Palliative Care Questionnaire / Palliative Care Outcome Scale

Abstract

The Palliative Care Outcome Scale (POS)—known in Dutch validation and clinical research as the Vragenlijst voor de palliatieve zorg—is a pioneering, multidimensional patient-reported outcome measure (PROM) and clinical assessment tool developed by Julia Hearn and Irene J. Higginson in 1999 at King’s College London. Designed specifically to capture the holistic needs of patients with advanced, life-limiting illnesses and their families, the POS measures physical symptom burden, psychological distress, spiritual and existential well-being, practical concerns, and information/communication satisfaction over a 3-day recall window. The instrument consists of 10 core psychometric items rated on a 5-point Likert scale (ranging from 0 to 4), supplemented by two open-ended qualitative prompts designed to capture idiosyncratic patient concerns and clinical priorities. A defining structural hallmark of the POS is its dual-perspective parallel design, comprising both a patient self-report questionnaire and an independent healthcare professional (proxy/staff) version, facilitating multi-informant discrepancy analysis in palliative and end-of-life care settings.

Extensive international validation across oncology, advanced organ failure, and neurodegenerative conditions demonstrates robust psychometric integrity. Construct validity is supported by significant moderate-to-high correlations with legacy instruments, including the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30), the Support Team Assessment Schedule (STAS), and the Hospital Anxiety and Depression Scale (HADS). The instrument exhibits acceptable internal consistency across its multifaceted domains (Cronbach’s alpha typically ranging from 0.65 to 0.78 for total score composites, reflecting the intentionally heterogeneous nature of palliative constructs, with subscale alphas reaching 0.75 to 0.88), excellent test-retest reliability (intraclass correlation coefficients [ICC] > 0.75), and marked responsiveness to clinical interventions within specialized palliative care. This article provides a comprehensive academic review of the POS, examining its theoretical architecture, psychometric properties, factor structure, cross-cultural adaptation into Dutch, and clinical utility.

Keywords

Palliative Care Outcome Scale, POS, Vragenlijst voor de palliatieve zorg, Patient-Reported Outcome Measures, End-of-Life Care, Total Pain, Psychometrics, Irene Higginson, Quality of Life, Symptom Burden, Dual-Informant Assessment

Authors

The Palliative Care Outcome Scale was conceptualized, developed, and initially validated by an interdisciplinary research team led by:

• Julia Hearn, PhD — Department of Palliative Care and Policy, Guy’s, King’s and St Thomas’ School of Medicine, King’s College London, United Kingdom.
• Irene J. Higginson, PhD, BMBS, FFPHM, FRCP — Professor of Palliative Care and Policy, Cicely Saunders Institute of Palliative Care, Policy & Rehabilitation, King’s College London, United Kingdom.

The Dutch translation and cross-cultural validation—titled the Vragenlijst voor de palliatieve zorg—was spearheaded by academic palliative care researchers in the Netherlands and Belgium in collaboration with the Cicely Saunders Institute, including investigators affiliated with the Amsterdam University Medical Centers (VUmc), Erasmus MC, and the Center for Palliative Care (Centrum voor Palliatieve Zorg), seeking to validate the tool within Dutch palliative consultation services, hospices, and hospital inpatient units.

Purpose

The fundamental purpose of the Palliative Care Outcome Scale (POS) / Vragenlijst voor de palliatieve zorg is to provide a standardized, clinically actionable, and psychometrically robust mechanism for capturing the multidimensional health status and unmet needs of patients receiving palliative and end-of-life care. Historically, palliative medicine relied heavily on performance status metrics (such as the Karnofsky Performance Scale) or disease-specific oncological quality-of-life inventories (such as the EORTC QLQ-C30 or FACT-G). While valuable, these legacy instruments were often excessively lengthy (causing substantial respondent burden for frail, deteriorating individuals), failed to evaluate non-physical domains (e.g., spiritual angst, time lost to medical appointments), and neglected the essential perspectives of informal caregivers and families.

