Introduction
Palliative care centers on enhancing the health-related quality of life (QOL) for individuals facing advanced illnesses. A crucial aspect of achieving this goal is the accurate and meaningful measurement of QOL itself. This review delves into the application of QOL questionnaires within palliative care, emphasizing their clinical and research utility. We will explore how these instruments are vital for assessing patient well-being, evaluating treatment effectiveness, and, importantly, for calculating quality-adjusted life years (QALYs). QALYs are increasingly recognized as a critical tool in healthcare, particularly in palliative care, for measuring health outcomes and informing resource allocation decisions. This article provides an updated overview of QOL measurement in palliative care, with a focus on defining QOL, examining the practical uses of QOL instruments, reviewing validation methodologies, and discussing the role of QALYs in utility analyses.
Understanding Quality of Life in Palliative Care
Quality of life in the context of health extends beyond the mere absence of disease. It is a multifaceted concept, reflecting an individual’s subjective response to life events that influence both the quality and duration of their life [1]. It represents a personal evaluation of life satisfaction and overall well-being. Significantly, individuals may live with a disease without perceiving themselves as “ill” or experiencing a diminished QOL. Conversely, some without diagnosed illnesses may report feeling unwell and have a poor QOL [2–4]. This inherent subjectivity means that patients with similar disease stages can exhibit vastly different QOL scores. Consequently, when interpreting QOL outcomes at a group level, it’s essential to acknowledge that these may not accurately reflect the diverse individual experiences of living with illness [5–7].
QOL stands as a paramount outcome measure in clinical trials, serving as a key determinant of treatment utility and cost-effectiveness. Indeed, for palliative care interventions, improving QOL is often the primary objective [8–12].
The Patient’s Perspective: What Defines QOL?
Despite the development of numerous QOL questionnaires in palliative care, a universally applicable instrument remains elusive. Patient perspectives on QOL are diverse and encompass various domains. These include: the fulfillment of personal aspirations, effective management of physical symptoms, emotional equilibrium, the capacity to maintain a normal lifestyle and self-identity, social engagement (family and societal roles), existential and spiritual fulfillment, the discovery of meaning in life, adaptability and resilience, and the adjustment of values and goals in response to disease progression [13].
The World Health Organization’s definition resonates deeply with this understanding: “an individual’s perception of their position in life in the context of the culture (family and society) and value systems in which they live and in relation to their goals, expectations, standards and concerns” [14]. Individuals reporting the lowest QOL are often those who feel least successful in achieving their personal goals, expectations, standards, and addressing their concerns. This aligns with the “Calman’s gap” concept [15], which posits QOL as the ratio or gap between expectations and reality. A narrow gap signifies good QOL, where expectations align with reality, while a wide gap indicates poor QOL. It is important to note that strategies like collusion or therapeutic misperceptions, such as presenting an unrealistically optimistic prognosis, can artificially narrow this gap, falsely inflating perceived QOL [16].
Palliative care aims to reduce this gap authentically by facilitating a “response shift” in QOL. This involves assisting patients in recalibrating their expectations and re-prioritizing domains of importance as their disease progresses. Therapies that bolster a patient’s sense of self, along with excellent symptom management and functional rehabilitation, also contribute to narrowing this gap and improving QOL.
QOL is not static; it evolves as the disease progresses [17]. Changes occur in both reflective questionnaire items and causative items that directly impact QOL scores. While fatigue, physical function, and mood often decline, pain and certain gastrointestinal symptoms may be managed effectively with progressive disease. Fixed-item QOL instruments, where each item carries equal weight, may not fully capture these shifts in importance [18]. The European Organization for Research and Treatment of Cancer QOL questionnaire (EORTC-QLQ-C30), commonly used in cancer trials, has been adapted for palliative care (EORTC QLQ-C15-PAL) to better reflect end-of-life QOL [19]. The Schedule for the Evaluation of Individual Quality of Life-Direct Weighting (SEIQOL-DW) questionnaire is designed to measure changes in QOL domains over time, accommodating recalibrations and response shifts. This tool can detect shifts in priorities and patient resilience. However, a limitation is that the SEIQOL-DW index score does not directly correlate with objective health, functional status, or demographic and clinical parameters [20]. Its strength lies in reflecting a patient’s capacity to value life domains beyond health, despite health challenges. The primary drawback is the inability to quantitatively measure the minimal clinically important difference (MCID) for therapeutic response assessment or to use it for utility estimation in cost-effectiveness analyses.
