| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the Department of Anesthesia, Indiana University School of Medicine, Indianapolis, Indiana.
Address correspondence to Thomas R. Vetter, MD, Department of Anesthesia, Riley Hospital for Children, Room 2001, 702 Barnhill Drive, Indianapolis, Indiana 46202. Address e-mail to tvetter{at}iupui.edu.
Abstract
BACKGROUND: Pain is a complex and individual experience that is often difficult for patients to fully describe using a conventional pain intensity scale. Health-related quality of life is an additional metric by which to assess patients’ subjective perspective on their chronic pain experience and its adverse effect on their lives. Health-related quality of life encompasses those aspects of health and well-being valued by patients, specifically, their physical, emotional, and cognitive function, and their ability to participate in meaningful activities within their family, workplace, and community.
METHODS: A methodical search of the medical literature was undertaken to identify the most commonly applied health-related quality of life measurement instruments. These measurement instruments were then assessed within the context of chronic pain medicine clinical practice and research.
RESULTS: This primer provides an overview of the concept of health-related quality of life as a clinical measurement and the specific means by which to measure health-related quality of life across various cultures in adults, as well as in children and adolescents, suffering from chronic pain conditions.
CONCLUSIONS: We have the ability and impetus to routinely assess adult and pediatric health-related quality of life in chronic pain medicine. However, further attention needs to be focused on overcoming barriers to the more widespread measurement of health-related quality of life. A valid preference-based, utility measure of health-related quality of life is a requirement for performing a cost-utility (cost-effectiveness) analysis and undertaking formal decision analysis modeling.
Pain is a complex and individual experience that is often difficult for patients to fully describe using a conventional unidimensional self-reported pain intensity scale (1). The measurement of health-related quality of life (HRQoL) is another way to assess patients’ subjective perspective on their pain experience and its adverse impact on their lives. The concept of HRQoL can also be incorporated into the pain treatment options presented to patients. For example, HRQoL and resulting patient preference can play an important role in clinical decision making in the treatment of chronic low-back pain, specifically in deciding between surgical management and nonoperative symptom control (2).
The measurement of HRQoL, and the effect of pain on HRQoL, in particular, is in keeping with the growing belief that a patient-centered social contract should be developed between the patient and the treating physician. This social contract affords patients the right to determine how they want to live as they survive rather than what they are willing to lose in order to extend survival (3). This greater involvement of patients in determining their chronic plan of care is based largely upon patients’ self-perceived HRQoL (4,5).
This primer is intended to provide an overview of HRQoL within the context of chronic pain medicine research and clinical care. The paper will first discuss the fundamentals of HRQoL. It will then briefly discuss the tools by which to measure HRQoL across various cultures in adults as well as in children and adolescents (6–8).
HRQOL AND THE ADVERSE EFFECTS OF PAIN
The World Health Organization (WHO) has defined health to be "a state of compete physical, mental, and social well-being, and not merely the absence of disease or infirmity" (9). Patients seek out medical care because their illness or disease has detrimentally affected not only their health and well-being but also their attendant quality of life (10). Chronic pain is often associated with a reduced sense of well-being. An extensive cross-sectional survey of Finnish adults (11) revealed a 14.3% prevalence of daily chronic pain, with the relative frequency of chronic pain being an independent predictor of self-rated poor health.
Health and quality of life are inherently interrelated, thus giving rise to the concept of HRQoL (Fig. 1) (12). HRQoL is the assessment of quality of life within the context of clinical medicine and clinical research (13). In contrast to the more objective physiologic, laboratory, and radiologic assessments of disease, the measurement of HRQoL provides additional important information about the subjective adverse impact of a chronic condition, especially one resulting in pain and suffering (14–16). HRQoL encompasses physical, emotional, and cognitive function, as well as the ability to participate in meaningful activities within the family, workplace, and community (17). HRQoL also inherently carries with it the notion of value, with one state of health being perceived as superior to another (10).
