Abstract
Health-related quality of life (HRQOL) is an important measure of a patient’s perception of his/her illness. Over the past 3 decades, numerous instruments have been developed to measure HRQOL in various patient populations, with 2 basic approaches: generic and disease-specific. While generic measures have broad application across different types and severity of diseases, disease-specific measures are designed to assess particular diseases or patient populations. All HRQOL instruments, however, must be valid and have high reliability and responsiveness. Validity ensures that the instrument measures what it is supposed to measure. Reliable instruments are able to reproducibly differentiate between subjects. Responsive evaluative measures are able to detect important changes in HRQOL during a period of time, even if those changes are small. HRQOL measures should also be interpretable, meaning that the differences in scores that correspond to small, moderate, and large HRQOL changes are easily identifiable. This article describes the steps in the development of HRQOL instruments from the conceptual framework to creation and testing. Several examples of generic and disease-specific instruments commonly used to evaluate HRQOL in patients with immune-mediated inflammatory diseases (IMID) are provided.
- HEALTH-RELATED QUALITY OF LIFE
- RESPONSIVENESS
- RELIABILITY
- GENERIC MEASUREMENT INSTRUMENTS
- DISEASE-SPECIFIC
Health-related quality of life (HRQOL) is a multidimensional concept referring to patients’ perceptions of the influence of disease and treatment on their physical, psychological, and social function and well-being1. To that end, various HRQOL instruments have been developed2. HRQOL questionnaires can be self- or interviewer-administered. They are used to measure either differences in quality of life between patients at any point in time or changes in HRQOL within patients during a certain period of time3.
HRQOL instruments are valuable for a number of purposes, including evaluating the impact of disease and/or treatment on a patient’s overall well-being in clinical trials and daily practice4. HRQOL tools can also be used for assessing health-related gaps across different segments of the population and for measuring and comparing the effectiveness of healthcare interventions for various conditions5. Further, both clinicians and policymakers are recognizing the importance of measuring HRQOL in the daily management of patients with various diseases, as well as in making policy-related decisions3. In the current environment with drugs and biologics that cost more than older compounds, improvements in HRQOL are usually a requirement for coverage of these medications. HRQOL instruments can be important tools to confirm that highly statistically significant improvement in a disease is associated with improvement in HRQOL.
INSTRUMENT DEVELOPMENT
The process of the development of a HRQOL instrument entails a series of steps, including:
-
Identification of concepts and development of a conceptual framework
-
Instrument creation
-
Assessment of instrument properties, and
-
Instrument modification6.
Each step requires a different study construct. While the initial phases seek to establish the content and the format of the instrument, the later phases aim to determine the measurement properties of the instrument.
Identification of concepts and development of conceptual framework
The initial, and probably the most important, step in instrument development is to identify concepts and domains of importance7. The most commonly considered domains include psychological, physical, and social functioning as well as somatic comfort7,8,9. When deciding on these, one must consider the intended population (i.e., adults, children), condition, timeframe, and research application. Expected relationships among concepts are to be hypothesized.
Instrument creation
During this phase, a first draft of the instrument is generated. Table 1 outlines a list of items and considerations included in the process.
Items for inclusion in the measure are usually derived from a literature review, accompanied by a series of interviews with patients and experts in the field. Administrative methods, response period, and response scales are chosen. It is important to keep a manageable number of items by identifying and choosing those that are (1) most pertinent to the purpose of the measure, (2) most important to the patient, and (3) most frequent. The selection of the number of items is generally performed by a panel of experts in a particular therapeutic field. Thereafter, the instrument is formatted and the instructions are drafted. It is important to ensure that the items and their description make sense, are easily understood, and are easy to complete within an allocated timeframe. The types of scoring of an instrument can vary, namely: a single rating on a single concept; a single score combining multiple ratings of related domains or individual concepts (index); multiple uncombined scores of multiple related domains (profile); multiple uncombined scores of independent concepts (battery); or a combination of single rating, index, profile, or battery (composite). Once the initial version is developed, it must undergo pilot testing, after which the instrument and the procedures are refined.
Assessment of instrument properties
The third phase in the development of an HRQOL instrument includes testing to ensure its reliability, validity, and ability to detect change. The reliability of an instrument to consistently measure the characteristic of interest includes the concepts of both reproducibility (rater agreement) and internal consistency (correlation among questions composing an instrument so that different questions on the same concept are in agreement). Administrative and respondent burden is evaluated and items are revised accordingly. After the identification of meaningful differences in scores, the format, scoring procedures, and training material are finalized.
Instrument modification
Finally, during the modification phase, the concepts measured, populations studied, research application, and instrumentation or method of administration can be modified further, if necessary.
EVALUATION OF AN INSTRUMENT AND ITS PROPERTIES
Once the initial version of an instrument is developed, it must be tested to ensure that it is reliable, valid, and responsive to changes.
The Outcome Measures in Rheumatoid Arthritis Clinical Trials (OMERACT) filter
OMERACT is an international organization that focuses on rheumatology outcome measures10. To be accepted as an OMERACT-endorsed outcome measure for use in a clinical trial, the instrument must pass through the OMERACT filter in its intended setting10. The filter has 3 component criteria, each represented by a question to be answered about that measure: truth, discrimination, and feasibility.
In addition, the following 5 criteria should be applied (Table 2): (1) Feasibility ensures that the instrument is used efficiently. It takes into consideration administration time, reading and understanding level, and multicenter administration.