Hearn and Higginson (1999) engineered the POS to address these critical clinical and methodological voids. Specifically, the instrument serves three complementary functions across clinical and research paradigms:

• Routine Point-of-Care Clinical Assessment: To quickly identify, prioritize, and monitor the physical, emotional, and social needs of the patient during clinical consultations, home visits, or inpatient ward rounds, alerting the interdisciplinary team to hidden distress.
• Clinical Audit and Quality Improvement: To serve as a standardized outcome benchmark across health services, palliative care networks, and hospices, allowing programs to evaluate quality of care, track longitudinal outcomes, and detect systemic deficits in symptom management or communication.
• Comparative Effectiveness and Intervention Research: To provide an empirically sensitive outcome measure for clinical trials evaluating novel palliative interventions, complex supportive care models, pharmacological protocols, and holistic psychosocial therapies.

Crucially, the POS introduces a dual-perspective framework featuring a Patient Version (self-completed by the patient whenever cognitive and physical capacities permit) and an independent Staff Version (completed by the primary treating nurse or physician). Both versions operationalize a short recall period—the preceding 3 days (or past 48 to 72 hours)—which minimizes recall bias and accommodates the rapidly fluctuating clinical trajectories inherent to advanced terminal illnesses. By comparing patient and staff ratings, clinicians can immediately identify communication discrepancies, unrecognized suffering, and under- or over-estimated symptom severity.

Psychological Construct

The POS operationalizes a complex, multidimensional construct encapsulating the multifaceted experience of advanced life-limiting illness. Rather than viewing distress as an isolated physiological manifestation, the instrument treats quality of dying and living as an integrated biopsychosocial-spiritual phenomenon. The construct decomposes into five distinct yet interrelated dimensions across its core items:

1. Physical Symptom Burden and Somatic Distress

This dimension captures the severity and impact of both primary nociceptive/neuropathic pain and other debilitating secondary physical symptoms (such as nausea, vomiting, dyspnea, fatigue, constipation, severe pruritus, or cachexia). Rather than merely logging the physiological presence of a symptom, the construct measures the degree to which physical suffering intrudes upon the individual’s consciousness and daily functioning over the preceding 3 days.

2. Psychological and Emotional Well-Being

The psychological domain evaluates core affective distress, focusing predominantly on two pervasive emotional states: severe anxiety and depressive despair. Anxiety is operationalized through questions probing patient worry, dread, and apprehension regarding their illness, trajectory, and anticipated death. Depressive affect is captured through feelings of pervasive sadness, low mood, or feeling that “things were getting on top of” the individual. Importantly, the construct also measures family/carer anxiety as perceived by the patient and staff, recognizing that psychological suffering in palliative contexts is fundamentally systemic and relational.

3. Spiritual, Existential, and Meaning-Based Concerns

Palliative care constructs must acknowledge existential dissolution. The POS targets spiritual and existential distress by assessing whether the patient feels “at peace” and whether they perceive life as retaining meaning, purpose, or worthwhileness despite severe biological decline. This existential axis reflects modern empirical models of demoralization, dignity-related distress, and spiritual coping, providing insight into whether the patient is experiencing existential isolation, loss of dignity, or spiritual serenity.

4. Practical, Financial, and Logistical Burdens

Advanced illness often forces patients and their families to navigate complex, fragmented medical ecosystems. The POS captures the pragmatic realities of illness by assessing practical problems (such as financial strain, transport difficulties, home care logistics, or social dependency) and evaluating “wasted time”—the existential and practical toll incurred when patients spend substantial portions of their remaining days attending bureaucratic medical appointments, waiting for transport, or navigating inefficient care systems.

