This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Somerfield, M. R. Articles by Loprinzi, C. Search for Related Content PubMed PubMed Citation Articles by Somerfield, M. R. Articles by Loprinzi, C. Social Bookmarking                            What's this?

Wherefore Global Quality-of-Life Assessment?

American Society of Clinical Oncology, Health Services Research, Alexandria, VA

To the Editor: The recent article by Sloan et al¹ offers a comparative analysis of four global quality-of-life (QOL) instruments. The study reported by Sloan et al is methodologically sound, but I question the validity of their conclusions about the value of the single-item, global Spitzer Uniscale (UNISCALE) tool they endorse for assessing change in QOL, as well as their corresponding claims about the clinical utility of this measure.

On the basis of analyses of changes in QOL scores over time, Sloan et al argued that the UNISCALE measure, a single-item tool that asks patients to rate their QOL during the past week, is superior to the longer (22-item) Functional Living Index-Cancer (FLIC) measure for assessing change in QOL. The UNISCALE's greater sensitivity to change in QOL in their sample of patients with advanced colorectal cancer is a function of the length of this tool, according to Sloan et al: The UNISCALE demonstrated greater sensitivity because change in just one item is required for a change to appear empirically. Obversely, the multi-item Functional Living Index-Cancer is less sensitive to change because multiple items have to change for a change in QOL to appear empirically; in addition, there is the risk of having change in one item of a multi-item measure "wash out" change in another item of that measure.

This line of reasoning is troublesome. The strategy of using a single, global QOL measure to assess change would seem to place a premium on brevity at the expense of clinical utility. Few could challenge the suggestion that a single, global QOL measure is to be preferred over a multi-item measure for detecting change, if the primary interest is in detecting change for the sake of detecting change. But it is generally believed from a clinical standpoint that global measures of QOL are of little to no value for assessing change. The reason is that one cannot determine the source of the change in a global score from one assessment to another.² Of note, the lack of change in a global score from one time to the next is as difficult to interpret.

Just what do people call to mind when they are asked to rate their overall QOL? No one knows for certain. In the Sloan et al sample of patients with advanced disease (median survival of 4.5 months), a fair number of patients probably answered this item in terms of their current physical functioning; this might explain the concordance rate between the UNISCALE and Eastern Cooperative Oncology Group performance status (52%, not the 38% listed in Table 2). Other patients might have answered in terms of emotional well-being, others in terms of spiritual issues, others in terms of family functioning, and yet others in terms of some "mental calculus"³ through which they attempted to provide a synthesis of their QOL across various dimensions. Is the referent for the rating the same from time 1 to time 4? The fact remains that, with global items, no one knows what individuals are referring to from one assessment to the next, which renders interpretation of change with global measures impossible. This same point has been made with respect to global assessments of coping with cancer, a research area replete with largely uninterpretable findings of little or uncertain clinical value.⁴

Is there a more informative and clinically useful strategy that is still brief enough to be used by practitioners? Is the development of a "simple yet meaningful"¹ tool for routine clinical assessment of multidimensional QOL even a realistic goal? This has not been studied adequately and should be high on the agenda of QOL researchers.⁵

There has been consensus for years that QOL measures should assess physical, emotional, and social well-being, that they should be multidimensional.³ Viewed in this light, the continued investigation of global QOL is puzzling at best. At worst, continued investigation of global QOL could undermine efforts to introduce or maintain multidimensional assessments of QOL in cancer clinical trials to the extent it reinforces perceptions that QOL data offer nothing of practical clinical value.

REFERENCES

1. Sloan JA, Loprinzi CL, Kuross SA, et al: Randomized comparison of four tools measuring overall quality of life in patient with advanced cancer. J Clin Oncol 16:3662-3673, 1998[Abstract]

2. Cella DF, Bonomi AE: Measuring quality of life: 1995 update. Oncology 9:47-60, 1995 (suppl)[Medline]

3. Lazarus RS: Theory-based stress measurement. Psychol Inquiry 1:3-13, 1990

4. Somerfield MR: The utility of systems models of stress and coping for applied research: The case of cancer adaptation. J Health Psychol 2:133-151, 1997[Abstract]

5. Ganz PA: Impact of quality of life outcomes on clinical practice. Oncology 9:61-65, 1995 (suppl)[Medline]

Mayo Clinic, Rochester, MN

In Reply: We thank Dr Somerfield for his thought-provoking letter regarding our recent article.¹ This provides us with the opportunity to revisit some of the major tenets of the work and to present some supplementary material. An important point to make is that the results presented in the present article have been replicated recently using three other clinical trials and two other QOL instruments (the Functional Assessment of Cancer Therapy-General and the European Organization for Research and Treatment of Cancer [EORTC] QOL questionnaire [QLQ] C30).² Patient populations involved were advanced lung, colorectal, and various advanced tumors.

First, and most importantly, as we stated in the article, we are not eschewing the use of multi-item QOL assessment tools, nor are we suggesting that a global score is sufficient to garner the details of QOL change characteristics. Rather, our intent was to assess the relative merits of using a single item versus many items to assess global QOL.

Somerfield's impression that "the only purpose of the global score is to detect change" needs to be clarified. Norms have been established for global QOL scores in the literature.³ In particular, the issues of adaptation and response shift have demonstrated that even in acutely ill individuals, the global QOL score can be surprisingly high, centering around the 75th percentile.⁴ Hence, if one were to observe a global score at baseline for an individual well below such a level, then the result in itself is worthy of investigation. We need not wait for a change to take place.