|
Transforming the concept of HRQoL into a practical measurement tool requires identifying the various elements, formally referred to as the "domains" or "dimensions" of an individual’s existence, that constitute HRQoL (Table 1) (18). In an effort to standardize clinical outcomes measurement, the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) identified six core outcome domains to be considered when designing chronic pain treatment efficacy and effectiveness trials (19). It is worth noting that two of the six IMMPACT core outcome domains, namely, physical functioning and emotional functioning, directly overlap with the conventional domains of HRQoL.
|
THE SELECTION OF HRQOL MEASUREMENT INSTRUMENTS
An objective selection process was applied to determine which specific adult and pediatric HRQoL measurement instruments to include in this primer (Fig. 2). An initial PubMed search was undertaken in August 2006 of the MEDLINE database from 2000 to the present, using the MeSH-terms "Health Status Indicators" and "Quality of Life." This initial search was limited to English language review articles but included all ages. The published abstracts of the 544 citations generated via this search were individually read by the author, from which 33 review articles from a variety of disciplines (see Appendix) were identified as focusing on specific generic and preference-based HRQoL measurement instruments. A full-text reading of these 33 review articles generated 21 adult generic and preference-based HRQoL instruments and 28 pediatric generic HRQoL instruments.
|
This compilation was supplemented by six other adult generic HRQoL measurement instruments described in the "General Health Status and Quality of Life" chapter of the most recent (2006) edition of McDowell’s text, "Measuring Health: A Guide to Rating Scales and Questionnaires" (20). Finally, an additional three adult and two pediatric generic HRQoL instruments were identified from the online Patient-Reported Outcome and Quality of Life Instruments Database (www.qolid.com), maintained by the MAPI Research Institute and Trust (www.mapi-research.fr).
In an effort to determine how commonly each of these 30 adult generic HRQoL instruments and 30 pediatric generic HRQoL instruments has been applied, a subsequent series of individual PubMed searches were undertaken of the MEDLINE database from 1966 to the present. The results of this series of MEDLINE searches (Table 2) provide an empirical estimate of the historic use of the adult and pediatric HRQoL instruments throughout the international medical community. These data on the use of specific adult HRQoL measurement instruments are consistent with a previous bibliographic study of applied patient-assessed health outcome measures (42).
|
THE THREE BASIC CATEGORIES OF HRQOL INSTRUMENTS
HRQoL instruments fall into three basic categories: 1) generic measures, 2) condition-specific measures, and 3) preference-based measures. Each category of HRQoL instrument serves a specific purpose in clinical practice and clinical research (Fig. 3) (43).
|
Generic and condition-specific health measures are constructed using traditional psychometric testing principles, such as reliability, validity, and responsiveness (44,45), whereas preference-based measures are rooted in utility theory (46).
In the last 30 yr, there has been a proliferation of HRQoL measurement instruments (47,48). These numerous HRQoL instruments have varied widely in their development, content, applicability, and quality. In an attempt to address this lack of standards, the Scientific Advisory Committee of the Medical Outcomes Trust has devised a set of eight key attributes and corresponding review criteria for HRQoL measurement instruments (Table 3) (49).
|
Generic HRQOL Instruments
Generic HRQoL instruments provide an overall, comprehensive perspective, and thus have the ability to assess the complex continuum between well-being, disability, and death (48,50). Generic measures emphasize breadth over specificity by focusing on the common elements of health that transcend any one disease (51). This breadth allows generic measures to be used both in and across a variety of medical and surgical conditions and related treatments.
Generic measures of HRQoL are applicable in both clinical care and clinical research (10). Patient-reported generic HRQoL scores can augment the clinician’s duly noted signs and symptoms of disease, and other conventional diagnostic test data (52). A generic HRQoL measure can likewise strengthen a clinical trial by providing outcomes data on the beneficial and adverse effects of treatment from the study subjects’ perspective (53).
