control”). It reflects secondary psychological reactions to chronic illness as well as primary emotional distress.

Sleep (S)

Comprising 5 items, this dimension assesses subjective sleep architecture disruption, insomnia, and nocturnal rumination. The items evaluate sleep onset latency (“It takes me a long time to get to sleep”), early morning awakenings (“I’m waking up in the early hours of the morning”), sleep maintenance problems (“I lie awake for most of the night”), cognitive hyperarousal (“Worry is keeping me awake at night”), overall sleep dissatisfaction (“I sleep badly at night”), and pharmacologic dependence (“I take tablets to help me sleep”). Sleep disturbances captured here often mediate the interaction between physical pain and emotional distress.

Social Isolation (SI)

The Social Isolation dimension (5 items) evaluates perceived interpersonal disconnection, alienation, and communicative friction. Rather than measuring the objective size of an individual’s social network, it gauges subjective loneliness (“I feel lonely”), lack of intimate social attachment (“I feel there is nobody I am close to”), relational distress and friction (“I’m finding it hard to get on with people”), social withdrawal (“I’m finding it hard to make contact with people”), and perceived interpersonal burdensomeness (“I feel I am a burden to people”). This dimension captures the profound social withdrawal that often accompanies chronic physical impairment.

Physical Abilities (PA)

This 8-item subscale evaluates mobility impairments and limitations in fundamental activities of daily living (ADLs). Items address functional ambulation (“I’m unable to walk at all”, “I can only walk about indoors”, “I need help to walk about outside”), transfer and movement challenges (“I have trouble getting up and down stairs and steps”, “I find it hard to bend”, “I find it hard to reach for things”), static endurance (“I find it hard to stand for long”), and self-care independence (“I find it hard to dress myself”). It quantifies the degree of physical restriction experienced in one’s personal habitat.

Part II: Specific Life Areas

Part II supplements the six core dimensions by assessing whether current health problems directly disrupt 7 distinct life roles: paid employment (work), domestic management (looking after the home), social interaction (social life), family relationships (home life), intimate partnerships (sex life), recreational activities (interests and hobbies), and leisure travel (holidays). This section isolates specific contextual role handicaps resulting from physical and emotional impairment.

6. Theoretical Framework

The Nottingham Health Profile is rooted in medical sociology, the biopsychosocial model of medicine formulated by George Engel, and the phenomenological tradition of subjective health status measurement. Historically, health was operationalized primarily through mortality and morbidity indices. However, in 1948, the World Health Organization (WHO) redefined health as “a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity.” The NHP operationalized this definition into a psychometrically sound, standardized self-report metric.

A critical foundational theory underlying the NHP is the conceptual distinction between disease, impairment, disability, and handicap, as formalized by Philip Wood and subsequently adopted in the WHO’s International Classification of Impairments, Disabilities, and Handicaps (ICIDH). While traditional clinical measures capture biological pathology (disease) or structural dysfunction (impairment), the NHP focuses on perceived restriction in functional activity (disability) and the resulting disadvantage in fulfilling social roles (handicap). The instrument explicitly conceptualizes health not as an absolute, objective clinical phenomenon, but as a perceived, lived reality mediated by individual expectations, social roles, and psychological appraisal.

Methodologically, the construction of the NHP was governed by two theoretical and psychometric principles:

1. Lay-Derived Item Generation: Unlike many contemporary instruments constructed deductively by physicians and clinical experts, the NHP used an inductive, patient-centered methodology. Statements were extracted from hundreds of unstructured, qualitative depth interviews conducted with primary care patients, chronically ill individuals, and members of the general public. Over 2,200 unique patient statements detailing the subjective impact of illness were collected and categorized, ensuring that the vocabulary, phrasing, and cognitive framing of items reflected authentic lay language rather than professional medical jargon.
2. Thurstone’s Law of Comparative Judgment: To score the items within each domain, the developers rejected arbitrary Likert weighting and applied Louis Leon Thurstone’s method of paired comparisons. Items within each subscale were presented in pairs to broad cross-sections of the general population and clinical samples, who judged which statement of the pair represented a more severe health problem. Mathematical scaling converted these paired preferences into interval-level weights, ensuring that severe impairments (e.g., “I’m unable to walk at all”) receive substantially higher numerical penalties than milder difficulties (e.g., “I find it hard to bend”).

7. Validity

The psychometric validity of the Nottingham Health Profile has been substantiated across hundreds of empirical investigations involving community samples, primary care populations, and diverse clinical cohorts spanning rheumatology, cardiology, neurology, pulmonary medicine, oncology, and geriatrics.

