Health Assessment Questionnaire (HAQ)

Abstract

The Health Assessment Questionnaire (Health Assessment Questionnaire (HAQ)), introduced by Dr. James F. Fries and colleagues at Stanford University in 1980, represents a foundational paradigm shift in clinical evaluative science and rheumatology. Developed as a core component of the Arthritis, Rheumatism, and Aging Medical Information System (ARAMIS), the HAQ was engineered to quantify functional status and physical disability directly from the patient’s perspective, providing a rigorous empirical alternative to conventional laboratory markers and radiographic staging. The central component of the instrument, the Health Assessment Questionnaire Disability Index (HAQ-DI), assesses functional capacity across 20 specific physical activities grouped into eight domains of daily living: Dressing & Grooming, Arising, Eating, Walking, Hygiene, Reach, Grip, and Common Daily Activities. Items are scored on a four-point ordinal difficulty scale ranging from 0 (without any difficulty) to 3 (unable to do), supplemented by an algorithmic adjustment that incorporates the use of assistive devices or personal assistance. Over four decades of empirical investigation have demonstrated that the HAQ possesses exceptional psychometric properties, including high internal consistency (Cronbach’s alpha = 0.88–0.98), robust test-retest reliability ($r = 0.85$–$0.98$), and profound predictive validity for long-term health outcomes, including work disability, joint replacement, healthcare costs, and all-cause mortality. The instrument forms an essential element of the American College of Rheumatology (ACR) and European Alliance of Associations for Rheumatology (EULAR) core response criteria, standing as the international gold standard for measuring patient-reported functional outcomes in chronic disabling conditions.

Keywords

Health Assessment Questionnaire, HAQ, HAQ-DI, functional disability, activities of daily living, rheumatoid arthritis, patient-reported outcome measures, psychometrics, physical functioning, James Fries

Authors

The Health Assessment Questionnaire was conceptualized, developed, and validated by the investigative team at the Stanford University School of Medicine led by:

• James F. Fries, MD — Professor of Medicine (Emeritus), Division of Immunology and Rheumatology, Stanford University School of Medicine, Stanford, California, USA; Principal Investigator of ARAMIS. Contact: [email protected].
• Patricia W. Spitz, RN, MPH — Senior Research Associate, Division of Immunology and Rheumatology, Stanford University School of Medicine, Stanford, California, USA.
• Donald Y. Young, MD — Clinical Assistant Professor of Medicine, Stanford University School of Medicine, Stanford, California, USA.
• Richard G. Kraines, MS — Senior Biostatistician and Computer Systems Architect, Stanford University School of Medicine, Stanford, California, USA.
• Halsted R. Holman, MD — Guggenhime Professor of Medicine, Stanford University School of Medicine, Stanford, California, USA.

Purpose

The primary clinical and psychometric purpose of the Health Assessment Questionnaire is to provide an objective, reliable, and standardized measurement of physical disability and functional status over time in individuals suffering from chronic degenerative, musculoskeletal, and systemic rheumatic conditions. Prior to the dissemination of the HAQ in 1980, clinical evaluation in disorders such as rheumatoid arthritis (RA), osteoarthritis, and systemic lupus erythematosus relied almost exclusively on biological proxies, including erythrocyte sedimentation rate (ESR), rheumatoid factor titers, swollen joint counts, and radiographic erosion scores. While these parameters reflect pathophysiological processes, they correlate poorly with an individual’s actual capacity to function in domestic, vocational, and social spheres.