Two caveats with generic measures of HRQoL are the potential for a ceiling effect versus a floor effect and the relative lack of sensitivity to change (i.e., responsiveness) in any one given health domain or dimension (54). A ceiling effect is present when a significant proportion of subjects initially rate themselves as being in good health, and thus subsequent improvement is not readily discernable. Conversely, a generic HRQoL instrument may not be able to discern either poor health from very poor health or further deterioration, resulting in a floor effect (13,55). By design, a generic instrument assesses overall HRQoL, and thus may not detect a clinically significant change in a specific body system or symptom; hence, the impetus for a concomitant condition-specific instrument (56).
Moreover, even if pain is assessed as a separate dimension on a generic HRQoL instrument, the relative intensity of a patient’s chronic pain can have a nonlinear adverse effect on other assessed health dimensions, such as physical functioning, sleep, and mood (57,58). In patients with osteoarthritis, low back pain, and peripheral neuropathy, the qualities and spatial characteristics of pain also appear to have independent effects on patient-perceived levels of physical and emotional functioning (59).
Five of the most commonly applied and widely cited adult generic HRQoL instruments (Table 4) are discussed here (60,61).
|
The 36-Item Short-Form Health Survey
The 36-Item Short-Form Health Survey (SF-36) (http://www.sf-36.org/) (21,62,63) is the most commonly used generic HRQoL measure in the world (61,64). This phenomenon is due in large part to the widespread multicultural testing and adaptation of the original American English SF-36 by the International Quality of Life Assessment Project (http://www.iqola.org/) (60,65,66). Not surprisingly, the IMMPACT has recommended that the SF-36 be incorporated "as a generic measure of physical functioning because of the large amount of data available to permit comparisons among different disorders and treatments" (67).
Despite its preeminent stature, the SF-36 is not without potential limitations, particularly in assessing the effect of pain on HRQoL. Even though the SF-36 may be applied to evaluate the overall outcome of a multidisciplinary pain treatment regimen, only 2 of the 36 items (questions) on the SF-36 specifically assess bodily pain (Table 5) (21). It is thus quite possible that the therapeutic effect of a given analgesic regimen (or lack thereof) will be overshadowed by the other seven SF-36 health-related domains or dimensions (54).
|
Furthermore, in a study of health status measures in low back pain patients, the SF-36 appeared to have a floor effect in those patients who deteriorated between baseline and follow-up at 3 and 6 mo. In addition, the SF-36 physical functioning subscale demonstrated poor responsiveness to both improvement and deterioration when compared with the Health Utility Index (HUI), the EuroQol (EQ-5D), and the Oswestry Low Back Pain Disability Questionnaire (69).
In an effort to address these shortcomings, a low back-specific version of the SF-36 physical functioning scale has been created and applied (70). In a cohort of patients with low back and associated lower extremity pain of variable duration, the Low-Back SF-36 PF18 exhibited improved responsiveness to clinical change, and floor and ceiling effects were essentially eliminated (70).
The Dartmouth Primary Care Cooperative Information Charts
The Dartmouth Primary Care Cooperative Information (COOP) charts were originally developed by a network of community medical practices involved in primary care research projects (http://www.dartmouth.edu/~coopproj/) (28). Designed to be analogous to the familiar Snellen visual acuity screening chart, the COOP Charts are a series of nine pictorial, user-friendly subscales (71). One of the nine subscales on the COOP Charts directly assesses pain (Table 5).
Although more extensively applied in a primary care setting, the COOP Charts appear to be a viable option to assess HRQoL in a high-volume, predominantly interventional chronic pain medicine practice. To this end, the COOP Charts have been shown to be a suitable alternative to the SF-36 as a patient-reported outcome measure in the longitudinal treatment of chronic low back pain (72).
The Nottingham Health Profile
The Nottingham Health Profile (NHP) was developed specifically for use in community-based ecologic studies of chronic disease and attendant unmet health care needs (73). Pain is assessed on the NHP by way of eight simple yes/no questions, which focus on the persistency of pain and the changes in pain with various physical activities (Table 5).
The major problem with the NHP is that it was not designed to be used in clinical trials (23). The NHP has also exhibited a floor effect in conditions with severe symptoms, undermining its applicability in many chronic pain conditions (11). The NHP is thus generally considered a historically innovative, but now seldom-used, generic health survey instrument (61).