At the end of life, key domains influencing QOL and the quality of dying include physical and psychological comfort, the option to live in a preferred location, maintaining hope and pleasure, positive relationships with the medical team and family, preserving independence, being treated with respect, and a sense of life fulfillment [21]. Many of these domains are not comprehensively assessed by standard questionnaires like the EORTC-QLQ-C30 or the EuroQol five-dimension questionnaire (EQ-5D), posing a challenge to utility assessment at the end of life [8, 22–26]. Furthermore, as delirium is common in the final stages of life, caregiver ratings are sometimes used to assess patient QOL [27]. There is a clear need for utility measures specifically designed to assess the quality of dying, as existing tools like the EQ-5D are designed to gauge health status, not the dying process. Further research is essential to develop and validate QOL questionnaires tailored to different stages of the disease trajectory.
The Rationale for QOL Measurement
QOL questionnaires serve diverse purposes beyond research. Clinically, they can identify crucial domains and areas of patient distress, initiating conversations about what matters most to the individual. The Missoula-VITAS QOL Index is a validated tool well-suited for this purpose [28–30].
In palliative care intervention trials, QOL questionnaires are often primary or secondary outcomes. Ideally, these questionnaires should be validated, reliable, and, critically, sensitive and responsive to changes in QOL over time – possessing sound clinimetric properties. Establishing the MCID for the specific population being assessed is paramount [31].
QOL questionnaires also play a role in service delivery management and quality monitoring. Instruments like the EORTC-QLQ-C15 PAL, the Palliative Care Outcome Scale (POS), and the Support Team Assessment Schedule (STAS) are used to audit palliative care quality outcomes [32, 33]. As healthcare increasingly emphasizes value, QOL questionnaires will be instrumental in informing policymakers about the cost-effectiveness (utility) of interventions. Utility, often represented by quality-adjusted life years (QALYs) saved [34], presents a unique challenge in palliative care, which we will discuss further. Certain QOL questionnaires, such as the EQ-5D and the Short Form-36 (SF-36), are utility tools capable of differentiating levels of health [35–38].
Reliability, Validity, and Responsiveness: Unique Challenges in Palliative Care?
Assessing the validity of a QOL questionnaire involves several approaches. Face validity involves expert review of the questions to ensure they intuitively make sense. Content validity assesses whether the questionnaire comprehensively covers the relevant domains. Construct validity measures the questionnaire’s correlation with expected outcomes, either by comparison to another validated questionnaire or to objective measures (e.g., performance scores). This external validation can be convergent (predicting an outcome) or divergent (discriminating between distinct groups). It’s important to note that QOL may not always change with disease progression, and the validity of a questionnaire is highly context-dependent, varying by disease state. Questions suitable for one stage might become irrelevant later, leading to patient confusion and blank responses (e.g., questions about appetite when a patient relies on a feeding tube at the end of life) [39]. Therefore, understanding the validated population for a questionnaire is crucial. A significant challenge in palliative care QOL assessment is the heterogeneous patient population, encompassing diverse diagnoses, disease stages/prognoses, and languages. Validating questionnaires for specific populations is therefore particularly important. Unfortunately, palliative care research often lacks adequately powered validation studies for existing QOL questionnaires, leaving many underdeveloped.
Reliability refers to the consistency and freedom from random error of a questionnaire, encompassing internal consistency and reproducibility (test-retest and inter-rater reliability). Internal consistency measures how well items within a domain or subscale correlate with each other, typically assessed using Cronbach’s alpha or factor analysis [40]. High reliability (Cronbach’s alpha > 0.7) is beneficial for power calculations using a questionnaire’s MCID [41]. Adequate internal consistency requires a sufficient sample size (at least 100 individuals) and item categories (ideally 7 or more) [40]. Items within domains can be reflective or causative (causative items like pain directly influence QOL). Causative items can cluster, making factor analysis less suitable for internal consistency measurement. Reliability also includes repeatability, assessed when QOL is stable. Test-retest reliability generally requires at least 50 patients [40, 42–44]. However, in palliative care, QOL is often not stable. If repeatability is assessed too soon (within days), recall bias is likely. The optimal interval for determining reliability in palliative care remains undefined [40]. Inter-rater reliability becomes important with multiple assessors, measured by the intraclass correlation coefficient (Kappa), which represents the proportion of agreement beyond chance. Kappa ranges from -1 to +1, with positive values indicating agreement better than chance [45]. It is applicable to binary or nominal scale ratings.
