Duke Health Profile (Duke-PH)

Abstract

The Duke Health Profile (DUKE), often designated in clinical epidemiology and health services research as the Duke-PH, is a 17-item, patient-reported outcome measure (PROM) designed to evaluate generic health-related quality of life across adult populations in primary care, clinical trials, and epidemiological investigations. Developed by George R. Parkerson Jr. and his colleagues at the Department of Community and Family Medicine at Duke University Medical Center, the instrument represents an empirically derived, refined short-form adaptation of the earlier 63-item Duke-UNC Health Profile (DUHP). The DUKE operationalizes health according to the World Health Organization multidimensional conceptual model, capturing functioning along biological, psychological, and social axes. It generates scores across eleven discrete, validated dimensions: six positive health measures (physical health, mental health, social health, general health, perceived health, and self-esteem) and five health dysfunction measures (anxiety, depression, combined anxiety-depression, pain, and disability). Each subscale is transformed linearly to a standardized metric ranging from 0 to 100, wherein higher scores in the health measures denote superior functional capacity and well-being, while higher scores in dysfunction dimensions reflect greater functional impairment and symptom severity. Administered through a brief three-point polytomous response architecture across diverse temporal frames (concurrent/today, past week), the profile requires approximately three to five minutes for completion. Psychometric evaluations demonstrate robust convergent and discriminant validity against established health status measures (including the SF-36, Sickness Impact Profile, and Beck Depression Inventory), acceptable-to-strong internal consistency reliability (Cronbach’s alpha coefficients typically ranging from 0.55 to 0.83 across multidimensional subscales), and high test-retest reproducibility (intraclass correlation coefficients and Pearson correlations between 0.65 and 0.82). The instrument has demonstrated cross-cultural invariance across dozens of validated international translations, affirming its utility as a brief, standardized assessment tool for ambulatory medicine, population health monitoring, and economic health evaluation.

Keywords

Duke Health Profile, DUKE, Duke-PH, health-related quality of life, patient-reported outcome measures, primary care psychometrics, functional health status, multidimensional health assessment, ambulatory care evaluation, health dysfunction

Authors

The Duke Health Profile was developed by an interdisciplinary team of clinical epidemiologists, family physicians, and biostatisticians affiliated with the Department of Community and Family Medicine at Duke University Medical Center, Durham, North Carolina, USA.

• George R. Parkerson Jr., MD, MPH: Professor Emeritus and former Chair of the Department of Community and Family Medicine, Duke University School of Medicine. Principal investigator responsible for the conceptual design, longitudinal psychometric testing, and clinical deployment of the Duke Health Measures suite.
• W. E. Broadhead, MD, PhD: Associate Professor of Community and Family Medicine and clinical epidemiologist, Duke University Medical Center; investigator specializing in psychosocial determinants of primary health care outcomes, social support operationalization, and depression screening.
• C. K. J. Tse, MPH: Senior biostatistician and data analyst, Department of Community and Family Medicine, Duke University Medical Center; co-developer involved in structural equation modeling, scaling transformations, and factorial validation of the 17-item profile.

Institutional contact for licensing and academic inquiry: Department of Community and Family Medicine, Duke University Medical Center, Box 2914, Durham, NC 27710, USA.

Purpose

The primary clinical and epidemiological objective of the Duke Health Profile (DUKE) is to provide a brief, psychometrically sound, patient-completed measurement tool capable of evaluating the broad spectrum of adult functional health status and health-related quality of life within high-volume primary care and ambulatory research settings. Prior to its construction, most established health status measures—such as the 136-item Sickness Impact Profile (SIP), the Nottingham Health Profile (NHP), and the 63-item Duke-UNC Health Profile—imposed significant respondent burden, requiring between 15 and 45 minutes to administer. Such length rendered routine clinical implementation and high-throughput epidemiological tracking impractical. Parkerson and colleagues sought to develop an ultra-efficient instrument that minimized patient burden while retaining multidimensional sensitivity, enabling clinicians and investigators to evaluate both positive health attributes and disabling pathological symptoms.

