Hospital Anxiety and Depression Scale (HADS)

Abstract

The Hospital Anxiety and Depression Scale (HADS) is one of the most widely utilized and psychometrically validated self-report instruments designed to detect, screen, and quantify symptom severity of anxiety and depression among patients in non-psychiatric clinical environments. Developed in 1983 by Anthony S. Zigmond and R. Philip Snaith, the scale was engineered to address a pervasive diagnostic dilemma in consultation-liaison psychiatry and behavioral medicine: the confounding overlap between somatic manifestations of physical illness (such as fatigue, insomnia, anorexia, psychomotor slowing, and autonomic arousal) and the classic somatic diagnostic criteria for mood and anxiety disorders. Comprising 14 brief items divided equally into two distinct 7-item subscales—the Anxiety subscale (HADS-A) and the Depression subscale (HADS-D)—the instrument evaluates states over the preceding week using a four-point Likert-type response format (scored 0 to 3). Each subscale yields a continuous score ranging from 0 to 21, with conventional cut-off scores categorizing presentations into normal (0–7), borderline abnormal (8–10), and clinically significant or probable caseness (11–21). Extensively scrutinized across hundreds of psychometric evaluations involving hundreds of thousands of medical inpatients, outpatients, and general population cohorts, the HADS demonstrates robust internal consistency (Cronbach’s alpha typically ranging from .80 to .93 for HADS-A and .76 to .90 for HADS-D) and consistent test-retest reliability across diverse intervals. While its original two-factor orthogonal structure is widely confirmed, structural equation modeling and exploratory and confirmatory factor analyses have also revealed alternative latent dimensionalities—including unidimensional psychological distress models and tripartite hierarchical models (autonomic anxiety, negative affectivity, and anhedonia)—highlighting ongoing psychometric debate. This comprehensive academic review provides an exhaustive analysis of the scale’s historical purpose, theoretical foundations, psychometric construct architecture, diagnostic validity, reliability parameters, structural dimensionality, administrative scoring guidelines, licensing parameters, and the complete authentic English item inventory.

Keywords

Hospital Anxiety and Depression Scale, HADS, psychometrics, anxiety screening, depression screening, medical illness, somatic symptom contamination, anhedonia, factor analysis, construct validity

Authors

The Hospital Anxiety and Depression Scale was developed by Anthony S. Zigmond and R. Philip Snaith, faculty members and clinical researchers associated with the Department of Psychiatry at the University of Leeds and St. James’s University Hospital, Leeds, United Kingdom. R. Philip Snaith (1933–2003) was an internationally renowned British psychiatrist whose scholarly career centered on the phenomenology, clinical classification, and measurement of affective disorders, clinical anxiety states, and anhedonia. Throughout his tenure at the University of Leeds, Snaith contributed decisively to behavioral medicine, the development of the Snaith-Hamilton Pleasure Scale (SHAPS), and the refinement of rating scales in clinical psychosomatics. Dr. Anthony S. Zigmond worked extensively in consultation-liaison psychiatry, focusing on psychological morbidity within acute hospital wards and general medical clinics.

Purpose

The foundational imperative underlying the creation of the Hospital Anxiety and Depression Scale (HADS) was the resolution of a critical diagnostic challenge in medical settings: somatic symptom contamination. Prior to the early 1980s, clinicians and researchers attempting to screen for emotional distress in medical inpatients and outpatients relied heavily on classic psychometric instruments such as the Beck Depression Inventory (BDI), the Hamilton Rating Scale for Depression (HAM-D), the Zung Self-Rating Depression Scale, and the State-Trait Anxiety Inventory (STAI). While these instruments possessed sterling psychometric credentials in psychiatric populations, their diagnostic validity deteriorated significantly when applied to patients suffering from organic physical illnesses, such as cardiovascular disease, cancer, chronic obstructive pulmonary disease (COPD), end-stage renal disease, systemic lupus erythematosus, and neurological conditions.