The HAQ was designed to operationalize the patient’s daily lived experience of functional limitation into a mathematically sound, interval-like metric. By measuring fine motor performance, gross motor locomotion, and coordination, the HAQ serves several essential purposes:

• Clinical Trial Efficacy Endpoint: It acts as a primary or key secondary outcome variable to detect structural and symptomatic therapeutic response to disease-modifying antirheumatic drugs (DMARDs), biologic therapies, and targeted synthetic small molecules.
• Longitudinal Observational Monitoring: Within clinical practice and population registries, routine HAQ administration enables clinicians to evaluate disease trajectories, track subclinical functional deterioration, and implement treat-to-target therapy adjustments.
• Prognostication and Risk Stratification: Baseline HAQ-DI scores demonstrate profound prognostic capability. Elevated HAQ scores reliably forecast premature cessation of employment, early institutionalization, requirement for total joint arthroplasty, escalated healthcare utilization, and premature death.
• Health Economics and Quality-Adjusted Life Years (QALYs): HAQ functional indices serve as fundamental inputs in pharmacoeconomic modeling, allowing health economists to calculate utility indices (such as through mapping algorithms from HAQ to EuroQol EQ-5D or HUI-3) for cost-effectiveness analyses.

Psychological Construct

The core psychological construct quantified by the HAQ is functional disability within the sphere of physical activities of daily living (ADLs). Within the classification architecture of the World Health Organization’s International Classification of Functioning, Disability and Health (ICF), physical health spans structural impairments (e.g., joint synovitis), activity limitations (e.g., inability to manipulate fine objects), and participation restrictions (e.g., vocational exclusion). The HAQ-DI specifically isolates activity limitations that are common, meaningful, and vulnerable to musculoskeletal or connective tissue dysfunction.

The construct is operationalized across eight distinct conceptual dimensions, each reflecting discrete biomechanical and anatomical functional units:

1. Dressing and Grooming: Measures fine motor coordination of the distal upper extremities, manual dexterity, digit flexion/extension, and bilateral coordination necessary for manipulating fasteners, tying laces, and reaching the cranium (e.g., dressing oneself, shampooing hair).
2. Arising: Evaluates lower extremity proximal muscle strength (quadriceps and gluteal musculature), axial skeletal stability, and weight-bearing hip and knee joint integrity required to transition between seated and standing postures or negotiate bed transfers.
3. Eating: Quantifies forearm pronation and supination, grip strength, and upper extremity coordination during feeding mechanics (e.g., cutting meat, lifting a full beverage, opening a standard container).
4. Walking: Assesses ambulatory mobility, aerobic endurance, lower extremity alignment, and ankle-foot mechanics under diverse conditions (e.g., walking outdoors on even flat surfaces, climbing stairs).
5. Hygiene: Captures complex full-body range of motion, axial flexibility, balance, and reach required for personal sanitation (e.g., washing and drying the entire body, tub bathing, toilet transfers).
6. Reach: Evaluates shoulder joint abduction and flexion, scapular mechanics, and upper torso extension against gravitational loads (e.g., reaching for and lowering a five-pound weight from overhead, bending down to recover floor-level items).
7. Grip: Focuses explicitly on hand power, finger strength, cylindrical and spherical grasp, and wrist rotational torque (e.g., opening car doors, turning tightly fitted faucets, breaking seals on previously opened jars).
8. Common Daily Activities: Synthesizes multiple physical domains into instrumental activities of daily living (IADLs), reflecting endurance, executive mobility, vehicular access, and heavier domestic physical tasks (e.g., running community errands, automobile entry and egress, domestic chores such as vacuuming or yard maintenance).

Crucially, the construct recognized by Fries and colleagues accounts for human behavioral adaptation. When an individual experiences functional limitation, they frequently adopt compensatory strategies, such as utilizing specialized physical aids (e.g., reachers, jar openers, canes) or relying on human assistance. The HAQ-DI incorporates this compensatory reality directly into the construct: if an individual performs an activity without subjective difficulty solely because they rely on an adaptive device or another person, their construct score is adjusted upward to reflect latent functional limitation.

Theoretical Framework

The theoretical framework underlying the Health Assessment Questionnaire is Dr. James Fries’ Hierarchy of Patient Outcomes, often known as the “Five Ds” model of health status. Fries posited that clinical medicine must orient its scientific focus away from isolated surrogate biological markers toward terminal, human-centered outcomes that directly encompass patient welfare. The five ultimate negative outcomes identified in this framework are:

• Death: Premature mortality and reduced life expectancy.
• Disability: Loss of capacity to perform normal, essential physical and social functions.
• Discomfort: Subjective physical pain, fatigue, and psychological distress.
• Drug Toxicity: Iatrogenic adverse consequences arising from medical interventions.
• Dollar Cost: Direct and indirect economic burdens incurred by the patient and society.