Clinically, the DUKE serves three primary purposes: screening, treatment monitoring, and outcome prediction. As an ambulatory screening instrument, the profile identifies latent psychological morbidity—specifically depressive and anxiety spectra—that frequently present somatically in primary care medicine. By capturing physical disability, pain severity, sleep disturbances, fatigue, and social engagement simultaneously, the instrument alerts primary care providers to hidden vulnerabilities that may complicate medical adherence and recovery. In longitudinal disease management, the DUKE is utilized to monitor functional changes in patients diagnosed with chronic degenerative or inflammatory conditions, including type 2 diabetes, coronary artery disease, chronic obstructive pulmonary disease, osteoarthritis, and fibromyalgia. Its 0 to 100 standardized scoring scale provides quantitative indices that allow clinicians to track therapeutic efficacy, postoperative functional recovery, and rehabilitation progress over defined temporal intervals.

In health services research and clinical epidemiology, the profile is widely employed as a primary or secondary endpoint in randomized clinical trials, comparative effectiveness research, and population-level health surveys. Health economists utilize DUKE scores to track generic utility decrements associated with multimorbidity and to evaluate the cost-effectiveness of psychosocial, pharmaceutical, or surgical interventions. Furthermore, the instrument provides an empirical bridge between biological disease severity and subjective illness experience, permitting researchers to assess how socioeconomic determinants, psychological resilience, and social support networks modulate functional impairment.

Psychological Construct

The Duke Health Profile operationalizes health as a complex, non-orthogonal, multidimensional construct encompassing eleven specific dimensions. These dimensions are conceptually divided into six health measures and five dysfunction measures, acknowledging that the presence of health is not merely the absence of disease, nor is dysfunction solely the inverse of well-being.

Positive Functional Health Dimensions

• Physical Health (Items 8, 9, 10, 11, 12): Reflects biological integrity, physiological endurance, and basic ambulatory mobility. This subscale measures the absence of somatic limitations when executing foundational physical tasks (such as ascending stairs or running a football field) alongside the absence of systemic somatic disruptions, such as severe pain, sleep fragmentation, and exertional fatigue.
• Mental Health (Items 1, 4, 5, 13, 14): Represents emotional equilibrium, cognitive stability, and psychological resilience. It evaluates positive self-appraisal, persistence in goal-directed activity, attentional focus, and the relative absence of debilitating sadness and autonomic nervousness.
• Social Health (Items 2, 6, 7, 15, 16): Quantifies interpersonal functioning, socialization frequency, and integration within community networks. This dimension measures communicative ease with others, satisfaction with familial dynamics, interpersonal comfort, and active participation in civic, religious, or recreational activities.
• General Health: A composite overarching index derived as the unweighted arithmetic mean of the Physical Health, Mental Health, and Social Health subscales. It embodies an integrated metric of total functional equilibrium across biological, psychological, and social spheres.
• Perceived Health (Item 3): A single-item subjective global assessment measuring the respondent’s cognitive appraisal of their somatic resilience and constitutional vitality (“I am basically a healthy person”). This dimension serves as a direct proxy for global self-rated health, a validated predictor of long-term mortality.
• Self-Esteem (Items 1, 2, 4, 6, 7): Evaluates intrapsychic regard, self-efficacy, and perceived interpersonal value. Items reflect self-acceptance (“I like who I am”), perceived interpersonal compatibility, resilience against defeatism, and harmonious integration within close relationships.

Functional Health Dysfunction Dimensions

• Anxiety (Items 2, 5, 7, 10, 12, 14): Captures cognitive, affective, and somatic manifestations of distress, characterized by concentration deficits, interpersonal tension, hyperarousal, insomnia, diurnal fatigue, and subjective nervousness.
• Depression (Items 4, 5, 10, 12, 13): Operationalizes the cognitive, motivational, and affective components of unipolar depressive episodes, including low frustration tolerance, impaired concentration, vegetative sleep disruption, anergia/fatigue, and pervasive dysphoria.
• Anxiety-Depression (Items 4, 5, 7, 10, 12, 13, 14): A combined affective morbidity dimension capturing generalized psychological distress and internalizing symptomatology, reflecting the shared variance between clinical anxiety and depressive spectrums.
• Pain (Item 11): A direct assessment of physical discomfort, measuring the frequency and burden of body aches and musculoskeletal or systemic hurting experienced during the past week.
• Disability (Item 17): Measures acute functional confinement and systemic role limitation, specifically recording the number of days within the past week the individual was housebound, hospitalized, or confined to a nursing facility due to physical illness, injury, or psychological incapacitation.