The Sickness Impact Profile
The Sickness Impact Profile (SIP) (http://www.outcomes-trust.org/instruments.htm) was one of the earliest (1975) self-reported general health surveys. The SIP was designed primarily to assess the effect or burden of illness on an individual’s activities of daily living and behavior (25). Only two of the 136 items on the SIP directly assess pain (Table 5). Although the SIP remains applicable in medical conditions involving severe mobility impairments, including those accompanied by chronic pain (e.g., osteoarthritis), (74) the instrument has generally been supplanted by other, more respondent-friendly tools (61).
The WHO Quality of Life Surveys
The WHO quality of life (WHOQOL) project has from its inception in 1991 involved 15 collaborative centers worldwide (27). The resulting WHOQOL with 100 Questions (WHOQOL-100) and its abbreviated version, the WHOQOL-BREF (http://www.who.int/evidence/assessment-instruments/qol/), assess individuals’ perceptions of their quality of life within the context of their culture and its value systems in addition to their own personal goals, standards, and concerns (75–77).
Although the WHOQOL-100 is a very comprehensive instrument (78,79), its length makes it impractical for repeated use in clinical trials or routine clinical practice (61). The WHOQOL-BREF was therefore created from the pool of items on the WHOQOL-100 (80). Both the WHOQOL-100 and the WHQOL-BREF include pain and discomfort as one facet of the physical health domain (Table 5). The WHQOL-BREF is considered a viable and valid alternative to the SF-36 as a cross-cultural, patient-centered HRQoL measurement instrument (81).
Condition-Specific HRQOL Instruments
Condition-specific HRQoL instruments are designed to assess specific diagnostic groups, usually with the goal of determining clinically significant responsiveness to treatment or disease progression (82). Condition-specific measures are formulated to identify small incremental changes in the most relevant domains or dimensions of a particular disease (64). This sensitivity makes condition-specific measures particularly attractive to clinicians and health outcomes researchers seeking to identify tangible interventional benefits (54). While a generic health status measure like the SF-36 includes overall pain as one domain of HRQoL, a properly constructed condition-specific measure can assess in greater detail the impact of a chronic pain condition on HRQoL.
However, this specificity has at least two potential pitfalls. A condition-specific measure may fail to discern the impact of the disease or illness on a patient’s general function and well-being (61). For example, a medication may reduce the symptoms of postherpetic neuralgia but not improve the ability to perform activities of daily living because of the general frailty of the patient. Second, given the vast array of applied condition-specific measures, it is often difficult or impossible to compare the treatment effects from studies of a different or a single disease. This has lead to the overarching premise that generic health measures provide a common metric for comparisons across different treatments or conditions, and condition-specific measures can serve to complement, rather than replace, generic measures (83,84).
Chronic Pain Condition-Specific HRQOL Measures
While a discussion of their individual properties and frequency of use is beyond the intended scope of this primer, there are a growing number of chronic pain condition-specific HRQoL and patient-reported outcomes measures. Two such chronic pain condition-specific instruments are the Roland–Morris Back Pain Questionnaire (RMQ) and the Western Ontario and McMaster University Osteoarthritis Index (WOMAC). These two instruments were chosen here simply to serve as illustrative examples for common chronic-pain conditions.
The RMQ was created by culling 24 pain and function items from the SIP considered to be relevant to chronic low back pain (85,86). The 24 questions on the self-administered RMQ are answered with a simple yes or no, generating a total score from 0 (least disability) to 24 (greatest disability). The RMQ can be completed in 5 min or less and can be readily scored by the clinician, thus posing minimal respondent and administration burden (87). The RMQ is available in 12 languages. The reliability, validity, and responsiveness of the RMQ have been well established in its targeted low back pain population (88,89).