Responsiveness measures an instrument’s sensitivity to detect changes over time that are clinically or patient-centered [46]. This requires measuring random measurement error and changes in mean scores pre- and post-intervention, or differences between groups. Sensitivity is measured using distributional and anchor-based approaches [6, 47–53], which will be discussed further. Finally, delirium, common in the final weeks of life, can impact QOL assessment. Surrogate assessments introduce further variability as patient and surrogate QOL ratings are not perfectly correlated.
Measuring Meaningful Changes in QOL: The Importance of MCID
Clinicians often find it challenging to interpret QOL scores and changes. Determining the magnitude of change that constitutes clinical significance is particularly crucial in palliative care, where QOL is a primary outcome. Regular QOL measurement, at each clinical visit or consistently over time, is essential to detect patient improvement, stability, or deterioration in response to palliative interventions. However, QOL measurement is not routinely practiced, often poorly collected in clinical trials, and frequently relegated to a secondary or tertiary outcome [54].
Responsiveness can be evaluated through repeated QOL questionnaires in observational studies, single-arm intervention studies, or randomized trials. Minimal important differences (MIDs) are estimated using statistical methods like the paired T-test, Cohen’s effect size, standardized response mean, and responsiveness statistics [31, 51, 52, 54]. The fundamental statistical principle involves dividing the pre- and post-test mean difference (assuming normality) by intra- or intergroup random error or variability. For Cohen’s standardized mean difference, it’s the difference between group means pre- and post-intervention divided by the standard deviation of baseline values. A MID effect size is often defined as a standardized mean difference (SMD) of 0.2 (though some advocate for 0.5), with moderately important differences at 0.5 and large differences at 0.8 [55, 56]. Distributional methods offer advantages: they are readily obtainable, don’t require a clinical anchor, and provide insight into data distribution. Confidence intervals can be established for precision, unlike anchor methods. However, the 0.2 SMD benchmark for MID is debated [57, 58], and the distributional approach can seem abstract to clinicians. Nonetheless, SMD can be translated into the number needed to treat (NNT), provided reliability is acceptable (Cronbach’s alpha ≥ 0.7) [59, 60]. Distributional approaches are best used alongside anchor methods for determining meaningful QOL score changes [60]. Attempts have been made to correlate important differences between these methods. One standard error of the mean or 1/2 standard deviation difference between pre- and post-test mean QOL scores correlates with an effect size of 0.5 if Cronbach’s alpha > 0.75 and 0.2 if it’s > 0.9 [59–61]. However, this relationship holds best when patients are at moderate, not extreme, QOL levels [62, 63]. MCIDs from anchor-based methods have not been directly compared to distributional approaches for most QOL questionnaires.
The MCID is defined as “the smallest difference in a score corrected for random noise within a domain of interest which patients perceive as beneficial and which would mandate, in the absence of troublesome side effects or excessive costs, a change in patient management” [53]. Unlike distributional methods, anchor methods are patient-centered. One way to determine MCID is using patient-rated global perceived improvement or decline in QOL. This anchor approach enables proportional response outcome analysis (“responder’s analysis”), which is not possible with distributional methods [7, 53]. Population-based approaches use external outcomes as anchors to define treatment responses. For instance, a 10-point improvement in a QOL questionnaire might correlate with a 20% improvement in function and survival. The MCID would then be a 10-point improvement on the questionnaire scale. Results can be expressed as absolute or relative (percentage) changes, with absolute changes being easier to analyze [64–66]. Another common anchor method uses global impressions of patient-perceived change (e.g., scales from -7 to +7), dichotomized to construct receiver-operating characteristic curves for cutoff values, determining sensitivity and specificity [5, 53, 67–70].
Anchor methods have limitations. Meaningful changes are not always bidirectional, are anchor-dependent, and are not immutable [52, 71–73]. This lack of bidirectional symmetry is seen in tools like the SF-36 and the Functional Assessment of Cancer Therapy [52, 74]. A universal MCID for all populations is unlikely [61, 75, 76]; however, due to limited research, available MCIDs are often liberally applied beyond their original populations. The chosen threshold for cutoff in Likert scales (e.g., -7 to +7) in global QOL assessments influences the MCID. Anchors may not account for precision, potentially placing the MCID within measurement error [63, 77]. Discrepancies can also arise between population-based anchors and individual perceived global changes and disease course. For example, SF-36 QOL changes more closely mirror patient global anchors and physician assessments than actual disease course in rheumatoid arthritis [78]. Global anchors are susceptible to response shifts as patients adapt to their illness and to regression toward the mean, where patients with the worst QOL may show improvement over time without intervention [77, 79].