5. Information Sharing and Communication Satisfaction

The communication dimension evaluates whether the patient has been provided with as much information as they desire regarding their condition, prognosis, and treatment options. The construct recognizes that information needs are idiosyncratic: some individuals seek exhaustive prognostic details, whereas others prefer selective avoidance. Consequently, the construct evaluates congruence between desired and received information, as well as the ability of the patient to openly discuss their fears and feelings with their family and social network.

Theoretical Framework

The theoretical architecture of the POS is anchored in two foundational concepts: Dame Cicely Saunders’ multidimensional theory of “Total Pain” and Avedis Donabedian’s classic healthcare quality framework of Structure, Process, and Outcomes.

Cicely Saunders’ Construct of “Total Pain”

In founding modern palliative medicine, Dame Cicely Saunders postulated that human suffering at the end of life cannot be understood simply through the neurophysiology of tissue damage. Instead, Saunders formulated the concept of Total Pain, asserting that pain and distress are an amalgam of four inextricably linked domains: physical pain, psychological pain (e.g., terror, loss of autonomy, grief), social pain (e.g., loss of social roles, economic catastrophe, interpersonal strain), and spiritual/existential pain (e.g., crisis of faith, questioning the meaning of one’s existence, unresolved guilt). The POS directly reflects Saunders’ model by weighting physical, psychological, social, and spiritual items equally within a single unified clinical index, preventing the reductionist biological bias common in acute healthcare metrics.

Donabedian’s Quality of Care Framework

Methodologically, the POS operationalizes Avedis Donabedian’s quality-of-care paradigm. Palliative services often evaluate their performance using structural indicators (e.g., nurse-to-patient ratios, hospice bed availability) or process indicators (e.g., completion of advance care directives, time to first clinical consultation). However, Donabedian argued that the ultimate measure of clinical quality is the outcome: changes in health status, symptom relief, and subjective well-being directly attributable to care. The POS was engineered as a true outcome measurement tool, sensitive to longitudinal shifts resulting from interdisciplinary palliative care interventions.

Multi-Informant Discrepancy Theory

The conceptual framework of the POS also incorporates multi-informant discrepancy theory. Terminal illness frequently precipitates cognitive impairment, severe weakness, or delirium, rendering patients intermittently unable to self-report. By creating structurally parallel Patient and Staff versions, the POS operationalizes the dual-observer paradigm. In classical measurement theory, proxy reports were traditionally viewed as imperfect substitutes for patient self-reports. However, the theoretical framework of the POS treats the staff report not as an error-prone surrogate, but as a complementary, independent observational perspective. Discrepancies between patient self-reports and staff ratings serve as valuable clinical signals: for instance, staff systematically underestimate existential distress and family anxiety while overestimating physical functional limitations compared to patient self-ratings.

Validity

The psychometric validity of the POS—both in its original English version (Hearn & Higginson, 1999) and its validated Dutch translation (Vragenlijst voor de palliatieve zorg)—has been empirically established across diverse inpatient, outpatient, and community palliative care cohorts.

Content and Face Validity

Content validity was established through rigorous qualitative and Delphi consensus methods involving palliative care clinicians, oncologists, nurses, social workers, spiritual caregivers, and palliative patients and their informal caregivers. Initial item generation prioritized relevance, comprehensiveness, and minimizing respondent burden. Pilot testing confirmed that patients and healthcare providers considered the items highly relevant to the realities of advanced illness, praising the inclusion of existential well-being and practical disruption alongside standard somatic symptoms.

Construct Validity: Convergent and Discriminant

Construct validity has been demonstrated through extensive cross-instrument correlation with established, legacy psychometric inventories:

• Physical Symptoms: The POS physical symptom items (pain and other symptoms) correlate moderately to strongly with the pain and physical functioning subscales of the EORTC QLQ-C30 (Pearson’s r = 0.58 to 0.72, p < 0.001) and the physical distress scales of the Edmonton Symptom Assessment System (ESAS; r = 0.65 to 0.76).
• Psychological Distress: The POS anxiety and depression items exhibit strong convergent validity with the Hospital Anxiety and Depression Scale (HADS-A and HADS-D), yielding correlation coefficients ranging from r = 0.60 to 0.74 (p < 0.001).
• Discriminant Validity: Discriminant validity is supported by weaker correlations between non-overlapping domains. For example, correlations between the POS “wasted time/appointments” item and the EORTC physical nausea scales are negligible (r < 0.20), confirming that the instrument taps distinct, non-redundant facets of the patient experience.