The WOMAC (http://www.womac.org) is a multidimensional, self-administered health status instrument for patients with osteoarthritis of the hip or knee (90). The WOMAC assesses the three dimensions of pain, disability, and joint stiffness using a series of 24 questions with either a five-point Likert or a 100 mm visual analog scale format. The WOMAC is self-administered and can be completed in an average of 12 min (91). The latest version of the instrument (WOMAC 3.1) is available in 65 languages and dialects. The WOMAC has been shown to be one of the most significant preoperative predictors of functional outcome after total knee arthroplasty (92). The WOMAC has also been reported to be the instrument of choice for evaluating patient-centered outcome after knee replacement surgery in osteoarthritis (93).
The Valuation of Health States Using Preference-Based Instruments
Preference-based instruments are fundamentally different from generic and condition-specific HRQoL instruments that simply describe health states (94). Preference-based measures of HRQoL expand upon generic and condition-specific measures by applying various methods to incorporate patient opinion concerning the utility or value of a particular health state (46).
The utility or value of a given state of health is determined by its relative desirability or preference as perceived by the individual at a particular point in time (95). Preference-based health measures are predicated on this perceived utility or value and the resulting rater’s preference for one state of health over another (96). In addition, and very importantly, only preference-based HRQoL scores can be used in performing a cost-utility (cost-effectiveness) analysis and undertaking formal decision analysis modeling (Fig. 3) (97,98).
Generating a utility score for a specific health state, such as suffering from diabetic neuropathic pain or chronic sciatica, involves applying either a direct or an indirect preference-based measurement method. A direct preference-based measure typically entails direct face-to-face respondent interviews. In contrast, an indirect preference-based measure involves the completion of a conventional health status questionnaire by study subjects or patients. The questionnaire responses in turn generate a valuation for the specific health state using a pre-established formula (99).
The Direct Preference-Based Methods for Generating a Health-Related Utility
The standard gamble, the time trade-off, and the visual analog rating scale are the three most common direct preference-based methods for assigning a utility or value score to a specific health state (99–103). The standard gamble method has been widely applied in the field of decision analysis and represents the original method for obtaining a health utility score (104,105). However, limited patient facility with probabilities and numbers can be a significant barrier to the meaningful routine assessment of health utility using either the standard gamble or time trade-off technique (106). The three direct preference-based health utility methods are thus usually supplanted by an indirect preference-based measure of HRQoL.
The Indirect Preference-Based Measures of HRQOL
An indirect preference-based measure of HRQoL consists of a simplified descriptive health survey and a predetermined preference-based scoring formula (107). In contrast to a multi-item generic HRQoL instrument, the simplified scale for each health domain or dimension (including pain) on an indirect preference-based HRQoL instrument often consists of only a single item or question (13,108). The predetermined preference-based scoring formula is generated during the initial instrument development using one of the above direct utility methods (e.g., the standard gamble) in a random sample of the general population (107).
An indirect preference-based instrument generates a single, aggregate health utility score, which conventionally ranges along a continuum between 0.0 (equal to death) and 1.0 (equal to perfect health). This single preference-based utility score compares the patient’s present health state to death and represents the relative value of that health state to the patient (15).
The four most widely applied indirect preference-based measures of HRQoL are included in this primer (Table 6). All four of these indirect preference-based generic measures were developed using multiattribute utility theory (110,111). Multiattribute utility theory holds that the individual health attributes or dimensions (including pain) that comprise HRQoL can be either 1) added together if they are assumed to be independent of one another, or 2) multiplied together if they are assumed to be interactive and thus interdependent, to arrive at a final, combined health utility score (13,108,112,113). In practical terms, if the instrument scoring formula is multiplicative and pain is an included health domain, then pain has a direct effect across all of the other health domains.
|
The Quality of Well-Being Scale.
The Quality of Well-Being Scale (QWB) scoring function is additive and is based upon visual analog scale preference scores obtained from a random sample of the general population in San Diego, CA (http://www.outcomes-trust.org/instruments.htm) (99,114). The oldest of the four measures, the QWB was originally developed within the framework of a general health policy model focusing on health care resource allocation (30,115). Pain on the QWB is not assessed as a separate dimension, but instead is grouped in a logical manner with other commonly associated symptoms in various regions of the body (Table 5).