Construct and Convergent Validity

Construct validity was established through systematic correlations with physiological indices, physician assessments, functional status metrics, and health service utilization patterns. In validation studies by Hunt, McKenna, McEwen, and colleagues (1980, 1981), NHP scores demonstrated strong, statistically significant correlations with objective morbidity markers. For example, patients presenting to primary care with acute or chronic illness demonstrated significantly elevated scores across all six dimensions compared to healthy community controls ($p < .001$). Furthermore, domain scores correlated systematically with the frequency of general practitioner consultations, prescription medication volume, and rates of certified sickness absenteeism from employment.

Convergent validity has been established through multi-trait multi-method comparisons with other validated health status instruments. NHP dimensions correlate moderately to strongly with corresponding subscales of the SF-36: the NHP Physical Abilities subscale correlates strongly with the SF-36 Physical Functioning scale ($r = .70$ to $.82$), the NHP Pain subscale correlates with the SF-36 Bodily Pain scale ($r = .74$ to $.85$), and the NHP Emotional Reactions subscale correlates with both the SF-36 Mental Health dimension ($r = -.65$ to $-.78$) and the General Health Questionnaire (GHQ) ($r = .62$ to $.71$).

Discriminant and Known-Groups Validity

The NHP demonstrates known-groups validity, distinguishing between groups known to differ in health status:

• Age and Gender Stratification: In large-scale population surveys, NHP domain scores—particularly Physical Abilities, Pain, and Energy Level—systematically increase across advancing age cohorts, reflecting age-associated biological degeneration and comorbidity prevalence. Females consistently report slightly higher distress on Pain, Sleep, and Emotional Reactions than males, matching recognized epidemiological distributions.
• Disease Severity Gradients: Among patients with peripheral arterial occlusive disease, chronic obstructive pulmonary disease (COPD), congestive heart failure, and osteoarthritis, NHP dimension scores track functional staging systems (such as the New York Heart Association functional classification and Fontaine stages for claudication) with high fidelity.

Criterion and Predictive Validity

NHP scores demonstrate strong predictive utility regarding downstream health outcomes. Prospective investigations have revealed that elevated distress on the NHP—specifically high scores on Energy Level, Physical Abilities, and Emotional Reactions—serves as an independent predictor of long-term mortality, emergency hospital readmission, failure to return to employment following major surgical interventions, and subsequent nursing home admission among older adults.

8. Reliability

The reliability of the Nottingham Health Profile has been documented in both cross-sectional and longitudinal studies across diverse cultural and language translations.

Test-Retest Reliability

Given that health status naturally fluctuates, evaluating test-retest reliability requires assessing medically stable patients over intervals short enough to preclude genuine biological change, yet long enough to minimize recall bias (typically 24 hours to 4 weeks). In the benchmark reliability study conducted by Hunt, McKenna, and Williams (1981) among patients with stable osteoarthrosis across a 4-week retest interval, Pearson correlation coefficients ($r$) and Spearman rank correlations for the six dimensions demonstrated stability:

• Energy Level: $r = .77$ to $.85$
• Pain: $r = .86$ to $.91$
• Emotional Reactions: $r = .77$ to $.83$
• Sleep: $r = .85$ to $.89$
• Social Isolation: $r = .71$ to $.79$
• Physical Abilities: $r = .85$ to $.88$

Subsequent international validation studies evaluating European language adaptations (French, Spanish, Swedish, German, Italian, and Dutch) confirmed stability, reporting Intraclass Correlation Coefficients (ICC) consistently exceeding $.80$ for most dimensions among stable chronic populations.

Internal Consistency

Because the NHP uses dichotomous items scaled via Thurstone’s method rather than classic Likert summative scaling, traditional internal consistency coefficients (such as Cronbach’s alpha or the Kuder-Richardson Formula 20 [KR-20]) reflect both item homogeneity and the severity distribution of items across the underlying trait continuum. Across published studies, Cronbach’s alpha values typically fall between:

• Physical Abilities: $\alpha = .78 – .86$
• Pain: $\alpha = .80 – .90$
• Emotional Reactions: $\alpha = .75 – .84$
• Sleep: $\alpha = .72 – .81$
• Social Isolation: $\alpha = .65 – .78$
• Energy Level: $\alpha = .62 – .75$ (the lower value reflecting the 3-item length)

Guttman split-half coefficients and item-total correlations across the dimensions generally confirm acceptable to high internal homogeneity, satisfying psychometric standards for group-level comparative research.

9. Factor Analysis

The dimensional structure of the Nottingham Health Profile has been examined through both exploratory (EFA) and confirmatory (CFA) factor analyses, as well as modern Item Response Theory (IRT) and Rasch measurement models.