(2) Reliability certifies that the instrument consistently measures the characteristics of interest. It includes reproducibility and internal consistency. Repeated testings are performed to verify test-retest, intrarater, and interrater reliability. In summary, a reliability value of 0.00 indicates absence of reliability and value of 1.00 means perfect reliability. A reliability coefficient > 0.70 is desirable as it implies that 70% of the measured variances are reliable and 30% are owed to random error11. Cohen’s κ coefficient is a statistical measure of interrater agreement and values > 0.60 are desirable as they show good to excellent agreement between the 2 raters’ scores12.
Internal consistency represents the correlations among items composing an instrument. It verifies whether several items that propose to measure different aspects of the same trait produce similar scores. The goal of a reliable instrument is for scores on similar items to be related, but for each to contribute some unique information. Internal consistency is usually measured with Cronbach’s α, which ranges between 0 and 113. A rule of thumb is that α = 0.7 indicates acceptable, and α = 0.8 represents good reliability. However, reliabilities of α = 0.9 are not necessarily desirable, as this indicates that the items may be redundant.
(3) Validity confirms that the instrument measures what it is supposed to by demonstrating appropriate correlations with other measures, usually based on a prior prediction of the degree of such correlations. Validity of content, criterion, and construct are the standard categories.
(4) Responsiveness verifies the ability of an instrument to detect small but clinically important change. This is particularly important where subjective reports of health status are one of the primary outcomes of the trial. There are several ways to present responsiveness: difference from baseline, treatment difference, relative percentage improvement, mean response, and relative efficiency. Receiver operator characteristic curves (Figure 1)14 evaluate how a given change in score can discriminate patients who improve from those who do not.
(5) Interpretability ensures that the minimal clinically important difference (MCID) is established. Table 3 provides an overview of established MCID for 2 commonly used HRQOL instruments: the Medical Outcomes Study Short-Form Survey (SF-36) and EuroQoL (EQ-5D)15,16,17,18,19,20.
GENERIC VERSUS DISEASE-SPECIFIC INSTRUMENTS
HRQOL measures can be divided into generic and disease-specific21.
Generic instruments
Generic instruments are designed to assess HRQOL in a broad range of populations with or without chronic illness. While generic instruments may not be sufficiently sensitive to detect changes in HRQOL in any specific illness, they are used to compare HRQOL in populations with different diseases. Further, when these measures are applied to healthy populations, normative data can be gathered. These data can then be used to compare the burden of disease of a specific condition to that of other chronic illnesses, as well as to healthy controls. Commonly used generic measures include the SF-3622 and EQ-5D23.
The SF-36 is a short-form measure of generic health status in the general population (Table 4). It is designed for self-administration and it can be administered to anyone over the age of 14 years. It consists of 36 items divided into 8 health profiles. The SF-36 has been translated and adapted in many countries.
EQ-5D consists of descriptions of health status according to 5 dimensions (Table 5)24. Each dimension is divided into 3 levels. By combining different levels from each dimension, EQ-5D defines a total of 243 health states. These states may be converted to a score using “sets of values” derived from general population samples. EQ-5D provides a simple descriptive profile and generates a single index value for heath status on which full health is assigned a value of 1 and death a value of 0.
Disease-specific instruments
Disease-specific instruments focus on concerns relevant to a particular illness. These instruments measure changes in HRQOL over time or with treatment, which is not possible with generic measures. Some of the common disease-specific instruments used in IMID include: ASQoL for ankylosing spondylitis25, PsAQoL for psoriatic arthritis26, RAQoL for rheumatoid arthritis27, PDI (Psoriasis Disability Index)28, IBDQ (Inflammatory Bowel Disease Questionnaire)29, and IBDQoL30.
Specific instruments can also be considered in a broader context, namely system- or organ-specific instruments. For example, the most frequently used instrument to assess HRQOL in patients with psoriasis is the Dermatology Life Quality Index (DLQI). This is neither a generic (it is centered on skin) nor a disease-specific instrument. An advantage with this type of questionnaire is its application for assessment of other dermatological diseases; physicians not involved in research prefer it to using and interpreting 10 or 20 different disease-specific questionnaires. However, it also has disadvantages. For example, some skin diseases also involve other organs (psoriasis involves skin and joints) and the DLQI does not assess the influence of the joint component on QOL.
CONCLUSION
HRQOL instruments allow broader assessment of the effects of disease and intervention on patients. They are applicable to all phases of trial assessment and require active patient participation. They also provide a standardized tool for comparison with other studies. The incorporation of HRQOL instruments in a study improves the likelihood of uptake by decision-makers and healthcare providers. Results obtained from QOL assessments are also considered valuable and are often used in pharmacoeconomic evaluations.
Footnotes
-
Supported by an unrestricted grant from Abbott Canada. Dr. Wells has received funds from BMS, Abbott, Amgen, and UCB. Dr. Russell has received consultant fees from Amgen/Wyeth, Schering-Plough, UCB, and Abbott Canada, and speaking fees from Roche and BMS. Dr. Haraoui has acted as consultant for Abbott Laboratories, Amgen, Bristol-Myers Squibb Canada, Roche Canada, Schering-Plough Canada, UCB Pharma, and Wyeth Pharmaceuticals; received grant/research support from Abbott Laboratories, Amgen, Bristol-Myers Squibb Canada, Roche Canada, Schering-Plough Canada, UCB Pharma, and Wyeth Pharmaceuticals; and received honoraria from Abbott Laboratories, Amgen, Bristol-Myers Squibb Canada, Roche Canada, Schering-Plough Canada, UCB Pharma, and Wyeth Pharmaceuticals. Dr. Bissonnette received research grants or honoraria from Abbott, Amgen, Astellas, Merck, Pfizer, and Schering.