Within this hierarchy, Fries asserted that disability occupies the pivotal intermediate position. It represents the chronic cumulative consequence of untreated biological disease, while simultaneously acting as the primary driver of discomfort, medical expenditures, and early mortality. To capture these outcomes comprehensively, the full Stanford HAQ battery originally integrated five dimensions: the Disability Index (HAQ-DI), the Discomfort/Pain Scale (measured via visual analog scales), the Drug Toxicity Index, General Health/Well-being ratings, and Economic/Cost assessment questionnaires.

Furthermore, the HAQ is anchored theoretically within the Disablement Process Model articulated by sociologists and epidemiologists such as Saad Nagi, Lois Verbrugge, and Alan Jette. This model delineates a linear progression: Pathology (active disease process) $\rightarrow$ Impairment (dysfunction at the organ or joint level) $\rightarrow$ Functional Limitation (restriction in performing discrete physical actions) $\rightarrow$ Disability (inability to execute socially defined roles and activities). The HAQ-DI was explicitly engineered to assess the intersection of functional limitation and disability, establishing a psychometrically robust behavioral metric that reflects structural impairment without being confounded by subjective emotional distress or personality traits.

Validity

The construct, criterion, convergent, discriminant, and predictive validity of the HAQ have been subjected to extensive empirical verification across diverse cultural and diagnostic populations.

Construct and Convergent Validity

Convergent validity has been established through strong, statistically significant correlations between the HAQ-DI and both clinical disease activity indices and objective physical performance metrics. In patients with rheumatoid arthritis, HAQ-DI scores correlate robustly with:

• Swollen and tender joint counts ($r = 0.40$ to $0.65$, $p < 0.001$).
• The Disease Activity Score 28 (DAS28) and Simplified Disease Activity Index (SDAI) ($r = 0.55$ to $0.72$).
• Visual Analog Scale (VAS) pain assessments ($r = 0.50$ to $0.75$).
• Objective physical performance measures, including grip strength measured via dynamometry ($r = -0.50$ to $-0.68$), 50-foot walking speed ($r = 0.60$ to $0.78$), and the Keitel Functional Test ($r = 0.70$ to $0.85$).
• The Physical Component Summary (PCS) score of the Short Form-36 (SF-36), where inverse correlations consistently exceed $-0.75$, demonstrating excellent construct concordance.

Discriminant Validity

The HAQ demonstrates clear discriminant validity by differentiating between functional physical limitations and purely affective or psychological distress. Correlational analyses between HAQ-DI scores and the Mental Component Summary (MCS) of the SF-36 or dedicated depression measures (e.g., the Beck Depression Inventory) typically yield weak-to-modest coefficients ($r = 0.15$ to $0.32$). This confirms that the instrument isolates somatic and locomotive dysfunction from independent psychological disturbances. Moreover, the HAQ successfully discriminates between clinical cohorts characterized by varying stages of structural anatomical damage (Steinbrocker functional classes I through IV; $p < 0.0001$).

Predictive and Criterion Validity

The predictive validity of the HAQ represents its most striking psychometric attribute. In seminal longitudinal cohort studies conducted by Pincus et al. (1989) and Fries et al. (1990), baseline HAQ-DI scores proved to be the single most potent predictor of 5-year, 10-year, and 20-year all-cause mortality in rheumatoid arthritis—outperforming traditional laboratory markers (ESR, rheumatoid factor), disease duration, comorbidities, and initial radiographic damage scores. An incremental elevation of 0.25 units on the HAQ-DI is independently associated with an adjusted hazard ratio for mortality of 1.15 to 1.30. Furthermore, baseline HAQ-DI scores demonstrate profound predictive capacity for severe outcomes:

• Permanent work disability and vocational cessation within 3 to 5 years (odds ratios ranging from 2.5 to 4.2).
• Subsequent total hip and knee replacement surgery.
• Future direct medical expenditures and cumulative hospitalization days.

