Abstract
The Alzheimer’s Disease Related Quality of Life (ADRQL) is a multidimensional, proxy-respondent psychometric instrument specifically developed to evaluate the health-related quality of life (HRQoL) of individuals diagnosed with Alzheimer’s disease and related progressive neurodegenerative dementias. Conceptualized and validated by Peter V. Rabins, Judith D. Kasper, and colleagues at the Johns Hopkins University, the instrument addresses the methodological limitations of self-report measures in populations with moderate-to-severe cognitive impairment. The ADRQL operationalizes quality of life through observable behaviors rather than subjective internal emotional states that require abstract introspection.
The instrument evaluates five distinct behavioral domains: Social Interaction, Awareness of Self, Feelings and Mood, Enjoyment of Activities, and Response to Surroundings. The measure utilizes a dichotomous response format (Agree / Disagree), completed by knowledgeable proxy informants such as family caregivers or institutional nursing staff who observe the patient across standard two-week assessment windows. Extensive psychometric evaluations demonstrate high internal consistency across subscales (Cronbach’s α ranging between 0.76 and 0.90) and acceptable test-retest reliability across clinical and community settings. Construct validity is supported by significant convergent associations with neuropsychiatric disturbance indices and activities of daily living, while preserving divergence from purely biological markers of cognitive staging. Weighting algorithms developed through standard gamble and conjoint analysis ensure that domain scores reflect explicit societal and patient-centered valuations, establishing the ADRQL as a gold standard in dementia clinical trials and health services research.
Keywords
Alzheimer’s disease, Quality of Life, ADRQL, Psychometrics, Dementia Assessment, Proxy-Report, Behavioral Neurology, Geriatric Psychiatry, Observational Rating Scale, Health Economics, Scale Validation, Neurocognitive Disorders
Authors
The development, psychometric formalization, and clinical standardization of the ADRQL were spearheaded by an interdisciplinary team of investigators at Johns Hopkins University:
- Peter V. Rabins, M.D., M.P.H.: Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine, Baltimore, Maryland. Co-founder of the Geriatric Psychiatry Program and pioneer in dementia neuropsychiatry.
- Judith D. Kasper, Ph.D.: Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland. Expert in health status measurement and long-term care epidemiology.
- Betty S. Black, Ph.D.: Department of Psychiatry and Behavioral Sciences, Division of Geriatric Psychiatry and Neuropsychiatry, Johns Hopkins University School of Medicine. Principal investigator overseeing instrument maintenance and licensing protocols (Email contact: [email protected]).
- Lawrence S. Kleinman, M.D., M.P.H.: Health services researcher and pediatric/preventive medicine methodologist involved in early psychometric calibration.
- Donald L. Patrick, Ph.D., M.S.P.H.: Department of Health Services, University of Washington, Seattle, Washington. World-renowned authority on health-related quality of life measurement and health outcomes methodology.
Purpose
The primary clinical and epidemiological objective of the Alzheimer’s Disease Related Quality of Life (ADRQL) measure is to provide a reliable, objective, and clinically sensitive assessment of well-being for individuals across all stages of progressive dementia, extending through moderate to profound cognitive deterioration. Evaluating quality of life within neurodegenerative illnesses introduces severe psychometric challenges. Traditional HRQoL instruments depend heavily on introspective self-reporting, semantic comprehension, working memory, and temporal orientation. As neurofibrillary tangles and amyloid plaques spread across the neocortex and hippocampus, individuals with Alzheimer’s disease experience progressive anosognosia (impaired disease awareness), communicative aphasia, and episodic memory collapse. Consequently, demanding that individuals synthesize their emotional well-being over a past timeframe using standard Likert scales produces profound measurement error, uninterpretable floor effects, or missing data.
The ADRQL overcomes these barriers by shifting the measurement paradigm from internal introspective evaluation to direct, observable behavioral assessment conducted by regular caregivers. Grounded in behavioral ecology and observable neuropsychiatric signs, the ADRQL captures observable expressions of comfort, distress, social connectedness, environmental engagement, and self-identity. It was designed to answer whether non-pharmacological interventions, pharmacotherapies, environmental redesigns, or supportive care protocols translate into tangible improvements in everyday lived experiences, regardless of whether underlying neurodegenerative trajectory can be halted.