Theoretical Framework

The conceptual architecture of the Duke Health Profile is rooted in the biopsychosocial model of medicine pioneered by George L. Engel (1977), integrated with the holistic definitions established in the Constitution of the World Health Organization (1948). Engel posited that conventional biomedical reductionism failed to capture the complete phenomenology of human disease because it isolated somatic pathology from psychological processes and social environment. In formulating the Duke-UNC Health Profile and subsequently refining it into the DUKE, Parkerson and colleagues translated these theoretical tenets into psychometric operations, positing that a patient’s clinical health status is an emergent property resulting from the interaction of biological integrity, cognitive-affective functioning, and interpersonal role enactment.

Underpinning the profile is the classical conceptualization of functional status advanced by sociologists and health services theorists such as Talcott Parsons and Donald Patrick. Parsons’ formulation of the “sick role” highlighted that illness manifests concretely as an institutionalized incapacity to execute normative social roles and expectations. The DUKE reflects this paradigm by distinguishing between capacity (e.g., physical ability to climb stairs), subjective symptoms (e.g., pain, nervousness), and role performance (e.g., community participation, hospitalization-induced confinement). Furthermore, the scale incorporates Albert Bandura‘s social cognitive theory of self-efficacy, reflected in the self-esteem and mental health dimensions (e.g., “I give up too easily” and “I like who I am”), postulating that perceived personal agency directly governs how individuals cope with physical limitation and chronic pathology.

A second foundational theoretical tenet of the DUKE is the multidimensional duality of health: positive health and negative dysfunction represent related, but partially distinct, functional continua rather than simple mirror images. Derived from the psychological well-being paradigms of Bradburn (1969) and Ware (1986), the model asserts that high physical or social health can coexist with episodic symptom burden (such as localized pain or mild anxiety), and conversely, that the total absence of physical symptoms does not automatically confer high self-esteem, robust social health, or high perceived vitality. By calculating independent positive and dysfunction scales from a shared, highly consolidated item pool using differential reverse-scoring algorithms, the DUKE captures this dual continuum within an efficient structural matrix.

Validity

The construct, convergent, discriminant, and predictive validity of the Duke Health Profile has been extensively documented in clinical epidemiology, ambulatory medicine, and community health cohorts.

Construct and Convergent Validity

In the seminal psychometric validation study conducted by Parkerson, Broadhead, and Tse (1990) involving 683 adult primary care patients at Duke University Family Medicine Center, construct validity was established through convergent correlation matrices against standardized external criterion instruments. The DUKE Physical Health scale correlated strongly with the physical dimension of the Sickness Impact Profile (SIP; r = −0.56, p < 0.001) and the Physical Functioning scale of the Medical Outcomes Study 20-item Short-Form (MOS SF-20; r = 0.67). The DUKE Mental Health scale exhibited pronounced negative correlations with the Beck Depression Inventory (BDI) (r = −0.64) and the Zung Self-Rating Depression Scale (r = −0.62), while the DUKE Depression subscale correlated positively with the BDI (r = 0.67, p < 0.001). The DUKE Anxiety dimension demonstrated convergent validity through strong correlations with the Spielberger State-Trait Anxiety Inventory (STAI; r = 0.61).

Further validation by Parkerson et al. (1991, 1992) comparing the DUKE against the MOS 36-Item Short Form Health Survey (SF-36) demonstrated substantial conceptual overlap across matching dimensions: DUKE Pain correlated with SF-36 Bodily Pain at r = −0.72; DUKE Social Health correlated with SF-36 Social Functioning at r = 0.54; and DUKE Perceived Health correlated with SF-36 General Health Perceptions at r = 0.65.