Criterion and Concurrent Validity

In concurrent validity studies, the POS demonstrated high agreement with the clinician-rated Support Team Assessment Schedule (STAS). Spearman rank correlations between parallel items on the POS and STAS typically range between r_s = 0.52 and 0.78. In the Dutch validation cohorts, concurrent assessments against the Dutch versions of the McGill Quality of Life Questionnaire (MQOL-NL) demonstrated robust convergence across overall quality-of-life composites.

Responsiveness to Clinical Change (Sensitivity)

A critical metric for any palliative PROM is responsiveness—the ability to detect clinically meaningful changes over time. Longitudinal trials evaluating specialized palliative team consultations demonstrate that the POS exhibits moderate-to-large effect sizes (Cohen’s d ranging from 0.45 to 0.82) within 48 to 72 hours of targeted palliative intervention, particularly in the relief of severe pain, dyspnea, and psychological crisis.

Reliability

Evaluating reliability in palliative and end-of-life populations introduces distinctive methodological complexities. Because clinical trajectories in advanced illness can shift rapidly, traditional test-retest intervals (such as two weeks) are inappropriate, as biological deterioration or acute clinical crises represent true clinical change rather than measurement error.

Internal Consistency

When evaluated as an aggregated unidimensional scale, the 10-item POS typically demonstrates moderate internal consistency. In the foundational validation study by Hearn and Higginson (1999), Cronbach’s alpha was calculated at α = 0.65 for the patient version and α = 0.70 for the staff version. Subsequent large-scale multicenter studies (e.g., Bausewein et al., 2008; Collins et al., 2015) have reported composite alphas ranging between 0.68 and 0.78.

Psychometricians emphasize that an alpha between 0.65 and 0.75 is theoretically expected and psychometrically appropriate for an instrument measuring heterogeneous clinical domains (pain, existential peace, financial issues, communication). A hyper-inflated alpha (> 0.90) would indicate item redundancy, which is clinically counterproductive in short screening instruments for frail populations. When subscales are isolated (such as the physical symptom cluster versus the psychosocial cluster), subscale internal consistency coefficients improve, demonstrating alphas between 0.75 and 0.85.

Test-Retest Reliability

Test-retest stability has been evaluated across short time windows (typically 24 to 48 hours) among stable palliative outpatients exhibiting no changes in medication or clinical status. Intraclass Correlation Coefficients (ICC) across individual items range from 0.73 to 0.89, and total score ICCs frequently exceed 0.82, demonstrating excellent measurement reproducibility in clinically stable cohorts.

Inter-Rater and Proxy-Patient Concordance

Inter-rater reliability between independent clinical observers (e.g., primary nurse versus palliative physician) completing the Staff Version is high, with weighted kappa coefficients (κ_w) ranging from 0.62 to 0.84. Concordance between the Patient Version and Staff Version displays systematic patterns: physical symptoms demonstrate moderate-to-high agreement (κ_w = 0.55 to 0.70), while internal psychological and existential domains (anxiety, depression, feeling at peace) display lower concordance (κ_w = 0.35 to 0.52), reinforcing the clinical necessity of obtaining patient self-report whenever possible.

Factor Analysis

The structural dimensionality of the POS has been extensively scrutinized using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) across diverse linguistic and cultural adaptations, including Dutch, German, Spanish, Portuguese, and Italian populations.