Reliability

The Health Assessment Questionnaire demonstrates exceptional reliability across multiple psychometric evaluation modalities, including internal consistency, test-retest reproducibility, and inter-rater/inter-method equivalence.

Internal Consistency

Internal consistency estimates for the HAQ-DI have been repeatedly evaluated using Cronbach’s alpha and McDonald’s omega. Across normative cohorts, clinical trial samples, and language-adapted translations, Cronbach’s alpha values for the 20 items consistently range between 0.88 and 0.98. Evaluation of the eight distinct category subscores similarly yields alpha coefficients exceeding 0.85. Split-half reliability coefficients, calculated via the Spearman-Brown prophecy formula, exceed 0.90, establishing that individual items sample the latent domain of physical limitation with minimal measurement error.

Test-Retest Reliability

The stability of the instrument over time in clinically stable populations has been demonstrated across varied intervals:

• Over short-term intervals (24 to 72 hours), test-retest Pearson correlation coefficients ($r$) and intraclass correlation coefficients (ICC) range from 0.92 to 0.98.
• Over two-week to one-month intervals in patients maintaining constant medication regimens, Pearson product-moment correlations remain between 0.85 and 0.95.
• Item-by-item test-retest concordance, measured via Cohen’s weighted kappa ($\kappa_w$), typically ranges from 0.65 to 0.85, indicating substantial to near-perfect agreement at the single-item level.

Inter-Rater and Administration Modality Agreement

The HAQ exhibits remarkable equivalence across different administration modes. Comparative studies examining self-administered postal questionnaires, in-person clinic self-reports, telephone interviews, proxy responses (e.g., spouses), and direct clinical interviews have yielded intraclass correlation coefficients consistently exceeding 0.85. The standard error of measurement (SEM) for the overall HAQ-DI score is estimated between 0.08 and 0.12 units, with the minimum clinically important difference (MCID) established across multiple clinical trials as a change of 0.22 units (or approximately a 10% to 15% shift on the 0 to 3 scale).

Factor Analysis

Extensive psychometric investigations utilizing exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and Item Response Theory (IRT) have confirmed the structural properties of the HAQ-DI.

Exploratory and Confirmatory Factor Structure

Early exploratory factor analyses frequently revealed a dominant first unrotated general factor accounting for 50% to 65% of the total item variance, supporting the conceptualization of the HAQ-DI as a unified measure of physical functional limitation. When multidimensional models are evaluated, factor analysis typically extracts two or three inter-correlated sub-dimensions reflecting anatomical clusters:

• Upper Extremity Function: Combining Dressing & Grooming, Eating, and Grip items, primarily loading on fine motor control and bilateral wrist/hand articulation (factor loadings: 0.65 to 0.88).
• Lower Extremity / Axial Mobility: Comprising Arising and Walking, driven by hip, knee, and ankle weight-bearing mechanics (factor loadings: 0.70 to 0.91).
• Complex / Global Trunk Function: Spanning Hygiene, Reach, and Common Daily Activities, which demand coordinated gross motor and multi-joint kinetic involvement (factor loadings: 0.58 to 0.82).

Modern Confirmatory Factor Analyses (CFA) demonstrate that a hierarchical bifactor model—featuring one overarching general disability factor alongside eight specific first-order category factors—provides superior model fit indices across international populations: Comparative Fit Index (CFI) > 0.96, Tucker-Lewis Index (TLI) > 0.95, and Root Mean Square Error of Approximation (RMSEA) < 0.06.