In clinical trials, the ADRQL functions as an indispensable secondary or co-primary outcome indicator alongside classical cognitive indices, such as the Mini-Mental State Examination (MMSE) and the Alzheimer’s Disease Assessment Scale–Cognitive Subscale (ADAS-Cog). In health services research and health economics, the ADRQL provides domain-specific and preference-weighted indices necessary to calculate quality-adjusted life years (QALYs) specifically attuned to the institutional and community realities of dementia care. Finally, in clinical psychiatric practice, it serves as an actionable framework for multidisciplinary teams to detect specific behavioral deficits, such as social withdrawal or agitation, and tailor holistic psychosocial care plans.
Psychological Construct
The overarching construct captured by the ADRQL is Health-Related Quality of Life in Dementia, defined as a multidimensional phenomenon manifested through observable behavioral responses reflecting an individual’s affective, social, relational, and physical interaction with their surrounding environment. Recognizing that quality of life is not a monolithic dimension, the instrument partitions this global construct into five interconnected yet distinct behavioral domains:
1. Social Interaction
This subscale captures the individual’s behavioral drive, willingness, and affective responsiveness in interpersonal contexts. Human beings remain innately social organisms, yet progressive cortical dementia often provokes profound social withdrawal (apathy) or hyper-reactive defensive behaviors. The items in this domain evaluate observable physical, verbal, and affective markers of interpersonal engagement. Specific behaviors include smiling or laughing when around other people, actively seeking physical proximity rather than retreating into isolation, initiating conversation, permitting affectionate physical contact (e.g., handshakes, embraces, comforting pats), and displaying receptive comfort when soothed by loved ones or professional staff. It also captures social receptivity to non-human stimuli, such as pleasant interactions with domestic pets or visiting children, and the capacity to enjoy group activities passively without demanding active functional performance.
2. Awareness of Self
Preservation of personhood, biographical identity, and personal agency represents a cornerstone of ethical and psychological well-being in dementia care. The Awareness of Self domain operationalizes personhood by examining behaviors indicating an ongoing connection to one’s life history, social roles, and current autonomy. Indicators include speaking about or acting out historical occupational routines, expressing clear awareness of familial status (such as identifying oneself as a spouse, mother, or grandfather), asserting personal preferences regarding routine daily choices (selecting preferred garments, choosing specific meals, or claiming a favorite chair), and retaining orienting responses to one’s own spoken name. Conversely, loss of self-awareness is indexed through persistent failure to express long-held moral attitudes, spiritual beliefs, or abandonment of historically cherished personal possessions that previously conferred security and emotional anchoring.
3. Feelings and Mood
Affective well-being in cognitive impairment cannot be reliably gathered via self-administered depression scales. The ADRQL operationalizes affect through observable psychomotor and non-verbal behavioral manifestations of anxiety, frustration, anger, contentment, and dysphoria over a specified two-week observational window. The construct captures distressing behavioral signs such as manual restlessness (wringing, twisting, or squeezing hands), vocal and aggressive outbursts (cursing, calling out, shrieking, leveling accusations), motoric combativeness (hitting, kicking, banging objects, pushing caregivers), and persistent affective dysregulation (frequent weeping, unprovoked anger, resisting essential personal hygiene activities). Symmetrically, the domain captures positive emotional balance, marked by outward expressions of calm satisfaction, peaceful posture, and general contentment within the living environment.
4. Enjoyment of Activities
This dimension operationalizes behavioral engagement and hedonic capacity through participation in structured, unstructured, solitary, or communal recreational pursuits. Apathy and anhedonia are pervasive neurobehavioral sequelae of frontal-subcortical and limbic disconnection in Alzheimer’s disease. This construct measures whether an individual sustains an active interest in preferred solitary pastimes (e.g., listening to classical music, watching television programs, sorting fabrics), displays discernible visual and behavioral expressions of pleasure when introduced to recreational activities, or exhibits profound behavioral demotivation characterized by chronic daytime hypersomnolence, prolonged motionless staring, or refusal to participate even when structured supportive prompting is provided.
5. Response to Surroundings
The physical and ambient setting exerts a profound impact on individuals experiencing neurocognitive fragmentation. This domain reflects an individual’s sense of environmental safety, orienting stability, and ecological comfort. Behavioral markers include expressing vocalized fears regarding personal security or theft of possessions, exhibiting acute confusion and motoric distress when transported outside familiar geographic settings, verbalizing anxious wandering desires (“I want to go home” despite being home), and voicing extreme existential anguish or suicidal thoughts. The subscale directly measures the degree to which an individual feels ecologically integrated or chronically alienated within their architectural and social setting.