Exploratory Factor Analysis (EFA)

Initial exploratory factor analyses utilizing principal component analysis with varimax and oblimin rotations universally reject a strict unidimensional structure. Instead, EFA consistently extracts a two-factor or three-factor latent structure accounting for 52% to 64% of total variance across cohorts:

• Factor 1: Physical and Functional Symptoms: Heavily loaded by Item 1 (Pain severity) and Item 2 (Other physical symptom burden), with cross-loadings occasionally observed for Item 9 (Time wasted on health/appointments). Typical factor loadings range from 0.68 to 0.84.
• Factor 2: Psychosocial and Existential Well-Being: Heavily loaded by Item 3 (Patient anxiety), Item 4 (Family anxiety), Item 5 (Depressed mood), and Item 6 (Feeling at peace). Factor loadings for this cluster range from 0.62 to 0.79.
• Factor 3: Communication and Practical Concerns: Captures Item 7 (Information received), Item 8 (Ability to share feelings), and Item 10 (Practical/financial matters). In two-factor solutions, these items typically collapse into the broader psychosocial dimension.

Confirmatory Factor Analysis (CFA) and Model Fit

Structural equation modeling and CFA conducted on large palliative cohorts confirm that multi-factor models exhibit superior model fit compared to single-factor models. Standard goodness-of-fit indices for a two-factor correlated model (Physical/Somatic Distress vs. Psychosocial/Spiritual/Information Distress) consistently meet modern psychometric criteria:

• Comparative Fit Index (CFI): 0.94 to 0.97 (exceeding the ≥ 0.90 benchmark for acceptable fit).
• Tucker-Lewis Index (TLI): 0.92 to 0.95.
• Root Mean Square Error of Approximation (RMSEA): 0.045 to 0.062 (with 90% confidence intervals well below the 0.08 threshold for close structural fit).
• Standardized Root Mean Square Residual (SRMR): 0.048 to 0.058.

Item Response Theory (IRT) and Rasch modeling analyses further demonstrate that the 5-point Likert response categories operate monotonically, with appropriate category threshold ordering across all 10 core items, validating the standard 0 to 4 scoring continuum.

Instrument / Measurement Tool

The Palliative Care Outcome Scale is structured as follows:

• Instrument Type: Multidimensional Patient-Reported Outcome Measure (PROM) and parallel Clinician-Reported Outcome Measure (ClinRO).
• Target Populations: Adults and older adults diagnosed with advanced, active, progressive, life-limiting illnesses (malignant and non-malignant conditions, such as advanced cancer, congestive heart failure, COPD, end-stage renal disease, and amyotrophic lateral sclerosis).
• Recall Period: The preceding 3 days (or past 48 to 72 hours).
• Dual Versions:
  • Patient Version: Self-administered by the patient via paper, digital tablet, or clinical interview.
  • Staff / Proxy Version: Completed independently by the primary nurse, palliative physician, or multidisciplinary care professional.
• Core Item Inventory: 10 standardized quantitative psychometric items spanning:
  • Item 1: Pain burden and physical discomfort
  • Item 2: Other physical symptoms (e.g., breathlessness, nausea, fatigue)
  • Item 3: Patient anxiety and worry regarding illness
  • Item 4: Family and caregiver anxiety/distress
  • Item 5: Depressive feelings and low mood
  • Item 6: Spiritual/existential peace (feeling at peace)
  • Item 7: Information availability and communication satisfaction
  • Item 8: Interpersonal communication (sharing feelings with family/friends)
  • Item 9: Time wasted on healthcare logistics and medical appointments
  • Item 10: Practical and financial matters
• Qualitative Supplementary Prompts: 2 open-ended narrative items:
  • Item 11: Identification of the patient’s main individual problem or priority concern over the past 3 days.
  • Item 12: Space for the patient or staff member to log any other concerns, questions, or contextual comments.
• Response Scale & Scoring Mechanics:
  • Core items are rated on a standardized 5-point Likert scale ranging from 0 to 4.
  • Anchor descriptors vary by question content (e.g., 0 = Not at all to 4 = Overwhelmingly / Severely; or 0 = Always to 4 = Never).
  • Directionality & Reverse Scoring: Higher numerical scores uniformly indicate greater distress, higher symptom burden, or worse outcomes. For positively phrased items (such as Item 6 [feeling at peace], Item 7 [information given], and Item 8 [sharing feelings]), the response scale is reverse scored so that 0 represents optimal well-being (e.g., “Best / Always”) and 4 represents maximal impairment/distress.
  • Total Score Calculation: The global POS score is computed by summing the 10 individual items, yielding a composite score ranging from 0 to 40, with higher scores reflecting greater overall palliative need and distress. Subscale scores for physical and psychosocial clusters can also be reported independently.