The HUI3.
The HUI3 (http://www.healthutilities.com/) scoring function is multiplicative and is based upon standard gamble preference scores obtained from a random sample of Canadian adults (99,113). The HUI3 is essentially an updated and expanded version of the HUI2. Both the HUI2 and HUI3 assess pain as a separate HRQoL dimension (Table 5) (26,116). The HUI3 scoring range (–0.36 to 1.0) includes negative health utility scores for health states perceived as being worse than death (61,117).
The EQ-5D.
The EQ-5D (http://www.euroqol.org/) scoring function is additive and is based upon time trade-off preferences scores obtained from a random sample of adults in the United Kingdom (22,99, 118–120). The EQ-5D incorporates pain/discomfort as a discrete HRQoL dimension (Table 5). Like the HUI3, the EQ-5D health utility scoring range (–0.59 to 0.0) includes negative values for health states considered by the respondent to be worse than death (61,117,118). The EQ-5D may be more likely than the HUI2 or HUI3 to exhibit a ceiling effect in the general adult United States population (121). There are also significant differences between the general population of the United States and the United Kingdom in the elicited time trade-off valuations underlying the EQ-5D (122). As a result, United States population-based preference weights for the 243 EQ-5D health states have recently been determined, thereby enhancing the internal validity of this widely applied instrument (123).
The Short-Form-6D.
The Short-Form-6D (SF-6D) converts six of the eight health dimensions, including bodily pain (Table 5), on the SF-36 into a single health utility score (124). Its scoring function is linear additive and the underlying utility is based upon standard gamble preference scores obtained from a random sample of adults in the United Kingdom (99). One advantage of the SF-6D is that it allows for the conversion of existing SF-36 data into single aggregate health utility scores for any population without directly measuring utility (31). However, due to its modeling parameters, the SF-6D has a more restricted range of health utility scores (0.30–1.0) and thus may not detect changes in patients in very poor health due to a resulting floor effect (61).
Comparing the Preference-Based Measures of HRQOL
A rigorous comparative review was undertaken of the QWB, HUI3, EQ-5D, and SF-6D (108). Even though all four of these preference-based HRQoL instruments "are designed to measure the same concept, each uses a different model of health, a different method of deriving preferences, and a different scoring formula" (108). Thus, not surprisingly, applying the QWB, HUI3, EQ-5D, and SF-6D can result in very divergent health utility scores being generated for the same health state (108).
Although none of the four preference-based measures was deemed by the reviewers to be the "best" instrument, several important selection recommendations were made. The EQ-5D and to a lesser extent the HUI3 exhibited a ceiling effect, making both less likely to discriminate among those individuals in very good health, while the QWB is the least likely of the four to exhibit a ceiling effect. In patients in poor health, using the SF-6D to convert an SF-36 health status score into a health utility score is likely to produce higher values than would have been obtained from the other three preference-based measures (108).
Moreover, because the assessment of pain is handled differently by these four preference-based instruments, the generated utility scores are not comparable. When applied in a large population of patients with a confirmed diagnosis of intervertebral disk herniation, spinal stenosis, or degenerative spondylolisthesis, baseline values on the EQ-5D, HUI2, HUI3, and SF-6D were significantly divergent, leading the authors to conclude that it may be inappropriate to compare health state values derived using different preference-based HRQoL instruments (125). No one instrument was found to be superior; therefore, researchers were advised to choose the preference-based measure that best fits the study design and the clinical condition under investigation (125).
Given their widespread use and their lack of an observed floor effect, and their ability to generate negative aggregate utility scores for particularly severe chronic pain health states, the EQ-5D and the HUI3 appear to be the more practical options for generating health utility scores for a chronic pain condition.
PEDIATRIC HRQOL
Growing attention has been focused on the measurement of HRQoL in children and adolescents (6,126–135). Routinely assessing pediatric HRQoL can facilitate patient–physician communication, improve patient and parent satisfaction, identify hidden morbidities, and assist in clinical decision-making (6). The measurement of HRQoL in pediatric clinical trials and clinical practice nevertheless remains limited as compared with the adult population (6,131).