Exploratory Factor Analyses

Early factor analytic evaluations utilizing principal component analysis with varimax and oblimin rotations confirmed that the 38 items of Part I load onto factors corresponding to the six empirically designated subscales. Items addressing functional movement (e.g., walking, stair climbing, bending) load onto a primary Physical Abilities factor (loadings typically ranging from $.52$ to $.84$). Bodily distress items load onto a Pain factor (loadings $.48$ to $.81$). Dysphoria, anhedonia, and affective distress load onto Emotional Reactions (loadings $.45$ to $.79$). Nocturnal awakenings and sleep architecture items load onto Sleep (loadings $.55$ to $.86$).

Factor analyses have also revealed higher-order structures. In several large clinical and population studies, the six dimensions condense into two major second-order meta-dimensions: a Physical Health Dimension (dominated by Physical Abilities, Pain, and Energy Level) and a Psychosocial Health Dimension (dominated by Emotional Reactions, Social Isolation, and Sleep). This aligns with the structural dualism found in other instruments, such as the SF-36 Physical and Mental Component Summary scores.

Confirmatory Factor Analysis and Structural Equation Modeling

Confirmatory factor analyses testing the oblique six-factor model have demonstrated acceptable to good model fit indices across diverse medical populations:

• Comparative Fit Index (CFI): typically ranges between $.91$ and $.96$
• Tucker-Lewis Index (TLI): $.90$ to $.95$
• Root Mean Square Error of Approximation (RMSEA): $.045$ to $.062$ ($90%\text{ CI } [.040, .068]$)
• Standardized Root Mean Square Residual (SRMR): $.048$ to $.065$

Inter-factor correlations in CFA models reveal moderate to high covariance between specific dimensions, notably between Pain and Physical Abilities ($r = .55 – .72$), and between Emotional Reactions and Social Isolation ($r = .58 – .74$). Conversely, correlations between Physical Abilities and Social Isolation are typically lower ($r = .28 – .42$), demonstrating divergence between physical impairment and social estrangement.

Rasch and Mokken Scaling Analyses

Modern psychometric critiques applying the Rasch unidimensional measurement model have highlighted both strengths and structural nuances of the NHP. While several subscales (such as Physical Abilities and Pain) satisfy Rasch criteria for unidimensionality and invariant item ordering, researchers have noted that the dichotomous items exhibit “floor effects” when applied to relatively healthy populations. Because items were selected to capture pronounced distress and dysfunction (e.g., “I’m unable to walk at all”, “I feel that life is not worth living”), the instrument functions with high precision among clinical populations with moderate to severe illness, but exhibits less variance among healthy, young community cohorts.

10. Instrument / Measurement Tool

• Instrument Name: Nottingham Health Profile (NHP).
• Instrument Type: Standardized generic patient-reported outcome measure (PRO) / health-related quality of life (HRQoL) survey.
• Administration Mode: Self-administered paper-and-pencil questionnaire, face-to-face structured interview, telephone interview, or computerized/electronic administration (ePRO).
• Target Population: Adults (16+ years) in general community populations, outpatient clinics, and hospitalized medical/surgical cohorts.
• Completion Time: Approximately 5 to 10 minutes.
• Structure: 45 total items divided into two independent sections:
  • Part I: 38 subjective health statements across 6 core dimensions:
    • Energy Level (EL): 3 items (Items 1, 12, 26)
    • Pain (P): 8 items (Items 2, 4, 8, 19, 24, 28, 36, 38)
    • Emotional Reactions (ER): 9 items (Items 3, 6, 7, 16, 20, 23, 31, 32, 37)
    • Sleep (S): 5 items (Items 5, 13, 22, 29, 33)
    • Social Isolation (SI): 5 items (Items 9, 15, 21, 30, 34)
    • Physical Abilities (PA): 8 items (Items 10, 11, 14, 17, 18, 25, 27, 35)
  • Part II: 7 items assessing health-related problems in specific areas of life (Items 39–45: paid employment, home tasks, social life, home relationships, sex life, hobbies, holidays).
• Response Format: Simple dichotomous forced-choice format: Yes / No. Respondents indicate whether each statement applies to them “at the present time” / “these days”.
• Scoring Methodology:
  • Unweighted Scoring: Raw domain scores can be calculated by summing the affirmative (“Yes”) responses within each subscale, yielding an index of the total number of perceived problems.
  • Weighted Scoring (Standard Thurstonian Method): Each item within Part I is assigned a predetermined Thurstonian mathematical weight derived from paired-comparison judgments. When a participant endorses an item (“Yes”), that item’s specific weight is credited; negative responses (“No”) receive a score of zero. The weights within each subscale sum to 100. Thus, each dimension yields a final score between 0 and 100, where:
    • 0: Indicates the absence of perceived health problems or distress in that domain (optimal perceived health).
    • 100: Indicates that all problems in that domain are endorsed (maximum distress / severe limitation).
  • Part II Scoring: Responses are scored as the proportion or absolute number of life areas disrupted by health (ranging from 0 to 7 affirmative endorsements), or analyzed as individual categorical binary outcomes.
  • No Single Global Index: Combining all 38 items of Part I into a single aggregate score is discouraged by the original authors because distinct dimensions represent qualitatively different aspects of human functioning; domain-specific profile scoring is standard practice.