Modern Psychometrics: Rasch and IRT Findings

Application of the Rasch measurement model and multi-parameter IRT has provided deeper insights into the performance of the HAQ:

• Item Difficulty Hierarchy: IRT calibrations consistently indicate that items vary across a broad spectrum of latent functional ability ($ heta$). Simpler items such as “Lift a full cup or glass to your mouth” represent very low difficulty (capturing severe disability), whereas items such as “Climb up five steps” or “Do chores such as vacuuming or yard work” represent higher difficulty thresholds (discriminating higher functioning levels).
• Ceiling and Floor Effects: In non-clinical populations or patients with very early, well-controlled disease, the HAQ-DI exhibits a noticeable floor effect (a score of 0.0), where mild functional deficits may go undetected. Conversely, in advanced, highly disabled populations, genuine ceiling effects are rare, as the score of 3.0 represents complete motor dependence.
• Differential Item Functioning (DIF): Minor DIF has been identified across gender and age for items such as “Do chores such as vacuuming or yard work” and “Open car doors,” reflecting normative differences in physical task distribution and baseline grip strength rather than clinical measurement bias. Overall, however, item calibrations remain remarkably invariant across cultural cohorts.

Instrument / Measurement Tool

The standard Health Assessment Questionnaire Disability Index (HAQ-DI) consists of 20 primary activity items distributed across eight categories, supplemented by assistive device and personal assistance indicators, as well as companion visual analog scales.

Key Structural Characteristics

• Test Type: Patient-Reported Outcome Measure (PROM); standardized self-administered functional capacity questionnaire.
• Administration Format: Paper-and-pencil, computer-administered (ePRO), clinical interview, or telephone-administered.
• Number of Items: 20 primary disability items; 8 categories of assistive devices; 8 categories of personal assistance; plus two accompanying 100 mm Visual Analog Scales (VAS) for Pain and Global Health. (An abbreviated 8-item short form is also utilized in specific settings).
• Recall Period: Past week (“over the past week”).
• Response Scale (Primary Items): Four-point ordinal difficulty rating:
  • 0 = Without ANY Difficulty
  • 1 = With SOME Difficulty
  • 2 = With MUCH Difficulty
  • 3 = UNABLE To Do

Standard Scoring Algorithm

The standard scoring algorithm for the HAQ-DI balances individual item responses with the functional impact of assistive devices and human help:

1. Determine Category Subscores: For each of the 8 functional categories, identify the highest item score circled by the respondent. For example, in the Eating category, if item 5 is scored 1, item 6 is scored 0, and item 7 is scored 2, the preliminary score for Eating is 2. (If multiple consecutive numbers are circled for a single item, code the higher number; if non-consecutive, code as missing).
2. Adjust for Aids, Devices, and Personal Assistance: Check the companion sections for assistive devices and personal help. If the respondent indicates using an adaptive device (e.g., jar opener, button hook, cane) OR requiring personal assistance for a given category:
   • If the preliminary category score is 0 or 1, adjust that category score upward to 2.
   • If the preliminary category score is already 2 or 3, it remains unchanged (retains the score of 2 or 3).
3. Calculate Total HAQ-DI Score: Sum the 8 final category scores and divide by the number of valid categories answered (a minimum of 6 of the 8 categories must be completed for a valid score). This yields a continuous disability index ranging from 0.00 (no functional disability) to 3.00 (maximum functional disability).
4. Interpretive Benchmarks:
   • 0.0 to 0.5: Mild to no functional disability (normal functional status).
   • 0.5 to 1.5: Moderate functional disability.
   • 1.5 to 3.0: Severe to extreme functional impairment.
   • Minimum Clinically Important Difference (MCID): $\Delta ge 0.22$ points.

Permissions & Fee and Test Year

• Initial Publication Year: 1980 (by Dr. James F. Fries, Patricia W. Spitz, Richard G. Kraines, and Halsted R. Holman at Stanford University).
• Intellectual Property & Copyright: The Health Assessment Questionnaire is copyrighted by Stanford University. However, Dr. James F. Fries maintained a firm commitment to open clinical science, making the instrument widely accessible to clinicians and independent academic investigators.
• Licensing and Fees:
  • Academic, Clinical, and Non-Profit Research: Free of charge. Academic investigators, clinicians, and non-profit institutions may reproduce and utilize the questionnaire without royalty fees.
  • Commercial Use and Industry-Sponsored Clinical Trials: Pharmaceutical corporations, contract research organizations (CROs), and commercial entities typically require formal licensing and user agreements through the Stanford University Office of Technology Licensing (OTL) or designated management entities.
• Contact and Resources: Inquiries regarding official translations, authorized digital implementations, and administration guidelines may be directed through the Stanford University Department of Medicine or historical ARAMIS archival repositories.

