Within the last few decades, outcomes research, and in particular quality of life (QoL) outcomes research, has become integrated into clinical and research practices. QoL outcomes are important to dermatology, because many diseases carry significant psychosocial burdens and morbidity from appearance and symptoms with few cases of mortality. In this article, the authors discuss the future directions in QoL. Important areas are the determination, clinical significance and interpretation of measured QoL values. Additionally, the development of proxies for preference-based QoL measures and modules for health status QoL measures are discussed.
Within the last few decades, outcomes research, and in particular quality of life (QoL) outcomes research, has become integrated into clinical and research practices. This change has transformed medicine from an old model, which had an emphasis on targeting objective measures such as blood pressure, to a new model where subjective measures such as QoL are of importance. Many dermatologic conditions are associated with clinical findings and symptoms that can negatively impact health-related QoL. This change in QoL is often the reason for which care is sought. Furthermore, the experience of health-related QoL is different for each person. This experience may be related to the patient’s disease or treatment, and it may incorporate differing cultural and historical experiences. For this reason, treatments are not only targeted to impact objective findings, but also to those that are more subjective such as itch, which is often the most bothersome aspect of the disease to the patient. Said another way, dermatologic diseases carry significant psychosocial burden and morbidity from appearance and impact on symptoms, function, and emotion, with few cases of mortality. Therefore, it is important within dermatology to address and incorporate QoL into clinical and research practices.
Fortunately, much work has been performed in the development and application of different types of QoL measures in dermatology. The paradigm of generic health versus skin-specific versus condition-specific health status instruments has been reviewed extensively elsewhere. Similarly, criteria as to how to choose a particular health status QOL instruments have been explored in other well-developed articles.
The next generation of dermatology QOL work, in the authors’ opinion, should focus not only on validating health status instruments and be rigorously tested for psychometric properties, but also concentrate on methods to assign clinical meaning to these instruments and to explore guidelines for the development of disease-specific measures. Additionally, less work has been performed on health preference measures of QoL (vs health status measures) that can be incorporated into cost-effectiveness analyses. In this atmosphere of limited health care resources, health policymakers will be forced to curtail certain treatments; dermatology needs to be able to join in these discussions by providing quantifiable metrics to incorporate the QoL improvement treatments offer. More specifically, outcomes researchers need to explore approaches to estimate preference-based QoL measures that are not as cumbersome as current methods to maximize available cost-effectiveness analyses.
Clinical significance of Q o L measures
The principles of evidence-based medicine are used increasingly in dermatology to guide clinical practice and resource allocation. With outcomes research blossoming, more emphasis is being placed on the results of these studies to inform clinical practice. The use of valid and clinically meaningful patient outcomes measures is vitally important to reaching this goal with quality and success. Therefore, there is a need to standardize scoring systems for measuring disease severity and other patient outcomes in the clinical trial setting. Lack of standardization and validity of measures can prohibit the meaningful interpretation of individual studies. Furthermore, wide variations of trial methodology that limit the comparison of data from different sources impact the production of therapy guidelines, compendiums of evidence, and systematic reviews. It has been noted that within the literature, there are almost as many scales as there are trials. Signs and symptoms are frequently mixed up together, and patient-centered outcomes measures, despite the awareness of QoL outcomes, seem to be neglected by an urge to measure objective signs. The clinical meaning of percentage changes in continuous objective scales will always be difficult to interpret and need to be interpreted in conjunction with other outcome measures, as in practice it is the patient who is treated and not the signs of the disease. Despite the known importance, QoL is assessed in a surprising minority of trials.
Interpreting results from outcomes data
Many health related QoL studies have focused on cross-sectional comparisons between groups, but great interest lies in the assessment of intraindividual change over time. To demonstrate the value and success of various treatments and care protocols, the research needs to show that the observed changes in patients’ outcomes are important and clinically substantial. Most studies attempt to assess clinical efficacy of a given intervention compared with another by making group comparisons and evaluating for statistically significant differences. Unfortunately, statistical significance does not in itself provide concise information about a given intervention’s clinically meaningful effects.
One approach is to assign clinical meaning to bands of scores of the QoL instrument such that 1 band of scores can be interpreted to have a particular amount of QoL impact. Hongbo and colleagues were among the first to apply this approach in dermatology. By anchoring the QOL scored to a global question of overall impairment to QOL, they were able to assign clinical meaning to bands of scores of the Dermatology Life Quality Index; for example, scores from 0 to 10 represented “no to a small QoL impact.” Thus, a change in score from 12 to 8 may represent a clinically meaningful change, while a change from 10 to 6 would not be clinically meaningful even if it were statistically significantly different. Other groups have applied this concept to the Skindex. The authors refer readers to Rogers’ article in this issue for a thorough review.