Unique conceptual and methodological issues have impeded the standardized assessment of HRQoL in chronically ill children and adolescents, including those suffering from chronic pain (132,136). A fundamental challenge in assessing pediatric HRQoL is the central role of child development and the associated dynamic social and psychological contexts in which a child or adolescent perceives health versus disease (137). This methodological challenge can be overcome, however, if sufficient consideration is given to the choice of an age-appropriate pediatric HRQoL instrument (128,138,139).
The use of an adult generic HRQoL measure should be avoided in chronic diseases of childhood because of its likely failure to tap important pediatric health domains and its response burden (140). A parental proxy assessment of a younger child’s HRQoL is a viable alternative. However, children aged 8–11 yr appear to report significantly lower HRQoL than their parents (141). Therefore, whenever possible, the pediatric patient’s own health perceptions in addition to those of a parental proxy should be elicited (127). If the health survey instrument is appropriately structured, a child as young as 4 yr can provide meaningful, even if only concrete, insight into their self-perceived health status. More subjective or abstract health domains can be self-reported by individuals aged 8 yr and older (126,132).
Pediatric Generic HRQOL Measures
On the basis of the two criteria of documented instrument reliability and validity and the availability of both a patient self-report and parental proxy report form, combined with the results of the present literature search, four generic pediatric generic HRQoL measures appear to warrant consideration in pediatric chronic pain medicine (Table 7).
|
Child Health and Illness Profile
The Child Health and Illness Profile (CHIP) (http://www.chip.jhu.edu/) has been primarily applied as a population-based measure of the effects of specific pediatric health service interventions on health status and behavior (34,142–144). While the CHIP includes the HRQoL domain of physical comfort, it does not directly assess pain. In keeping with its public health origin, the CHIP may be more applicable in the study of child health care policy and the effects of not only medical but also public health and other community interventions on children’s long-term health (145).
Child Health Questionnaire
The Child Health Questionnaire (CHQ) (http://healthactchq.com/) examines a wide array of physical and psychosocial domains, including bodily pain and discomfort (33,146). The CHQ has been shown to be a valid HRQoL tool in children with sickle cell disease, from both the patient and parental perspectives (147,148). However, given its substantial length and attendant respondent burden with serial administration, the CHQ is likely better applied as a population-based pediatric health screening survey (149).
The KINDL
The KINDL (http://www.kindl.org/) encompasses psychological well-being, social relationships, physical function, and everyday life activities (35). Only one item on the KINDL instruments directly assesses pain (150). In addition to conventional paper-and-pencil versions, two very appealing, computer-assisted, animated renditions of the KINDL known as the CAT-Screen are available (http://www.catscreen.de/). The KINDL has been applied in epidemiological studies of child and adolescent health, clinical studies involving acutely and chronically ill children, and in a pediatric rehabilitation program setting (150).
Pediatric Quality of Life Inventory
The Pediatric Quality of Life Inventory (PedsQL) (http://www.pedsql.org/) is a series of tools designed to measure the four primary chronic pain-related clinical outcomes of pain intensity, HRQoL, the family impact of a chronic disease, and parental satisfaction with medical care. The PedsQL 4.0 Generic Core Scales are the principal component of the PedsQL (32). Pain is specifically addressed within the physical function domain of the PedsQL Generic Core Scales.
The PedsQL Generic Core Scales have been successfully applied in a prospective analysis of pediatric cancer-related pain and emotional distress (151). The PedsQL Generic Core Scales have been used as well to demonstrate a longitudinal relationship among HRQoL, pain, and coping strategies in a cohort of 8- to 18-yr-olds with juvenile idiopathic arthritis (152). The PedsQL 4.0 Generic Core Scales have also been successfully incorporated into clinical decision making in cardiology, orthopedic, and rheumatology outpatient settings (153).