References

• Bruce, B., & Fries, J. F. (2003). The Stanford Health Assessment Questionnaire: Dimensions and practical applications. Health and Quality of Life Outcomes, 1(1), 20. https://doi.org/10.1186/1477-7525-1-20
• Fries, J. F. (1983). Toward an understanding of patient outcome measurement. Arthritis and Rheumatism, 26(6), 697–704. https://doi.org/10.1002/art.1780260601
• Fries, J. F., & Spitz, P. W. (1990). The hierarchy of patient outcomes. In B. Spilker (Ed.), Quality of Life Assessments in Clinical Trials (pp. 25–35). Raven Press.
• Fries, J. F., Spitz, P. W., Kraines, R. G., & Holman, H. R. (1980). Measurement of patient outcome in arthritis. Arthritis and Rheumatism, 23(2), 137–145. https://doi.org/10.1002/art.1780230202
• Fries, J. F., Spitz, P. W., & Young, D. Y. (1982). The dimensions of health outcomes: The Health Assessment Questionnaire, disability and pain scales. The Journal of Rheumatology, 9(5), 789–793.
• Goeppinger, J., Doyle, M. A., & Murdock, B. (1985). Self-administered function measures: The impossible dream? Arthritis and Rheumatism, 28(Suppl), S145.
• Lorig, K. R., Sobel, D. S., Ritter, P. L., Laurent, D., & Hobbs, M. (2001). Effect of a self-management program on patients with chronic disease. Effective Clinical Practice, 4(6), 256–262.
• 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
• Milligan, S. E., Hom, D. L., Ballou, S. P., & Kushner, I. (1993). An assessment of the Health Assessment Questionnaire functional ability index among women with systemic lupus erythematosus. The Journal of Rheumatology, 20(6), 972–976.
• Mobini, M., Mohammadpour, R. A., Khalilian, A. R., Anvari, F., & Lamsehchi, N. (2012). Measuring the validity and reliability of the Persian version of Stanford Health Assessment Questionnaire in Iranian patients with osteoarthritis. Zahedan Journal of Research in Medical Sciences, 16(9), 30–33.
• Pincus, T., Callahan, L. F., Brooks, R. H., Fuchs, H. A., Olsen, N. J., & Kaye, J. J. (1989). Self-report questionnaire scores in rheumatoid arthritis compared with traditional physical, radiographic, and laboratory measures. Annals of Internal Medicine, 110(4), 259–266. https://doi.org/10.7326/0003-4819-110-4-259
• Pincus, T., Summey, J. A., Soraci, S. A., Jr., Wallston, K. A., & Hummon, N. P. (1983). Assessment of patient satisfaction in activities of daily living using a modified Stanford Health Assessment Questionnaire. Arthritis and Rheumatism, 26(11), 1346–1353. https://doi.org/10.1002/art.1780261107
• Ramey, D. R., Raynauld, J. P., & Fries, J. F. (1992). The Health Assessment Questionnaire 1992: Status and review. Arthritis Care & Research, 5(3), 119–129. https://doi.org/10.1002/art.1790050303
• Rastmanesh, R., Rabiee, S., Shaabani, Y., Mazinani, H., Ebrahimi, A. A., & Jamshidi, A. R. (2010). Validation of the Persian version of the Stanford Health Assessment Questionnaire (HAQ) in patients with rheumatoid arthritis. Journal of Paramedical Sciences, 1(1), 16–25.
• Tagharrobi, Z., Sharifi, K., & Sooky, Z. (2013). Psychometric evaluation of Stanford health assessment questionnaire 8-item disability index (HAQ 8-item DI) in elderly people. Holistic Nursing and Midwifery Journal, 24(71), 9–20.

13. Items of the Scale (Questionnaire)