This diagnostic confounding occurred because conventional diagnostic instruments relied heavily on biological or neurovegetative signs of distress, including weight loss, altered appetite, sleep architecture disruption, generalized fatigue, somatic weakness, headaches, gastrointestinal upset, and psychomotor changes. In a patient undergoing cytotoxic chemotherapy, coping with advanced heart failure, or recovering from major abdominal surgery, these symptoms are ubiquitous physiological consequences of the underlying organic disease or its medical treatment, rather than unambiguous markers of an affective or anxiety disorder. When evaluated with traditional instruments, medical patients routinely produced artificially elevated scores, resulting in unacceptably high rates of false-positive diagnoses. Conversely, medical professionals frequently underdiagnosed psychiatric distress by dismissing emotional anguish as an understandable or secondary biological side effect of somatic pathology.

Zigmond and Snaith (1983) purposefully designed the HADS to circumvent this dilemma by strictly excluding physical, vegetative, and somatic symptomatology. The scale was constructed to measure the psychic, emotional, and cognitive manifestations of anxiety and depressive mood states. By stripping away bodily signs, the HADS provides a rapid, cost-effective, self-administered dimensional assessment that reliably distinguishes clinically significant mood disturbances from the expected physiological sequelae of physical illness. Its core clinical purpose is threefold:

• To serve as an efficient screening instrument in non-psychiatric hospital clinics, outpatient services, primary care, and post-surgical wards to identify patients who require formal psychiatric assessment or psychosocial intervention.
• To evaluate the dimensional severity of anxiety and depressive symptom profiles over time, facilitating the longitudinal tracking of therapeutic interventions, pharmacological treatments, psychological therapies, and palliative care regimens.
• To operate as an unambiguous, standardized outcome measure in epidemiological, psychosomatic, and health services research investigating the burden of mental health comorbidities across medical cohorts.

Psychological Construct

The Hospital Anxiety and Depression Scale conceptualizes psychological distress through two related yet clinically distinct dimensions: psychic anxiety and anhedonic depression. Each dimension is operationalized through a dedicated seven-item subscale that isolates psychological and emotional experience from physiological confounders.

1. The Anxiety Subscale (HADS-A)

The HADS-A subscale assesses psychic anxiety, capturing a subjective state of mental tension, generalized apprehension, internal restlessness, and autonomic alarm without relying on direct physiological indices. The construct taps into the cognitive and affective features of generalized anxiety disorder (GAD), panic vulnerability, and generalized neurotic tension:

• Cognitive Apprehension and Intrusive Worry: Evaluated by items such as “Worrying thoughts go through my mind” (Item 5) and “I get a sort of frightened feeling as if something awful is about to happen” (Item 3), measuring catastrophic forecasting, rumination, and pervasive dread.
• Tension and Inability to Relax: Captured by “I feel tense or ‘wound up'” (Item 1) and “I can sit at ease and feel relaxed” (Item 7, reverse-scored), assessing subjective strain, high baseline psychological arousal, and failure of emotional homeostasis.
• Subjective Somatic Projection (Autonomic Alarm): Although the HADS minimizes somatic criteria, Item 9 (“I get a sort of frightened feeling like ‘butterflies’ in the stomach”) includes a visceral metaphor. Zigmond and Snaith retained this single item because clinical interviews indicated that patients distinguish the subjective “butterflies” of sudden fright from primary organic gastrointestinal pathology.
• Motor Restlessness and Panic Propensity: Operationalized through “I feel restless as if I have to be on the move” (Item 11) and “I get sudden feelings of panic” (Item 13), capturing acute bursts of panic and psychomotor agitation without relying on cardiovascular parameters like tachycardia or diaphoresis.

2. The Depression Subscale (HADS-D)

The construct operationalized within the HADS-D focuses almost exclusively on anhedonia—the diminished capacity to experience pleasure, anticipation, and emotional responsiveness. Snaith argued that anhedonia represents the true, unadulterated psychopathological core of endogenous, biologically treatable depressive illness (melancholia), whereas somatic manifestations (insomnia, fatigue, weight changes) introduce noise into medical evaluations. The subscale evaluates distinct facets of anhedonia:

• Consummatory Anhedonia: Evaluated by items measuring present-moment hedonic responsiveness, such as “I still enjoy the things I used to enjoy” (Item 2) and “I can laugh and see the funny side of things” (Item 4). These items measure the loss of positive affective resonance in everyday social, recreational, or aesthetic contexts.
• Anticipatory Anhedonia: Operationalized via “I look forward with enjoyment to things” (Item 12), measuring the motivational component of reward processing—the cognitive ability to anticipate future pleasure, which is profoundly disrupted in clinical depressive episodes.
• Affective Blunting and Loss of Cheerfulness: Measured by “I feel cheerful” (Item 6), tracking the absence of baseline positive mood.
• Subjective Slowing and Behavioral Withdrawal: Evaluated through “I feel as if I am slowed down” (Item 8), which captures the cognitive and subjective experience of psychomotor inertia rather than observable physiological bradykinesia, and “I can enjoy a good book or radio or TV programme” (Item 14), which measures cognitive engagement with culture and leisure.
• Self-Care and Demoralization: Measured by “I have lost interest in my appearance” (Item 10), reflecting apathy, personal neglect, and social disconnection characteristic of severe affective withdrawal.

Theoretical Framework

The HADS is conceptually anchored in clinical phenomenology, psychiatric nosology, and contemporary affect theory—specifically, the Tripartite Model of Anxiety and Depression formulated by Lee Anna Clark and David Watson (1991), as well as Snaith’s psychopathological focus on hedonic deficit.

Historically, the relationship between anxiety and depression has posed a major challenge to dimensional psychometrics. In clinical populations, anxiety and depression exhibit enormous phenotypic comorbidity, frequently correlating between .60 and .80 on self-report questionnaires. Clark and Watson’s tripartite theory posited that this shared variance is driven by a broad, overarching higher-order construct: Negative Affectivity (NA), or general neurotic distress. Individuals high in NA experience pervasive subjective distress, irritability, worry, emotional lability, and cognitive vulnerability across diverse situations. However, to achieve discriminant validity, diagnostic instruments must capture the specific elements that differentiate the two clinical syndromes:

• Low Positive Affectivity (PA): Specifically characterizes depression. This manifest state includes anhedonia, lethargy, apathy, lack of enthusiasm, and loss of interest. It is largely independent of generalized anxiety.
• Physiological Hyperarousal (PH): Specifically characterizes anxiety disorders, manifesting as somatic autonomic activation, trembling, dizziness, panic, and bodily vigilance.

Although the HADS was developed nearly a decade before Clark and Watson formalized the Tripartite Model, Zigmond and Snaith’s intuitive clinical insights closely anticipated this affective paradigm. Zigmond and Snaith recognized that general medical illness artificially mimics Physiological Hyperarousal (via dyspnea, tachycardia, orthostatic hypotension) and elevates Negative Affectivity (via realistic concern about mortality, pain, and disability). Consequently, they deliberately purged PH items from the HADS-A (retaining only psychic tension and cognitive panic) and loaded the HADS-D exclusively with markers of Low Positive Affectivity (Anhedonia).

By defining depression through the selective lens of hedonic deficit rather than general distress or neurovegetative dysregulation, the HADS-D functions as an operationalization of Snaith’s concept of true melancholia. The theoretical framework assumes that while a medically ill patient may be in severe physical pain, functionally limited, and experiencing transient grief over their medical diagnosis, they will not present with true consummatory and anticipatory anhedonia unless a distinct depressive disorder has emerged.

Validity

The psychometric validity of the HADS has been examined across clinical cohorts, including oncology, cardiology, neurology, palliative care, rheumatology, chronic pain, and primary care, as well as general community samples.

1. Criterion and Predictive Validity

Criterion-related validity has been demonstrated by comparing HADS subscale scores against structured diagnostic psychiatric interviews based on the DSM and ICD (such as the SCID, SCAN, and MINI). In a landmark meta-analytic review of 747 papers by Bjelland, Dahl, Haug, and Neckelmann (2002), the sensitivity and specificity of both the HADS-A and HADS-D across diverse clinical environments were evaluated:

• Using the standard recommended cut-off score of 8 or higher, the HADS-A demonstrated an average sensitivity of 0.89 and an average specificity of 0.75 for identifying generalized anxiety disorder and related clinical anxiety states.
• For the HADS-D, using the cut-off score of 8 or higher yielded an average sensitivity of 0.83 and an average specificity of 0.79 against formal psychiatric diagnoses of major depressive episode or dysthymia.
• Receiver Operating Characteristic (ROC) analyses routinely generate Areas Under the Curve (AUC) between 0.84 and 0.94, confirming high diagnostic efficiency.
• When the more stringent threshold of 11 or higher is utilized (indicating “probable caseness”), specificity rises to 0.88–0.95, minimizing false positives in specialized psychiatric referral pathways.