The Pediatric Application of Preference-Based Health Measures
There has been minimal experience in applying preference-based HRQoL measures in children and adolescents (154,155). Though their use has been restricted primarily to adolescents, the HUI2 and the EQ-5D have been the most widely applied multiattribute, preference-based HRQoL instruments applied in the pediatric population (155–157). Of note, the HUI2 was actually devised to comprehensively describe the health status of survivors of childhood cancer (158).
WHAT IS A CLINICALLY SIGNIFICANT DIFFERENCE IN HRQOL?
Just as with any clinical research or patient care parameter, one needs to determine the threshold for a clinically significant difference in HRQoL scores (159,160). To provide greater meaning to HRQoL scores, investigators have applied two distinct strategies to identify this significance threshold: anchor-based methods versus distribution-based methods (161,162).
Anchor-Based Methods to Establishing Interpretability
Anchor-based approaches to interpreting HRQoL scores are in turn classified as single anchor methods, which are longitudinal and individual patient-focused, and multiple anchor methods, which are cross-sectional and population-focused (161,162). In light of its more tangible nature and innate clinical applicability, only the single anchor method will be discussed here.
Clinicians and patients naturally categorize an observed or reported difference in a clinical parameter as trivial, small but important, moderate, or large. The single anchor method identifies the demarcation between the trivial difference and the small but important difference, which is referred to as the minimum important difference (MID) in a HRQoL score (161).
The MID has been explicitly defined to be the smallest difference in the score on the health domain of interest "that patients perceive as important, either beneficial or harmful, and which lead the clinician to consider a change in the patient’s management" (161). The MID is essentially synonymous with the minimum clinically significant difference (163).
The MID has been determined for the most commonly used indirect preference-based measures of HRQoL (Table 6) (99,109). As in other clinical research fields, statistical significance in HRQoL studies is predicated largely on sample size and is not necessarily equivalent to clinical significance (164).
Distribution-Based Method to Establishing Interpretability
The distribution-based method for interpreting observed changes in HRQoL scores relies upon basic descriptive statistics, specifically, the sample sd, to describe the magnitude of effect or change (161,162). Conventionally, changes in the range of 0.2 sd represent a small change, those in the range of 0.5 sd represent a moderate change, and those in the range of 0.8 sd represent a large change (161,165). An inherent advantage to the distribution-based method of clinically interpreting health-status scores is the ready availability of a descriptive measure of sample variability. Thus, in the absence of anchor-based interpretive guidelines, a 0.5 sd likely represents the MID in the HRQoL score (166).
CONCLUSION
There are means to routinely assess HRQoL in clinical practice, including in pain medicine. However, the full potential of patient-reported HRQoL has yet to be fully realized (167). The routine use of patient-reported HRQoL in clinical practice will not significantly improve the process or the outcomes of patient care until existing logistical, communication, knowledge, and attitudinal barriers are overcome (167–169). This primer is intended to help overcome these barriers to the more widespread clinical use of HRQoL measures in pain medicine.
When the above-noted Medical Outcomes Trust instrument attributes and review criteria (49) were systematically applied to a series of widely recognized HRQoL measures, a group of leading health outcomes scholars concluded that the decision to use one instrument over another, to use a combination of two or more instruments, to use a preference-based measure, or to use a generic measure alongside a condition-specific measure should be determined primarily by the clinical or research purpose of the measurement (60).
ACKNOWLEDGMENTS
The author thanks Victoria L. Phillips, DPhil and Steven D. Culler, PhD, in the Department of Health Policy and Management at the Rollins School of Public Health, Emory University, Atlanta, GA, for her scholarly support, and their collective insightful and valuable critique of this manuscript.
APPENDIX: THE IDENTIFIED HEALTH-RELATED QUALITY OF LIFE REVIEW ARTICLES
As part of the objective selection process applied to determine which specific adult and pediatric health-related quality of life measurement instruments to include in this primer, 33 review articles from a variety of disciplines were identified as focusing on specific generic and preference-based HRQoL measurement instruments (see bibliography below).
Footnotes
Accepted for publication November 20, 2006.
Reprints will not be available from the author.
BIBLIOGRAPHY