Theoretical Framework
The design of the ADRQL is structurally anchored in three synergistic theoretical paradigms: the Need-Driven Dementia-Compromised Behavior (NDB) Model, Kitwood’s model of Dementia Care Mapping and Personhood, and modern Health Services Utility Theory.
The Need-Driven Dementia-Compromised Behavior (NDB) Model
Formulated by Algase and colleagues (1996), the NDB framework posits that disruptive, withdrawn, or unusual behaviors manifested by individuals with cognitive deficits are not merely meaningless random neurochemical byproducts of cerebral atrophy. Instead, these behaviors represent meaningful, goal-directed attempts to express unmet physical, physiological, or psychosocial needs within an environment that outstrips their neurocognitive processing capacities. In the ADRQL, behavioral indicators such as resisting care, pacing, or crying out are theorized as direct indices of subjective distress and unmet environmental equilibrium. Measuring these behaviors allows investigators to quantify the subjective experience of the patient through an ecologically valid behavioral lens.
Kitwood’s Dialectical Psychosocial Model of Personhood
Tom Kitwood’s seminal work at the Bradford Dementia Group fundamentally altered contemporary neurocare by conceptualizing dementia as a dialectical interplay between neurological impairment (neuropathology) and the interpersonal environment (malignant vs. supportive social psychology). Kitwood asserted that “personhood”—a status of standing or trust bestowed upon an individual by others—can be sustained even in severe cognitive collapse if the social environment supports five core psychological needs: comfort, attachment, inclusion, occupation, and identity. The ADRQL’s subscales align with Kitwood’s five structural pillars. By capturing whether caregivers comfort the patient, allow touch, preserve identity, and encourage occupation, the instrument measures the degree to which the patient’s lived reality reflects maintained personhood.
Health Economics and Multi-Attribute Utility Theory
Unlike purely clinical psychiatric scales that sum raw symptom counts, the ADRQL was simultaneously built upon Multi-Attribute Utility Theory (MAUT). HRQoL is fundamentally an evaluative construct reflecting human preferences for specific health states. The ADRQL incorporates explicit utility-based weighting algorithms. During development, Rabins, Kasper, and colleagues engaged expert panels, professional caregivers, and family members in formal decision-analytic elicitation protocols (including conjoint analysis and standard gamble tasks). These empirical weights reflect the relative social and clinical importance assigned to preserving social connectedness versus eliminating restlessness, transforming observed behaviors into a standardized, decision-analytic metric suitable for cost-effectiveness research.
Validity
The psychometric validity of the ADRQL has been demonstrated across multiple rigorous empirical validation trials in North American, European, and Asian cohorts, spanning home care settings, assisted living facilities, and specialized inpatient memory care units.
Construct and Convergent Validity
Construct validity was formally demonstrated by establishing hypothesized correlations between ADRQL subscale dimensions and validated external clinical criteria. In the validation cohort published by Kasper et al. (2009) involving diverse institutional and community settings, the ADRQL demonstrated significant moderate-to-high convergent correlations with the Neuropsychiatric Inventory (NPI). Specifically, the ADRQL Feelings and Mood and Response to Surroundings subscales correlated inversely with total NPI behavioral disturbance scores (r = -0.52 to -0.68, p < 0.001), showing that higher quality of life aligns with lower behavioral pathology.
Furthermore, convergent validity with depression measures was verified through significant associations with the Cornell Scale for Depression in Dementia (CSDD). Patients scoring in the clinically depressed range on the CSDD exhibited significantly depressed scores on the ADRQL Feelings and Mood subscale (Cohen’s d = 0.84, p < 0.001). Functional performance measures, such as the Psychogeriatric Dependency Rating Scale (PGDRS) and the Activities of Daily Living (ADL) Barthel Index, demonstrated selective, moderate correlations with Enjoyment of Activities (r = 0.41, p < 0.01) and Awareness of Self (r = 0.48, p < 0.001), reinforcing that functional autonomy contributes to, but does not completely determine, global quality of life.
Discriminant and Known-Groups Validity
Discriminant validity is demonstrated by the ADRQL’s relationship to biological indices of cognitive decline, such as the Mini-Mental State Examination (MMSE). While traditional HRQoL tools show artificial floor effects as MMSE scores drop below 10, the ADRQL preserves substantial score variance across severe cognitive strata. The correlation between the ADRQL global summary index and MMSE scores is typically weak-to-moderate (r = 0.20 to 0.32), proving that the ADRQL measures an experiential and affective construct distinct from raw neurocognitive performance. Patients with profound cognitive impairment can still experience positive quality of life if supported by an enriching, non-threatening social environment.