Discriminant and Known-Groups Validity

The DUKE consistently discriminates between clinical populations with known physiological or psychiatric morbidity and healthy reference cohorts. In studies by Parkerson, Broadhead, and Tse (1995, 1996), the profile demonstrated statistically significant score decrements across cohorts stratified by the Duke Case-Mix Severity System. Patients diagnosed with clinical unipolar depression exhibited mean DUKE Depression scores exceeding 48.0, compared to a mean of 14.2 in non-depressed primary care controls (p < 0.0001). Similarly, individuals with advanced chronic pulmonary disease, insulin-dependent diabetes with end-organ complications, or severe degenerative joint disease demonstrated mean Physical Health scores between 35.0 and 48.0, whereas healthy controls scored significantly higher (mean 82.4 ± 15.6, p < 0.001).

Predictive and Longitudinal Criterion Validity

Longitudinal research confirms that baseline DUKE scores reliably predict health services utilization, medical expenditure, disability leave, and all-cause mortality. In a prospective cohort of primary care attendees followed over 12 months (Parkerson et al., 1995), individuals with low baseline Perceived Health (scores ≤ 50) and elevated Disability scores exhibited a 3.4-fold increased relative risk of emergency department visits and inpatient hospital admissions, independent of biological age, gender, and comorbidity count. Receiver operating characteristic (ROC) curve analyses evaluating the DUKE Anxiety-Depression and Depression subscales as screening tools against structured clinical psychiatric interviews (DSM-III-R criteria via the DIS) yielded areas under the curve (AUC) between 0.81 and 0.86, confirming screening diagnostic utility.

Reliability

Psychometric reliability for the Duke Health Profile has been substantiated via multiple evaluations of internal consistency and test-retest reproducibility across general adult, primary care, geriatric, and chronically ill populations.

Internal Consistency Reliability

Because several DUKE subscales comprise a small number of items (ranging from 1 to 7 items per dimension) designed to maximize content breadth rather than redundant homogeneity, internal consistency estimates as measured by Cronbach’s alpha vary according to scale length:

• Physical Health (5 items): Cronbach’s α ranges from 0.64 to 0.74 across diverse validation samples.
• Mental Health (5 items): Cronbach’s α ranges from 0.70 to 0.80.
• Social Health (5 items): Cronbach’s α typically ranges from 0.55 to 0.65, reflecting the multidimensional nature of social interactions (e.g., family vs. broad community activities).
• Self-Esteem (5 items): Cronbach’s α ranges from 0.62 to 0.74.
• Anxiety (6 items): Cronbach’s α ranges from 0.68 to 0.77.
• Depression (5 items): Cronbach’s α ranges from 0.72 to 0.82.
• Anxiety-Depression (7 items): Cronbach’s α ranges from 0.78 to 0.84.

Single-item scales (Perceived Health, Pain, and Disability) cannot be evaluated via Cronbach’s alpha; their stability is assessed via test-retest reproducibility.

Test-Retest Reliability and Temporal Stability

In the original test-retest stability analysis reported by Parkerson et al. (1990), an ambulatory cohort was re-administered the DUKE at 1-week and 4-week intervals. Pearson product-moment correlation coefficients (r) and intraclass correlation coefficients (ICC) demonstrated high stability under conditions of clinical non-intervention:

• Physical Health: r = 0.78 (1 week), r = 0.72 (4 weeks)
• Mental Health: r = 0.75 (1 week), r = 0.68 (4 weeks)
• Social Health: r = 0.65 (1 week), r = 0.61 (4 weeks)
• General Health: r = 0.82 (1 week), r = 0.77 (4 weeks)
• Perceived Health: r = 0.73 (1 week), r = 0.66 (4 weeks)
• Pain: r = 0.71 (1 week), r = 0.64 (4 weeks)
• Disability: r = 0.68 (1 week), r = 0.62 (4 weeks)
• Depression: r = 0.77 (1 week), r = 0.70 (4 weeks)
• Anxiety: r = 0.74 (1 week), r = 0.67 (4 weeks)

These coefficients satisfy psychometric criteria for group-level comparisons and longitudinal monitoring in health services evaluation.

Factor Analysis

The factorial validity of the 17-item DUKE has been substantiated through exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) across varied international populations.