2. Convergent and Discriminant Validity

Convergent validity is documented by strong correlations with other validated anxiety and depression measures. The HADS-A correlates strongly with the State-Trait Anxiety Inventory (STAI-Trait) (r = .64 to .81) and the Beck Anxiety Inventory (BAI) (r = .60 to .75). The HADS-D correlates substantially with the Beck Depression Inventory (BDI-I and BDI-II) (r = .62 to .78) and the Center for Epidemiologic Studies Depression Scale (CES-D) (r = .65 to .80).

Discriminant validity is supported by its capacity to differentiate emotional distress from physical symptoms. In clinical trials of cancer and end-stage organ failure, HADS scores demonstrate modest or near-zero correlations with physiological markers, laboratory parameters (such as hemoglobin levels, inflammatory cytokines, or tumor staging), and standard somatic performance indices (such as the Karnofsky Performance Scale or ECOG status). This confirms that somatic disease severity does not artificially drive HADS totals.

Reliability

The reliability of the HADS has been examined across dozens of languages, cultural contexts, and clinical populations, consistently demonstrating high internal consistency and temporal stability.

1. Internal Consistency

Extensive psychometric investigations report high internal consistency across diverse settings:

• HADS-Anxiety (HADS-A): In their systematic review, Bjelland et al. (2002) identified that Cronbach’s alpha values for HADS-A across multiple studies ranged from .68 to .93, with a median coefficient of .83. Studies in oncology cohorts (e.g., Singer et al., 2008) and general medical wards typically yield alpha values between .82 and .88, indicating strong item interrelatedness without excessive redundancy.
• HADS-Depression (HADS-D): Cronbach’s alpha values for HADS-D ranged from .67 to .90, with a median of .82 across international validation studies. In chronic illness cohorts, alpha coefficients routinely fall within the .79 to .86 interval.
• Total Score (HADS-Total): When used as a composite measure of generalized psychological distress, the total 14-item scale exhibits high internal consistency, with alpha coefficients exceeding .88 to .93.

2. Test-Retest Reliability and Temporal Stability

Temporal stability assessments have demonstrated that the scale is stable over time when clinical conditions remain unchanged, yet sensitive to therapeutic changes:

• Over short retest intervals (under 2 weeks), intra-class correlation coefficients (ICC) and Pearson product-moment correlation coefficients typically range between r = .80 and .92 for HADS-A, and between r = .81 and .89 for HADS-D.
• Over moderate follow-up periods (2 to 6 weeks) in medically stable outpatients, test-retest correlations remain robust (r = .70 to .80).
• Importantly, the scale exhibits high responsiveness and sensitivity to change following evidence-based pharmacotherapy (such as SSRIs) or cognitive-behavioral interventions, with longitudinal effect sizes (Cohen’s d) often exceeding 0.70 to 1.20 in responders.

Factor Analysis

The underlying factor structure of the HADS has generated extensive empirical investigation and scholarly debate within modern psychometrics, resulting in several competing structural models evaluated via exploratory factor analysis (EFA) and confirmatory factor analysis (CFA).

1. The Classic Two-Factor Model

Zigmond and Snaith (1983) originally proposed a clean, orthogonal two-factor structure separating the 7 anxiety items from the 7 depression items. Dozens of studies utilizing principal component analysis and maximum likelihood factor analysis have confirmed this distinct two-factor structure across general hospital patients, primary care cohorts, and non-clinical samples. In standard confirmatory factor analytic evaluations, the two-factor model typically shows acceptable to good model fit: comparative fit index (CFI) > .90, Tucker-Lewis Index (TLI) > .90, and Root Mean Square Error of Approximation (RMSEA) < .06.