Known-groups validity was demonstrated by comparing patients living in specialized, sensory-adapted dementia units against those in non-adapted traditional institutional wards. As hypothesized, the ADRQL successfully distinguished between these environments, registering statistically superior scores in Response to Surroundings and Social Interaction for individuals residing in specialized environments (F = 14.22, p < 0.001), controlling for baseline cognitive impairment.
Reliability
The ADRQL exhibits robust psychometric reliability across internal consistency, inter-rater concordance, and temporal test-retest assessments:
Internal Consistency
Across validation cohorts, internal consistency reliability coefficients (Cronbach’s alpha) for the five individual subscales and the total composite scale have consistently met or exceeded established psychometric thresholds for clinical decision-making:
- Social Interaction: α = 0.84 to 0.88
- Awareness of Self: α = 0.76 to 0.82
- Feelings and Mood: α = 0.83 to 0.89
- Enjoyment of Activities: α = 0.77 to 0.83
- Response to Surroundings: α = 0.79 to 0.85
- Total ADRQL Scale Index: α = 0.88 to 0.92
These values demonstrate that items within each behavioral domain reflect a coherent, unified construct without exhibiting redundant item collinearity.
Test-Retest and Inter-Rater Reliability
Given that the ADRQL relies on proxy respondents, evaluating temporal stability and inter-informant concordance is essential. Test-retest stability was evaluated by administering the tool across a 7- to 14-day interval to the same proxy informants in stable clinical cohorts. Intraclass correlation coefficients (ICCs) ranged from 0.78 to 0.91 across all subscales, indicating stability in the absence of acute clinical interventions.
Inter-rater reliability was evaluated in institutional settings by comparing independent ratings generated by two professional nursing staff members who observed the same resident during the identical two-week period. Inter-rater ICCs across domains ranged from 0.69 to 0.82. Concordance between family proxies and professional institutional caregivers tends to be slightly lower (ICCs between 0.58 and 0.71), reflecting differences in observational contexts (e.g., brief personal visits vs. continuous round-the-clock nursing care), a common phenomenon in proxy-based assessment literature.
Factor Analysis
The structural dimensionality of the ADRQL was validated through rigorous Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) during its developmental and standardization phases.
Exploratory Factor Analysis (EFA)
Initial developmental cohorts underwent principal axis factoring and principal component analysis utilizing oblique rotation (Promax and Oblimin), reflecting the theoretical assumption that quality of life dimensions are intercorrelated rather than completely orthogonal. EFA yielded clear five-factor structures matching the hypothesized theoretical framework: Social Interaction, Awareness of Self, Feelings and Mood, Enjoyment of Activities, and Response to Surroundings.
Eigenvalues for the first five factors comfortably exceeded the Kaiser-Guttman criterion (eigenvalues > 1.5), collectively accounting for more than 54% of the total observational variance. Items loaded strongly on their primary conceptual factors, with primary factor loadings typically ranging between 0.45 and 0.82, while secondary cross-loadings remained low (< 0.25).
Confirmatory Factor Analysis (CFA)
Subsequent psychometric evaluations (e.g., Kasper et al., 2009) confirmed this structure via CFA. Structural equation modeling testing the five-factor oblique measurement model yielded good goodness-of-fit indices:
- Comparative Fit Index (CFI): 0.92 to 0.95 (exceeding standard ≥ 0.90 thresholds)
- Tucker-Lewis Index (TLI): 0.91 to 0.94
- Root Mean Square Error of Approximation (RMSEA): 0.048 to 0.056 (90% CI [0.042, 0.063]), falling well within acceptable fit boundaries (< 0.06)
- Standardized Root Mean Square Residual (SRMR): 0.051
Comparative model testing firmly established that the five-factor multidimensional model fit the observational data significantly better than alternative unidimensional models (Δχ² p < 0.001) or hierarchical higher-order models without correlated sub-dimensions, verifying the conceptual distinctness of each behavioral domain.
Instrument / Measurement Tool
The practical administration, scoring architecture, and operational specifications of the ADRQL are detailed below:
- Instrument Type: Proxy-reported observational behavioral rating scale.
- Target Population: Individuals diagnosed with probable or possible Alzheimer’s disease, vascular dementia, frontotemporal lobar degeneration, Lewy body dementia, or mixed neurocognitive disorders across mild, moderate, severe, and profound stages.