Permissions & Fee and Test Year

The Palliative Care Outcome Scale was originally published in 1999 by Julia Hearn and Irene J. Higginson. The intellectual property and copyright of the POS instrument family—including the POS, the Integrated Palliative care Outcome Scale (IPOS), and all authorized international translations (including the Dutch Vragenlijst voor de palliatieve zorg)—are held by Cicely Saunders International and King’s College London.

The instrument is provided free of charge for non-commercial clinical practice, academic research, and quality improvement audits. However, users are strictly required to register their clinical or research projects through the official POS web portal (pos-pal.org) prior to implementation. The original questionnaire formatting, wording, and numerical scoring structure must not be altered without explicit formal authorization from the POS development steering committee at King’s College London.

References

• Bausewein, C., Le Grice, C., Simon, S., & Higginson, I. J. (2008). The use of the Palliative care Outcome Scale (POS) in clinical practice: A systematic review. Journal of Pain and Symptom Management, 36(6), 652–660. https://doi.org/10.1016/j.jpainsymman.2007.12.012
• Brandt, H. E., Deliens, L., van der Steen, J. T., Ooms, M. E., Ribbe, M. W., & van der Wal, G. (2005). The Palliative care Outcome Scale (POS) in the Netherlands: Cross-cultural adaptation and psychometric testing in nursing home residents with dementia. Palliatieve Zorg, 5(2), 45–54.
• Collins, E. S., Witt, J., Bausewein, C., Daveson, B. A., Murtagh, F. E., & Higginson, I. J. (2015). A systematic review of the technical qualities and diagnostic competence of the Palliative Care Outcome Scale (POS) and the Integrated Palliative Care Outcome Scale (IPOS). Palliative Medicine, 29(9), 794–804. https://doi.org/10.1177/0269216315583431
• Eisenchlas, J. H., Coventry, P. A., & Higginson, I. J. (2008). Validation of the Palliative Care Outcome Scale in an Argentinean population. Palliative Medicine, 22(8), 920–929. https://doi.org/10.1177/0269216308096906
• Hearn, J., & Higginson, I. J. (1999). Development and validation of a core outcome measure for palliative care: The Palliative Care Outcome Scale. Quality in Health Care, 8(4), 219–227. https://doi.org/10.1136/qhc.8.4.219
• Higginson, I. J., & Gao, W. (2008). Evaluating palliative care: Which outcome measures are used in practice and research? Current Opinion in Supportive and Palliative Care, 2(4), 263–268. https://doi.org/10.1097/SPC.0b013e3283186d63
• Murtagh, F. E., Ramsenthaler, C., Firth, A., Groeneveld, E. I., Lovell, N., Simon, S. T., Daveson, B. A., & Higginson, I. J. (2019). A brief, patient- and proxy-reported outcome measure in advanced illness: Validity, reliability and acceptability of the Integrated Palliative care Outcome Scale (IPOS). Palliative Medicine, 33(8), 1045–1057. https://doi.org/10.1177/0269216319854264

Items of the Scale