At the heart of the matter is the tradeoff between patient burden and the level of detail required. While it would be inappropriate to always recommend the single-item approach, we believe it is equally untenable to suggest that a multi-item tool covering all aspects of QOL should be applied across all situations so that sufficient detail is obtained. First, it is a fallacy that the more questions you ask the more information you obtain.⁵ Response set, patient burnout, and redundant and irrelevant questions are all sources of concern with the indiscriminate use of multi-item tools. Furthermore, there is no clear mechanism for synthesizing the multidimensional results obtained from more complex instruments into clinically meaningful data.⁶

It is fairly well established that QOL is a multidimensional construct. It is also unlikely that every aspect of QOL will indicate changes for a given clinical situation or individual. For example, some of the work that we have done with our hospice population has indicated that, other than physical functioning variables, QOL among hospice patients remains remarkably stable from time of admission through to death.⁷ Hence, in the same manner that some follow-up test procedures (eg, computed tomography and magnetic resonance imaging) may be ordered only after an abnormality has been discovered by a more global and less detailed test procedure (eg, history and/or physical examination), so should QOL be applied in an efficient manner with consideration given to patient burden.

We disagree that, in general, "global measures of QOL are of little to no value for assessing change" (Somerfield's letter). In particular, the argument that one does not know "what people call to mind when they are asked to rate their overall QOL" could apply to any subjective measurement tool and undermines the entire research process of gathering information from patients in a reliable and valid manner. There is an element of trust in any patient communication that the patient understands the question and is able to provide accurate, reliable, and valid data. Without this assumption, we could not believe any subjective data garnered in any study. The fact that the UNISCALE and Eastern Cooperative Oncology Group performance status show a reasonable degree of concordance is in fact evidence of construct validity in that the UNISCALE is measuring something related to, but not redundant with, overall QOL.

Somerfield raises an important question regarding the possibility of finding an "informative and clinically useful strategy that is still brief enough to be used by practitioners." As our article (and others) indicates, it is possible to obtain initial estimates of patient QOL via simple, single-item tools. The clinical utility is inherent. If a patient provides an overall QOL score indicative of "normal" or "average" levels for a particular instrument, then no further energy needs to be expended exploring the myriad of QOL subconstructs. If, however, a patient reports a level of QOL that is either "abnormal" or "unusual," this could spark an obvious question as to the source and precise nature of the problem and might indicate a need for further clinical investigation. This would seem to us to be a simple and clinically relevant use of global QOL indicators and is analogous to what is routinely done with other clinical variables.

It is certainly not our intent to "undermine efforts to introduce or maintain multidimensional assessments of QOL in cancer clinical trials" by examining the application of global QOL tools. We see the two approaches as complementary, not competing. Other authors have pointed out that the economic and patient burden must be balanced against the need for QOL data within considerable fiscal restraints of oncology clinical trials.⁸ Not all QOL investigations need a complete multidimensional assessment to obtain clinically valuable information, nor will a single-item global score be sufficient to provide detail about complex interactions of the multidimensional constructs where such assessments are needed. Similarly, not all studies need to include a complete Minnesota Multiphasic Personality Inventory to diagnose a personality disorder. In fact, the MOS-SF36 is still one of the most often-used health status questionnaires in clinical trials despite the introduction of its more brief counterpart, the SF12.

For clarification, the concordance rate listed in Table 2¹ is correctly listed as 38%. The first value in each cell of Table 2 is a simple concordance proportion and the second is Kendall's coefficient, as is stated in the article.

Finally, it would seem that if nothing else, this discourse has indicated that there is much more work to do in the area of defining precisely how to measure patient QOL. We look forward to more fruitful discussions of complementary approaches as we move forward collegially in an attempt to answer this question.

REFERENCES

1. Sloan JA, Loprinzi CL, Kuross SA, et al: Randomized comparison of four tools measuring overall quality of life in patients with advanced cancer. J Clin Oncol 16:3662-3673, 1998

2. Sloan JA, Sargent DJ, Mahoney M, et al: When less is more: The role of single item global QOL assessments in oncology clinical trials. Proc International Society for Quality of Life Research, Baltimore, 1998, pp 662-663

3. Spilker B (ed): Quality of Life and Pharmacoeconomics in Clinical Trials. New York, NY, Raven, 1996

4. Bullinger M: Quality of life assessment in palliative care. J Palliat Care 8:34-39, 1982

5. Gonin R, Lloyd S, Cella D: Establishing equivalence between scaled measures of quality of life. Qual Life Res 5:20-26, 1996[Medline]

6. Aaronson NK: Methodological issues in assessing the quality of life of cancer patients. Cancer 67:844-850, 1991[Medline]

7. Sloan JA, Rummans TR, Bretscher M, et al: Measuring quality of life in hospice patients: Results of a pilot study. Presented at Manitoba Provincial Palliative Care Conference, Winnipeg, Canada, September 25-27, 1997

8. Klausner RD, Hubbard SM, Chappell J: Quality of life in clinical cancer trials. Monogr Natl Cancer Inst 20:1-111, 1996

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Somerfield, M. R. Articles by Loprinzi, C. Search for Related Content PubMed PubMed Citation Articles by Somerfield, M. R. Articles by Loprinzi, C. Social Bookmarking                            What's this?