While using clinical bands of scores is a useful first step to interpreting QoL instruments, those methods do not provide information for the minimally important difference (MID), that is the smallest change in QoL scores that the patients perceive is important. This change may be within a clinical band, or may straddle between 2 bands. This determination of the magnitude of intraindividual change necessary to establish clinical relevance remains to be determined for most measures. The methods for linking statistical evaluations and clinically meaningful standards for change are areas of future research. Although there are no agreed upon standards, the question of how to meaningfully interpret changes in QoL scales can be addressed with many existing methods. The authors outline several approaches and examples that future investigators may want to explore for existing validated dermatology measures.
Anchor-Based Approach for MID
One approach to establish the MID is to compare with an independent standard or anchor that is itself interpretable and to which the instrument under investigation bears at least a moderate correlation. Possible anchors for dermatology include changes in global ratings by subjects, change from systemic or high-potency topical treatments to low-potency topical maintenance treatments, fewer office visits, fewer missed days from work, or other improved measures of function. In this research, there is a shift from a focus on the mean difference to the difference in the proportion of patients who experience an improvement greater than the MID in the treatment and control groups. Said in another way, if the treatment is effective, more treated than control patients will show an improvement.
Reliable Change Index
The reliable change index (RCI) calculates whether the change in score from before to after treatment is caused by real change or by chance variation. The RCI is calculated using the standard error of measurement (SEM) to estimate the range of chance variation. RCI tables can be generated for the overall and subscales scores of the QoL measure. The RCI alone does not indicate clinical significance; it only expresses the amount of change between before and after treatment scores that would be statistically reliable.
Distribution-Based Methods of Estimating Clinical Significance
Distribution-based strategies estimate clinically significant change by using normative distribution statistics. One main approach involves estimating effect size in change scores, which involves subtracting an individual’s baseline scale score from the follow-up score, then dividing those results by the measure’s normative sample standard deviation. One weakness with this method is that the standard deviation, used in calculating effect sizes, is sample dependent and therefore not necessarily externally generalizable.
The other main distribution-based method involves using the SEM, because it theoretically does not vary sample to sample, and may be a more stable. The SEM is an estimate of the amount of test error associated with the test taker’s true score and is determined by multiplying the standard deviation of the measure by the square root of 1 minus its reliability coefficient. One SEM is then interpreted as the MID in individual patient scores. With this method, a range of true scores can be determined, and scores that fall outside that range would be considered rare scores. Another variant of the distribution approach was based on observations that studies that demonstrated one-half of the standard deviation of a measure represents the upper limit of the MID
Applications in Dermatology
Several of these approaches to determine clinical meaning in QoL scores have been applied in dermatology. The MID concept has been applied using the Dermatology Life Quality Index (DLQI) in chronic idiopathic urticarial. In this work, the authors used data from a phase 2 clinical trial and used both the anchor and distributional approaches for deriving estimates of MID. For the anchoring technique, the authors used a patient-assessed 5-point pruritus score as an anchor. Changes in the DLQI total score that correlated to changes in disease severity, defined as a mean change score in pruritus greater than or equal to 1, were considered the MID. The authors used the SEM method for the distribution-based approach. They found an estimate of 3.21 and 2.97 using the anchor approach and 2.24 to 3.1 using the distributional approach. They concluded by recommending a range of 2.24 to 3.10 as an MID for the DLQI in interpreting results for patients with chronic idiopathic urticaria.
Shikiar and colleagues estimated MIDs for the DLQI in psoriasis. Since there was no patient-based assessment of the smallest change that they would perceive to be beneficial, the data were used to explore 5 methods to estimate MIDs of the DLQI. The first 3 approaches involved anchor-based methods. They used 2 definitions of significant changes in Psoriasis Area and Severity Index (PASI) scores and compared DLQI change scores. They first defined significant PASI change on the percentage of patients who would perceive (25%) as beneficial and then defined PASI change on what clinicians would deem clinically relevant (50%–74%). The third anchor-based method for estimating MID relied on the association of changes of the DLQI with changes in the Physician’s Global Assessment (PGA). A minimal responder was defined as a patient whose PGA improved by either 1 or 2 points from baseline to week 12, whereas a nonresponder would have no difference. The DLQI MID was defined as the difference in the DLQI score between nonresponders and minimal responders.
The last 2 MID estimates were based upon the distributional methods. One used the method using the SEM, while the other estimated the MID using one-half of the standard deviation. Estimates for the DLQI MID ranged from 4.05 (for patient-based PASI difference) to 6.95 (for clinical PASI difference), while the SEM was 2.33, and one-half standard deviation was 3.59. In the conclusion, the authors disclaimed the clinical PASI difference and stated that they believed that it was too conservative for estimating the minimum change that patients will find beneficial. Instead, they modified their estimates and concluded that the MID of the DLQI in psoriasis should be in the range of approximately 2.3 to 5.7.
Related posts:
Health-Related Quality of Life in Dermatology: Introduction and Overview
Overview of Health Status Quality-of-Life Measures
Quality-of-Life Issues in Vitiligo
Melasma Quality of Life Measures
Health-Related Quality-of-Life Assessment in Dermatologic Practice: Relevance and Application
Clinical Meaning in Skin-specific Quality of Life Instruments: A Comparison of the Dermatology Life Quality Index and Skindex Banding Systems
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