Exploratory Factor Analyses

During initial instrument development from the parent 63-item DUHP, principal component and common factor analyses with varimax and promax rotations were conducted on candidate items. The analyses resolved a stable underlying structure corresponding to the primary domains of functioning: Physical Mobility/Capacity, Somatic Pain/Fatigue, Emotional Distress/Depressive Affect, Cognitive Stability/Self-Acceptance, and Interpersonal Functioning.

Cross-loading patterns were identified and deliberately retained where theoretically justified: for example, Item 10 (“Sleeping”) and Item 12 (“Getting tired easily”) load robustly onto the Physical Health axis (λ = 0.48 to 0.58) as somatic vegetative indicators, while simultaneously loading onto the Depression and Anxiety axes (λ = 0.44 to 0.59), reflecting their dual somatic-affective clinical etiology.

Confirmatory Factor Analytic Findings

Subsequent structural equation modeling across multinational cohorts—including French, Spanish, Norwegian, and Taiwanese adaptations—has evaluated the fit of the theoretical multi-factor hierarchical model. When testing a first-order multi-factor configuration comprising Physical, Mental, and Social health factors linked to a second-order General Health construct, confirmatory analyses consistently yield acceptable-to-good fit parameters:

• Comparative Fit Index (CFI): 0.91 to 0.95 across community and primary care samples.
• Tucker-Lewis Index (TLI): 0.90 to 0.94.
• Root Mean Square Error of Approximation (RMSEA): 0.046 to 0.062 (90% CI [0.040, 0.068]), indicating good model fit.
• Standardized Root Mean Square Residual (SRMR): 0.042 to 0.055.

Invariance testing confirms configural and metric invariance across adult age groups and genders, supporting valid comparative interpretations across demographic strata.

Instrument / Measurement Tool

• Test Type: Patient-Reported Outcome Measure (PROM); generic health-related quality of life and functional health profile.
• Target Population: Adults aged 18 years and older; adapted versions exist for adolescent and geriatric populations.
• Administration Format: Self-administered (paper-and-pencil, computer, tablet, or web platform) or interviewer-administered (in person or via telephone).
• Completion Time: 3 to 5 minutes.
• Item Count: 17 items (primary standard version); short-form variations include an 8-item brief clinical screener and a 15-item Duke Population Health Profile.
• Response Architecture: 3-point polytomous ordinal scales:
  • Items 1–7: (2) Yes, describes me exactly; (1) Somewhat describes me; (0) No, doesn’t describe me at all.
  • Items 8–9: (2) None; (1) Some; (0) A Lot.
  • Items 10–14: (2) None; (1) Some; (0) A Lot (for positive health calculations), or reverse-scored (for dysfunction calculations).
  • Items 15–16: (0) None; (1) Some; (2) A Lot.
  • Item 17: (2) None; (1) 1–4 Days; (0) 5–7 Days (reversed for disability: 0, 1, 2).
• Scoring and Transformation Formulas: Raw item responses are combined and multiplied by specific weighting factors to produce scale scores standardized from 0 to 100. Higher scores on the 6 health measures signify superior health; higher scores on the 5 dysfunction measures indicate greater impairment.
  • Physical Health (Items 8 to 12): Score = (Item 8 + Item 9 + Item 10 + Item 11 + Item 12) × 10 [Range: 0–100]
  • Mental Health (Items 1, 4, 5, 13, 14): Score = [Item 1 + (2 − Item 4) + (2 − Item 5) + Item 13 + Item 14] × 10 [Range: 0–100]
  • Social Health (Items 2, 6, 7, 15, 16): Score = [(2 − Item 2) + Item 6 + Item 7 + Item 15 + Item 16] × 10 [Range: 0–100]
  • General Health: Score = (Physical Health + Mental Health + Social Health) / 3 [Range: 0–100]
  • Perceived Health (Item 3): Score = Item 3 × 50 [Range: 0–100]
  • Self-Esteem (Items 1, 2, 4, 6, 7): Score = [Item 1 + (2 − Item 2) + (2 − Item 4) + Item 6 + Item 7] × 10 [Range: 0–100]
  • Anxiety (Items 2, 5, 7, 10, 12, 14): Invert responses where applicable: Score = [Item 2 + Item 5 + (2 − Item 7) + (2 − Item 10) + (2 − Item 12) + (2 − Item 14)] × 8.333 [Range: 0–100]
  • Depression (Items 4, 5, 10, 12, 13): Score = [Item 4 + Item 5 + (2 − Item 10) + (2 − Item 12) + (2 − Item 13)] × 10 [Range: 0–100]
  • Pain (Item 11): Score = (2 − Item 11) × 50 [Range: 0–100]
  • Disability (Item 17): Score = (Raw Confinement Score: None=0, 1–4 days=1, 5–7 days=2) × 50 [Range: 0–100]
  • Anxiety-Depression (Items 4, 5, 7, 10, 12, 13, 14): Score = [Item 4 + Item 5 + (2 − Item 7) + (2 − Item 10) + (2 − Item 12) + (2 − Item 13) + (2 − Item 14)] × 7.143 [Range: 0–100]