- Informant / Respondent Requirements: A primary informal caregiver (spouse, adult child) or professional care provider (nurse, certified nursing assistant) who has had direct, regular contact with the patient for at least several days per week over the preceding two weeks.
- Timeframe of Observation: Past two weeks (14 days) prior to the date of administration.
- Response Format: Dichotomous scale: Agree (behavior observed during the 2-week period) or Disagree (behavior not observed during the 2-week period). Informants choose whether each behavioral statement accurately describes the individual.
- Domain Structure & Item Count:
- Social Interaction: 12 items
- Awareness of Self: 8 items
- Feelings and Mood: 12 items
- Enjoyment of Activities: 4 items
- Response to Surroundings: 4 items
- Scoring and Weighting Rules:
- Items are balanced between positive indicators of well-being (e.g., “smiles or laughs”, “shows delight”) and negative indicators of distress (e.g., “throws, hits, kicks”, “says he/she wants to die”).
- Negative items are reverse-scored so that across all metrics, a higher numerical score consistently indicates better health-related quality of life.
- The instrument incorporates an empirically derived mathematical weighting system. Each item is multiplied by an established expert/caregiver utility weight.
- Weighted domain scores are calculated by summing weighted items, dividing by the maximum possible domain weight, and multiplying by 100, generating normalized subscale scores ranging from 0 to 100.
- A Total ADRQL Index Score (0–100) is calculated by computing the weighted average across all completed domains. Missing items are managed via domain-specific mean substitution if less than 20% of a given domain is unrated; otherwise, the domain is deemed invalid.
Permissions & Fee and Test Year
The foundational concepts, methodological architecture, and preliminary validation studies of the ADRQL were first published in 1999 by Dr. Peter V. Rabins, Dr. Judith D. Kasper, and colleagues at the Johns Hopkins University. A comprehensive psychometric re-evaluation and expanded standardization study followed in 2009.
The ADRQL is a copyrighted psychometric instrument held by the Johns Hopkins University. The instrument is accessible for academic research, educational initiatives, and non-commercial clinical practices. Researchers and clinicians wishing to use the ADRQL must request permission and obtain the official administration and scoring manual by contacting:
- Contact Person: Betty S. Black, Ph.D.
- Affiliation: Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine
- Email: [email protected]
- Official Resource Portal: Johns Hopkins Geriatric Psychiatry ADRQL Center
Commercial entities, pharmaceutical corporations, and contract research organizations (CROs) conducting funded trials are subject to institutional licensing agreements and associated administrative fees determined by the Johns Hopkins Technology Ventures office.
References
- Algase, D. L., Beck, C., Kolanowski, A., Whall, A., Berent, S., Richards, K., & Beattie, E. (1996). Need-driven dementia-compromised behavior: An alternative view of disruptive behavior. American Journal of Alzheimer’s Disease and Other Dementias, 11(6), 10–19. https://doi.org/10.1177/153331759601100603
- Black, B. S., Johnston, D., Rabins, P. V., Morrison, A., Lyketsos, C., & Samus, Q. M. (2012). Unmet needs of community-residing persons with dementia and their informal caregivers: The MIND-Montgomery County project. The American Journal of Geriatric Psychiatry, 21(6), 514–525. https://doi.org/10.1016/j.jagp.2012.10.020
- Kasper, J. D., Black, B. S., Shore, A. D., & Rabins, P. V. (2009). Evaluation of the validity and reliability of the Alzheimer’s disease related quality of life (ADRQL) assessment instrument. Alzheimer Disease and Associated Disorders, 23(3), 275–284. https://doi.org/10.1097/WAD.0b013e31819fe99b
- Kitwood, T. (1997). Dementia reconsidered: The person comes first. Open University Press.
- Rabins, P. V., Kasper, J. D., Kleinman, L., Black, B. S., & Patrick, D. L. (1999). Concepts and methods in the development of the ADRQL: An instrument for assessing health-related quality of life in persons with Alzheimer disease. Journal of Mental Health and Aging, 5(1), 33–48.
- Rabins, P. V., Kasper, J. D., & Black, B. S. (2000). The Alzheimer’s Disease Related Quality of Life (ADRQL) instrument: User’s manual. Johns Hopkins University.
- Shroff, H. B. (2014). Family resiliency, sense of coherence, social support and psychosocial interventions: Reducing caregiver burden and determining the quality of life in persons with Alzheimer’s disease (Doctoral dissertation, Florida Atlantic University). Florida Atlantic University Digital Commons. https://fau.digital.flvc.org/islandora/object/fau%3A28906