2. The Unidimensional Distress Model

Because the inter-factor correlation between HADS-A and HADS-D is frequently substantial (ranging from r = .50 to .75 in severe medical cohorts), several researchers (e.g., Razavi et al., 1990) have argued that the HADS operates most effectively as a general unidimensional screening scale for non-specific psychological distress. A single general factor accounting for the bulk of common item variance often emerges in psychiatric oncology, leading some clinicians to utilize the HADS-Total score (14 items, range 0–42) as a global index of distress.

3. Three-Factor and Tripartite Models

Subsequent psychometric evaluations utilizing advanced CFA techniques identified multidimensional complexity within the scale:

• Dunbar’s Three-Factor Model (2000): In line with Clark and Watson’s tripartite theory, Dunbar et al. demonstrated superior model fit by separating the items into three latent dimensions: Negative Affectivity (general distress/neuroticism: items 1, 5, 7, 11), Autonomic Anxiety / Hyperarousal (panic-like items: 3, 9, 13), and Anhedonia / Depression (items 2, 4, 6, 8, 10, 12, 14).
• Caci et al. (2003) and Norton et al. (2013): In an extensive structural evaluation involving meta-analytic structural equation modeling, Norton and colleagues evaluated over 20 competing models in the literature and found that while bifactor models (a general distress factor alongside specific anxiety and depression group factors) often show superior statistical indices, the original two-factor model remains the most clinically viable, parsimonious, and practically interpretable representation.

4. Item Cross-Loadings and Psychomotor Inconsistencies

Psychometric studies frequently note that two items exhibit cross-loadings or anomalous performance across cultures:

• Item 7 (“I can sit at ease and feel relaxed”): Although assigned to HADS-A, this item frequently cross-loads onto the Depression factor, likely because psychomotor restlessness and anhedonic apathy both impair peaceful relaxation.
• Item 8 (“I feel as if I am slowed down”): Assigned to HADS-D, Item 8 occasionally loads onto general somatic debility or anxiety in older medical patients, leading some psychometricians to recommend bifactor estimation when interpreting subtle subscale differences.

Instrument / Measurement Tool

The Hospital Anxiety and Depression Scale is configured as a brief, easily administered, 14-item paper-and-pencil or computerized self-report questionnaire. Its structured parameters are summarized below:

• Test Type: Self-administered psychological screening questionnaire / dimensional rating scale.
• Target Population: Non-psychiatric medical inpatients and outpatients, surgical patients, primary care patients, and general adult populations (typically aged 16 to 65+). Validated adolescent versions are also available.
• Administration Format: Self-report via paper-and-pencil, bedside clinical interview, or digital interfaces (web, tablet, clinical portals).
• Item Count: 14 items total (7 items measuring Anxiety, 7 items measuring Depression).
• Subscales:
  • HADS-A (Anxiety): Items 1, 3, 5, 7, 9, 11, 13
  • HADS-D (Depression): Items 2, 4, 6, 8, 10, 12, 14
• Recall Timeframe: “Past week” (instructions direct respondents to indicate how they have been feeling over the preceding 7 days).
• Completion Time: 2 to 5 minutes, minimizing cognitive and physical burden for ill individuals.
• Response Options & Scoring Logic: Each item is rated on a 4-point categorical response scale scored from 0 to 3. The scoring direction is balanced across items to mitigate acquiescence response bias:
• Directly Scored Items (0, 1, 2, 3): Higher score directly reflects greater symptom severity. This applies to items 1, 3, 5, 8, 9, 10, 11, and 13.
• Reverse-Scored Items (3, 2, 1, 0): The phrasing represents positive health or hedonic capability, meaning high affirmation indicates the absence of pathology. This applies to items 2, 4, 6, 7, 12, and 14. In these items, the most positive response is scored 0 and the most symptomatic response is scored 3.
• Subscale Range: Each subscale (HADS-A and HADS-D) produces a raw score sum ranging from 0 to 21.
• Total Score Range: When utilized as a global composite metric of emotional distress (HADS-Total), the sum ranges from 0 to 42.
• Clinical Interpretive Ranges / Cut-Off Scores:
  • 0 to 7: Normal (Non-case / asymptomatic range)
  • 8 to 10: Borderline abnormal (Mild symptoms / doubtful caseness / warranting clinical monitoring)
  • 11 to 21: Abnormal (Clinically significant symptom severity / probable case / indicating need for formal psychiatric or psychological diagnostic evaluation)

Permissions & Fee and Test Year

The Hospital Anxiety and Depression Scale was officially published in 1983 by Anthony S. Zigmond and R. Philip Snaith in the peer-reviewed psychiatric journal Acta Psychiatrica Scandinavica. For several decades following its initial publication, the scale was widely regarded as open-access and was reproduced freely in academic journals and public domain clinical toolkits.