Due to the limitations of both anchor-based and distribution-based approaches, using multiple methods to determine MCID in QOL questionnaires is recommended over relying on a single approach [52, 71, 77].
Palliative care QOL questionnaires are generally psychometrically tested, but many lack responsiveness (clinimetrics) [40, 80]. Clinicians often rely on statistically significant differences for reporting outcomes, misinterpreting statistical significance as clinical importance. Sound psychometrics are often mistakenly seen as indicators of clinical utility. Validity and reliability are initial steps in questionnaire development; clinical usefulness hinges on established responsiveness. Palliative care specialists must understand QOL change interpretation and the clinimetrics of these measures, especially in value-based healthcare systems where utility will be crucial for payers and policymakers.
Utility and its Relevance to Palliative Care: Quality-Adjusted Life Years (QALYs)
When QOL is used in palliative care, the endpoint should be both clinically and economically relevant. QOL questionnaires are adapted to reflect health states or levels that have utility in economic evaluations of palliative care. Palliative interventions should improve QOL, even without extending survival, thus positively impacting quality-adjusted life years (QALYs). The utility of an intervention is determined by its cost relative to the change in health state. Interventions with minimal benefit but high cost will not be considered valuable in population-based healthcare. Palliative care specialists face the challenge of demonstrating value to policymakers and payers by incorporating utility or cost-effectiveness data [81–84]. This is critical for the future of palliative care in value-based healthcare systems, where demonstrating value is essential for sustainability.
In value-based healthcare, the difference in QALYs before and after an intervention is used to estimate its impact on both quality and quantity of life. QALYs are calculated by multiplying the duration of time (in years) spent in a specific health state by the utility score associated with that state. Health state utility is derived from QOL questionnaire scores, adjusted for the relative importance of each QOL domain based on patient preferences. Some QOL questionnaires, like the EQ-5D and SF-36 (transformed into SF-6D), already have established weights based on population studies, making them particularly useful for utility analyses. The QALY difference with and without an intervention typically ranges from -1 to 1, where 0 represents a health state as bad as death for one year, 1 signifies perfect health, and negative scores indicate states worse than death. Cost-effectiveness is then assessed by dividing the intervention cost by the QALY difference, informing decisions about healthcare resource allocation. Utility assessment, in this sense, is patient-centered because it reflects patient-perceived QOL and benefits.
The SF-36, when transformed into the SF-6D utility tool, covers 6 QOL domains, each with 2-6 health levels: physical function, role limitation, social function, pain, mental health, and vitality. Health state values range from 0.35 to 1.0 [85]. The SF-6D MID is 0.03 (mean), with an effect size of 0.051 [86, 87].
The Assessment of Quality of Life (AQoL) instrument has also been adapted into a utility questionnaire, comprising 5 dimensions (illness, independent living, social relationships, physical function, psychological well-being). Four dimensions are combined in a multiplicative model weighted for community values, with health state values ranging from 0.04 to 1.0 [88].
The EQ-5D is the most widely used utility tool internationally, available in 3-level (EQ-5D-3L) and 5-level (EQ-5D-5L) Likert question forms. The EQ-5D-5L exhibits less of a ceiling effect. Scores range from -0.59 to 1, with negative values indicating states worse than death and 1 representing perfect health. Preference values vary across populations [89]. The MID values for SF-6D and EQ-5D are not equivalent and differ in absolute values. Both categorize health status similarly (very good to very bad), but disagreements exist, especially at the lower end of the health spectrum [90, 91].
QOL questionnaires with domains similar to EQ-5D or SF-36 can be mapped onto these utility measures in clinical studies [92–94]. Mapping allows QOL questionnaires to be used for both clinical outcomes and cost-effectiveness analyses, provided domain congruence exists. Mapped QOL questionnaires also require established MCIDs. Accurate mapping requires multiple predictors and specific characteristics: studies including both utility measures (e.g., EQ-5D) and QOL questionnaires, similar patient characteristics across studies using both, sufficient patient numbers completing both, well-defined clinical effectiveness or health outcomes, and outcomes relevant to the population [25, 95, 96].