11. Permissions & Fee and Test Year

Test Publication Year: The Nottingham Health Profile was initially developed and validated between 1979 and 1980, with foundational validation papers published in 1980 and 1981 by Sonja M. Hunt, James McEwen, and Stephen P. McKenna. The revised user’s manual was formally released in 1989 and 1991.

Copyright and Licensing: The Nottingham Health Profile is protected by copyright. Original rights were held by the research developers at the University of Nottingham and subsequent administrative entities (including Galen Research and Consultancy, Manchester, UK). The European versions and cross-cultural adaptations were harmonized under the auspices of the European Group for the Quality of Life and Health Measurement.

Academic and Clinical Use: The NHP is widely accessible for non-commercial academic research, public health surveillance, student theses, and non-funded clinical practice. However, researchers, healthcare organizations, and pharmaceutical sponsors conducting funded clinical trials or commercial studies must obtain formal licensing, scoring manuals, official translation certifications, and written user agreements prior to implementation. Requests for commercial licensing, standardized weighted scoring algorithms, and validated linguistic translations are managed through official copyright representatives and academic health outcome measurement repositories (such as Galen Research or authorized PRO distribution platforms).

12. References

• Dehdari, T., Hashemifard, T., Heidarnia, A., & Kazemnejad, A. (2005). The longitudinal effect of health education on health-related quality of life in patients with coronary artery bypass surgery. Faculty of Health, Tarbiat Modarres University, 15, 41–46.
• European Group for the Quality of Life and Health Measurement. (1992). European guide to the Nottingham Health Profile. Montpellier: The European Group for the Quality of Life and Health Measurement.
• Hunt, S. M., & McEwen, J. (1980). The development of a subjective health indicator. Sociology of Health & Illness, 2(3), 231–246. https://doi.org/10.1111/1467-9566.ep11340728
• Hunt, S. M., McEwen, J., & McKenna, S. P. (1984). Subjective health assessments and the perceived outcome of minor surgery. Journal of Psychosomatic Research, 28(2), 105–114. https://doi.org/10.1016/0022-3999(84)90004-9
• Hunt, S. M., McEwen, J., & McKenna, S. P. (1985). Measuring health status: A new tool for clinicians and epidemiologists. The Journal of the Royal College of General Practitioners, 35(273), 185–188. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1960256/
• Hunt, S. M., & McKenna, S. P. (1991). The Nottingham Health Profile user’s manual, revised. Manchester: Galen Research and Consultancy.
• Hunt, S. M., McKenna, S. P., & McEwen, J. (1989). The Nottingham Health Profile: User’s manual (Rev. ed.). Manchester: Department of Community Medicine, University of Nottingham.
• Hunt, S. M., McKenna, S. P., McEwen, J., Backett, E. M., Williams, J., & Papp, E. (1980). A quantitative approach to perceived health status: A validation study. Journal of Epidemiology & Community Health, 34(4), 281–286. https://doi.org/10.1136/jech.34.4.281
• Hunt, S. M., McKenna, S. P., McEwen, J., Williams, J., & Papp, E. (1981). The Nottingham Health Profile: Subjective health status and medical consultations. Social Science & Medicine. Part A: Medical Psychology & Medical Sociology, 15(3), 221–229. https://doi.org/10.1016/0271-7123(81)90005-5
• Hunt, S. M., McKenna, S. P., & Williams, J. (1981). Reliability of a population survey tool for measuring perceived health problems: A study of patients with osteoarthrosis. Journal of Epidemiology & Community Health, 35(4), 297–300. https://doi.org/10.1136/jech.35.4.297
• McDowell, I. (2006). Measuring health: A guide to rating scales and questionnaires (3rd ed.). Oxford: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195165678.001.0001
• McKenna, S. P., Hunt, S. M., & McEwen, J. (1981). Weighting the seriousness of perceived health problems using Thurstone’s method of paired comparisons. International Journal of Epidemiology, 10(1), 93–97. https://doi.org/10.1093/ije/10.1.93

13. Items of the Scale (Questionnaire)