Subsequently, the commercial copyright for the official authorized versions, scoring translations, and digital distributions of the HADS was acquired and managed exclusively by GL Assessment (part of Renaissance Learning) in the United Kingdom. Researchers and healthcare organizations wishing to deploy the official HADS in academic research, clinical trials, or electronic health record (EHR) platforms are historically required to obtain user licenses and pay associated royalty fees through GL Assessment.

However, significant scholarly and legal dialogue exists internationally regarding fair academic use. The original 14 items were published directly in the public academic literature in 1983 without redaction. While official proprietary translation bundles, scoring software, and certified clinical forms remain strictly protected intellectual property of GL Assessment, independent researchers frequently cite fair-use provisions for non-commercial academic research and non-profit public health screening. Practitioners and researchers are advised to verify regional intellectual property requirements and license compliance with GL Assessment before deploying the instrument within institutional, commercial, or enterprise health systems.

References

Bjelland, I., Dahl, A. A., Haug, T. T., & Neckelmann, D. (2002). The validity of the Hospital Anxiety and Depression Scale: An updated literature review. Journal of Psychosomatic Research, 52(2), 69–77. https://doi.org/10.1016/S0022-3999(01)00296-3

Caci, H., Baylé, F. J., Mattei, V., Boyer, P., & Dos Reis, G. (2003). How does the Hospital and Anxiety Depression Scale measure anxiety and depression in a normal population? Psychiatry Research, 118(1), 89–99. https://doi.org/10.1016/S0165-1781(03)00062-8

Clark, L. A., & Watson, D. (1991). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 100(3), 316–336. https://doi.org/10.1037/0021-843X.100.3.316

Dunbar, M., Ford, G., Hunt, K., & Der, G. (2000). A confirmatory factor analysis of the Hospital Anxiety and Depression Scale: Comparing empirical and theoretical models. British Journal of Clinical Psychology, 39(1), 79–94. https://doi.org/10.1348/014466500163121

Herrmann, C. (1997). International experiences with the Hospital Anxiety and Depression Scale—a review of validation data and clinical results. Journal of Psychosomatic Research, 42(1), 17–41. https://doi.org/10.1016/S0022-3999(96)00216-4

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

Norton, S., Cosco, T., Doyle, F., Done, J., & Sacker, A. (2013). The Hospital Anxiety and Depression Scale: A meta confirmatory factor analysis. Journal of Psychosomatic Research, 74(1), 74–81. https://doi.org/10.1016/j.jpsychores.2012.10.010

Razavi, D., Delvaux, N., Farvacques, C., & Robaye, E. (1990). Screening for adjustment disorders and major depressive disorders in cancer in-patients. British Journal of Psychiatry, 156(1), 79–83. https://doi.org/10.1192/bjp.156.1.79

Singer, S., Kuhnt, S., Götze, H., Hauss, J., Hinz, A., Liebmann, A., Krauss, O., Lehmann, A., & Schwarz, R. (2008). Hospital Anxiety and Depression Scale Cutoff Scores for Cancer Patients in Acute Care. British Journal of Cancer, 100(6), 908–912. https://doi.org/10.1038/sj.bjc.6604952

Snaith, R. P. (2003). The Hospital Anxiety and Depression Scale. Health and Quality of Life Outcomes, 1(1), 29. https://doi.org/10.1186/1477-7525-1-29

Zigmond, A. S., & Snaith, R. P. (1983). The Hospital Anxiety and Depression Scale. Acta Psychiatrica Scandinavica, 67(6), 361–370. https://doi.org/10.1111/j.1600-0447.1983.tb09716.x

13. Items of the Scale (Questionnaire)