Mapping QOL questionnaires to utility tools has drawbacks. Mapping regression is population-dependent, and regression algorithms vary, meaning “one size does not fit all” [96]. Certain questionnaires, like the Integrated Palliative care Outcome Scale (IPOS) and IPOS-Symptoms (IPOS-S), cannot be mapped to EQ-5D, precluding their use for utility assessment if EQ-5D is the chosen tool [97]. More broadly, patient and healthcare system perspectives on utility differ. Patients are concerned with the personal benefit and cost of palliative interventions, while utility measures primarily focus on medical system economics, often overlooking family financial burdens. QALYs assume a linear value of life duration, but patients may value time differently near the end of life, a nuance not captured by fixed utility questionnaires. Metrics of the quality of dying are also missed by EQ-5D and SF-6D. Utility tools cannot dissect the most valuable components of complex palliative care interventions [34]. Palliative care urgently needs utility measures tailored to its unique aspects.
Evidence that Palliative Care Enhances QOL and QALYs
A meta-analysis of recent studies examining specialist palliative care services in hospitals, hospices, or community settings found that palliative care had a small but statistically significant positive effect on QOL in 3 of 8 studies, with non-significant differences in the remaining four [98]. The overall effect size (SMD) was 0.16 (95% confidence interval 0.01–0.31). Sensitivity analysis yielded an SMD of 0.57 (95% confidence interval -0.02 to 1.15). For patients receiving early specialist palliative care, the effect size was 0.33 (95% confidence interval 0.05–0.61). Another study analyzing 10 palliative care studies (3 RCTs) of moderate to low quality, mostly in cancer patients, estimated an effect size of 0.27 in half of the studies. Observational studies showed larger effect sizes than RCTs [99]. A third study measuring end-of-life QOL using the Good Death Inventory [100] found improvements in 3 of 10 domains (preferred place of death, maintained hope and pleasure, comfortable environment), with effect sizes of 0.1, 0.1, and 0.09, respectively. Benefits were greater for patients with poorer performance scores (ECOG 3 and 4), with an effect size of 0.54.
These population-based effect sizes for palliative care’s impact on QOL might seem modest, sometimes below the previously discussed MID. However, population SMDs represent averages across individuals with dramatic improvements, modest gains, stable QOL, and those who worsen. Certain groups, like those receiving early palliative care and those with poorer performance scores at end-of-life, appear to benefit more. The challenge in assessing QOL benefits lies in the lack of data on individual responses achieving MCID, which would allow for gauging proportional or percentage responses [101]. This is a significant limitation in interpreting current studies.
QOL Tools for Consideration in Palliative Care
A recent systematic review of QOL questionnaires in palliative care recommended three measures with adequate psychometrics [40]: the McGill QOL Questionnaire, the QOL at the End of Life questionnaire, and the Quality of Death and Dying questionnaire [102–104]. Responsiveness was identified as the weakest psychometric property for these questionnaires. A survey of European palliative care specialists found the European Organization for Research and Treatment of Cancer QOL questionnaire (EORTC-QLQ-C30) and the Integrated Patient Care Outcome Scale (I-POS) to be the most frequently used [42].
McGill QOL Questionnaire
The McGill QOL Questionnaire consists of 16 items, each with 11 response options (0–10), covering psychological symptoms, existential well-being, support, and physical symptoms. Factor analysis validated these domains. Completion time is 10–30 minutes, with a 2-day recall period. Subscale scores are available. It has been validated in inpatient palliative care, demonstrating reliability (Cronbach’s alpha > 0.7, except for physical symptoms at 0.62) and intra-class correlation coefficients of 0.62–0.85. Validation against the Spitzer QOL Index and external validation against pain intensity showed correlations, with pain intensity correlating with a 0.56 change in the existential subscale and 0.66 in the total scale. Effect size differences were significant between good and average days and average and bad days for physical symptoms and support domains [104–107].
A shortened version of the McGill QOL questionnaire has been validated in 190 terminally ill patients, with a completion time of 3.3 minutes. Cronbach’s alpha ranged from 0.46 to 0.86 compared to the long form, with strong correlations across domains. External validation against hemoglobin levels and construct validity supported by principal component analysis were also demonstrated [108].