Permissions & Fee and Test Year

The standard 17-item Duke Health Profile was originally validated and published in 1990 by George R. Parkerson Jr., W. E. Broadhead, and C. K. J. Tse in Medical Care (Vol. 28, No. 11, pp. 1056–1072). The instrument and its family of health assessment tools (including the Duke-UNC Health Profile, the Duke Population Health Profile, and the Duke Severity of Illness Checklist) are copyrighted by Duke University (Durham, North Carolina, USA).

Duke University maintains an academic access policy: the instrument is generally accessible free of charge for non-commercial academic research, public health initiatives, medical training, and independent clinical practice. Commercial entities, pharmaceutical clinical trial sponsors, for-profit healthcare platforms, and software vendors seeking to embed the DUKE within proprietary electronic health record (EHR) systems are required to execute a formal licensing agreement and pay associated licensing fees. Inquiries regarding licensing, official language translations, and computerized scoring algorithms should be directed to the Department of Community and Family Medicine, Duke University Medical Center.

References

• McDowell, I. (2006). Measuring health: A guide to rating scales and questionnaires (3rd ed.). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195165678.001.0001
• Parkerson, G. R., Jr. (1994). User’s guide for the Duke Health Profile (DUKE). Department of Community and Family Medicine, Duke University Medical Center.
• Parkerson, G. R., Jr. (2002). User’s guide for Duke health measures. Department of Community and Family Medicine, Duke University Medical Center.
• Parkerson, G. R., Jr., Broadhead, W. E., & Tse, C. K. J. (1990). The Duke Health Profile: A 17-item measure of health and dysfunction. Medical Care, 28(11), 1056–1072. https://doi.org/10.1097/00005650-199011000-00007
• Parkerson, G. R., Jr., Broadhead, W. E., & Tse, C. K. J. (1991). Comparison of the Duke Health Profile and the MOS Short-Form in healthy young adults. Medical Care, 29(7), 679–683. https://doi.org/10.1097/00005650-199107000-00007
• Parkerson, G. R., Jr., Broadhead, W. E., & Tse, C. K. J. (1991). Development of the 17-item Duke Health Profile. Family Practice, 8(4), 396–401. https://doi.org/10.1093/fampra/8.4.396
• Parkerson, G. R., Jr., Broadhead, W. E., & Tse, C. K. J. (1992). Quality of life and functional health of primary care patients. Journal of Clinical Epidemiology, 45(11), 1303–1313. https://doi.org/10.1016/0895-4356(92)90171-t
• Parkerson, G. R., Jr., Broadhead, W. E., & Tse, C. K. (1995). Health status and severity of illness as predictors of outcomes in primary care. Medical Care, 33(1), 53–66. https://doi.org/10.1097/00005650-199501000-00005
• Parkerson, G. R., Jr., Broadhead, W. E., & Tse, C. K. J. (1996). Anxiety and depressive symptom identification using the Duke Health Profile. Journal of Clinical Epidemiology, 49(1), 85–93. https://doi.org/10.1016/0895-4356(96)00008-8
• Parkerson, G. R., Jr., Gehlbach, S. H., Wagner, E. H., Schmitt, M. G., & Baker, J. (1981). The Duke-UNC Health Profile: An adult health status instrument for primary care. Medical Care, 19(8), 806–828. https://doi.org/10.1097/00005650-198108000-00002

Items of the Scale