QOL at the End of Life Questionnaire
This questionnaire has 25 items with 5 response options each, covering domains like outlook, inspiration, spiritual activity, religion, and community. It uses a 1-week and 1-month recall period and requires a structured interview. Total and sub-section scoring systems are available. It has been validated in diverse populations, including heart failure, COPD, cancer, and end-stage renal disease patients, which is a strength. Internal consistency was > 0.7, correlating with the Functional Assessment of Chronic Illness-Spiritual subscale (> 0.6) and preparation subscale (0.4–0.6), and moderately correlating with the Missoula-VITAS QOL Index [103].
Quality of Dying and Death Questionnaire
This questionnaire assesses the quality of death from the perspective of bereaved caregivers or healthcare professionals. It contains 31 items with 11 response options (0–10), covering symptoms, personal care, preparation for death, moment of death, family, treatment preferences, and whole-person concerns. It is administered via semi-structured proxy interviews with retrospective recall. Cronbach’s alpha for internal consistency is 0.89. Factor analysis did not support subscale constructs based on questionnaire domains. Studies in dying patients showed higher scores associated with death at home, preferred location of death, better symptom treatment, adherence to patient preferences, family satisfaction with communication and care, and healthcare team availability [102].
Palliative Care Outcome Scale (POS)
This scale was developed through agreement between staff and patient ratings for 8 of 10 items and validated in 148 palliative care patients. Both staff and patient versions are available. The original scale has been expanded to include symptoms. Reliability is acceptable, with Cronbach’s alpha of 0.65 for patient and 0.74 for staff versions. Construct validity shows Spearman correlations from 0.48 to 0.8 with EORTC-QLQ-C30 and STAS questionnaires. Test-retest reliability is acceptable for 7 of 10 items. Scale score changes were noted with disease progression, but were not statistically significant [109, 110]. Validation in cancer patients using cognitive debriefing of patients and oncologists identified distinct domains: emotional well-being, disease consequences, informational support, anxiety, and burden of disease, although this is debated. Another study suggested two factors: psychological well-being and professional care-related items, with three independent items (family anxiety, symptoms, and pain). Some authors suggest POS reflects quality of care more than QOL [111]. POS has been applied to patients with HIV, neurological, pulmonary, cardiac, and kidney diseases, and used to study symptom prevalence and audit care quality [32, 112].
European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC-QLQ-C30)
The EORTC-QLQ-C30 has 9 multi-item domains, including 5 functional scales (physical, role, cognitive, emotional, social) and 3 symptom scales (fatigue, pain, nausea and vomiting) [113]. It includes global health and QOL scales with a 1-week timeframe and single-item symptom scales.
Initially validated in 305 patients across 13 countries, it takes approximately 11 minutes to complete. Only the functional scale (work and household work) failed to meet reliability standards (Cronbach’s alpha ≥ 0.7). Intrascale correlations were significant, suggesting distinct QOL domains. Discriminative validity was demonstrated based on ECOG performance scale. Statistically significant changes in expected directions were observed for physical and role function, global QOL, fatigue, nausea, and vomiting in patients with performance score improvements or declines. Psychometrics were consistent across countries and validated in terminally ill patients [114].
Validation against the Hospital Anxiety and Depression Scale, Edmonton Symptom Assessment Scale, and WHO-QOL Brief Scale showed Cronbach’s alpha > 0.7, except for social function (0.69), cognitive scale (0.57), and nausea and vomiting (0.69). This study confirmed its ability to discriminate patient groups by performance score, treatment, and education [115].
A review of 30 studies using EORTC-QLQ-C30 found it improved prognostication by 5–8% over clinical variables alone in various cancers, with nausea and vomiting, anorexia, dyspnea, role function, and physical function contributing to prognostication. Causative item clusters varied by disease. MCIDs have been established but vary by patient population [116].
The EORTC-QLQ-C30 has been adapted for palliative care as EORTC-QLQ-C15-PAL, based on interviews with 44 patients and 66 healthcare professionals. Item response theory was used to shorten the scale, prioritizing pain, physical function, emotional function, fatigue, global health status, QOL, nausea and vomiting, appetite, dyspnea, constipation, and sleep. Four original subscales were shortened to reduce burden [117]. Validation in 104 patients showed reliability ≥ 0.7, except for fatigue (0.58). It demonstrated convergent validity against the original scale, Brief Pain Inventory, and Beck Depression Inventory, and discriminated performance status [118]. Importantly, MCIDs are established for this scale in cancer patients [73]. This questionnaire is suitable for mapping to EQ-5D to assess palliative care intervention cost-effectiveness.
EQ-5D
The EQ-5D is a non-disease-specific measure describing health states and health-related QOL, complementing rather than replacing QOL measures [119]. It consists of a general health description across 5 domains, resulting in 243 health states plus “unconscious” and “dead.” Utility can be assessed using country-specific algorithms. Domains include mobility, self-care, usual activities, pain/discomfort, and anxiety/depression [120, 121]. As mentioned, it comes in 3-level and 5-level response formats, with the 5-level version less prone to ceiling effects [122]. It also includes a 20 cm visual analogue scale (VAS) anchored by “best imaginable” and “worst imaginable” health states (0–100). Relevant for both healthy and seriously ill individuals, it generates a single index value, with health states potentially rated as “worse than death” (negative values). Utility was established through individual values and “time trade-off” methods, leading to negative values. QALYs are generated using time trade-off methods and the VAS scale. It discriminates functional classes, correlating with the Health Assessment Questionnaire score [122]. EQ-5D’s purpose is to monitor patient health status over time, assess condition/disease severity, provide evidence for medical effectiveness in improving health states, and measure cost-effectiveness per QALY. It has been used to establish local, regional, and national health states [119].
Expert Commentary
Given that improving QOL is a central goal of palliative care, routine QOL measurement is essential. Numerous QOL questionnaires exist for palliative care, many with acceptable reliability and validity. However, a critical gap, particularly in palliative care, is responsiveness to change. MCIDs are largely undetermined, hindering interventional studies. Utility and cost-effectiveness analyses using QOL questionnaires and utility tools are also limited. QOL questionnaires are valuable for patient-clinician communication, outcome measurement, and quality audits, but policymakers and payers prioritize utility and cost-effectiveness. Developing utility measures tailored for palliative care outcomes is a future imperative.
Five-Year View
Currently, most palliative care QOL questionnaires have a fixed number of items. Collaborative research efforts are needed to: 1) further validate existing QOL questionnaires across disease trajectories; 2) define MCIDs for power calculations in clinical trials and result interpretation; and 3) assess clinical utility for cost-effectiveness analyses of palliative care interventions. Simultaneously, research is underway to develop personalized QOL questionnaires using computerized adaptive testing. This approach may reduce item burden and enable immediate electronic health record recording and reporting. A recent study demonstrated that routine patient-reported outcome assessment positively impacts not only QOL but also survival [123]. Further research is needed to fully explore the benefits of such interventions.
Key Issues
- The World Health Organization defines quality of life as “an individual’s perception of their position in life in the context of culture and value system in which they live and in relation to their goals, expectations, standards and concerns.”
- QOL is dynamic and changes with disease progression. Individual QOL perception can also evolve over time (response shift).
- Palliative care plays a vital role in enhancing QOL by improving patients’ ability to achieve personal goals, manage symptoms, and gain social support, while fostering realistic expectations through effective communication.
- Systematic reviews and RCTs demonstrate that palliative care referral and interventions improve QOL for patients with advanced diseases.
- QOL questionnaires are used for clinical needs assessment, quality monitoring, clinical trial outcomes, and utility/cost-effectiveness estimation.
- Validity and reliability are established for many palliative care QOL questionnaires. Validity assesses accuracy; reliability assesses reproducibility.
- Responsiveness to change and MCID are largely undetermined for most QOL questionnaires. Responsiveness reflects the ability to detect change over time; MCID defines the clinically meaningful change magnitude.
- QOL questionnaires can facilitate utility analyses to assess cost-effectiveness. This requires weighting QOL scores and transforming them into health states. More palliative care research is needed on utility and cost-effectiveness.
Acknowledgments
Funding
D Hui is supported in part by National Institutes of Health Grants (1R01CA214960–01A1, R21NR016736), an American Cancer Society Mentored Research Scholar Grant in Applied and Clinical Research (MRSG- 14–1418-01-CCE) and the Andrew Sabin Family Fellowship Award.
Footnotes
Declaration of interest
The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
Contributor Information
Mellar P Davis, Geisinger Medical Center, Danville, PA 17822.
David Hui, Department of Palliative Care, Rehabilitation and Integrative Medicine, MD Anderson Cancer Center, Houston, TX, USA, 77030.
References
(References are identical to the original article and are not repeated here for brevity but would be included in a full article.)
